Inventory and monitoring of wildlife habitat

BLM LIBRARY

88048148

^.---. is ■ '

'wJPJPP"*^!??*!

, I,:'' -

: ■&%&& \

mg&mMm

■

I WM

■ *■

INVENarORY fiy^D MOMTORIflllf i

~; PJ

a.; ' <

. $ I _\>

X

T7/* < ! J^

h0 v^

^•."^"i

iaffia&'J^ •-

>: S#

¥*'£**

•■/>"

if/m

Jf s

9TC&

^S

- J/^r

Iff

**m£

i&M

£&j£&

^

!| |^

•«'-V

•

sV-.Tn

...>.

. j • *f/^tei'

*****:*

MK&S

INVENTORY AND MONITORING
OF WILDLIFE HABITAT

This report should be cited as:

Cooperrider, A. Y., R.J. Boyd, and H. R. Stuart, eds. 1986. Inventory and monitoring of wildlife habitat. U.S. Dept. Inter..
Bur. Land Manage. Service Center. Denver, Co. xviii. 858 pp.

BLM/YA/PT-87/001 +6600

V

Lead Editor: •*>•.'"""

Janet Poorman j

Copy Editor: '1

Margaret McGinnis
Contract Editor:

William Dryer
Section paintings, chapter and cartoon illustrations

Peter Doran and Shirley McCulloch
Technical illustrations and cover:

Peter Doran
Graphs:

Geri McBeth and Laura Cuplin
Computer graphics

Rose Maruska
Chapter pasteups and chapter illustrations:

Laura Cuplin
Drawings for Amphibians and Reptiles chapter:

Lauren Porzer Keener
Graphics assistance:

Keith Francis and Herman Weiss
Editorial assistance- and word processing:

Marilyn Chatterton, June Johnson, Geri McBeth, and Margaret Trujillo
Computer assistance

Phyllis Ellu >tt and Surma Reitsma

For sale by the Superintendent of Documents. U.S. Government Printing Office, Washington, D.C. 20402

& ■* ° Q^

INVENTORY AND MONITORING
OF WILDLIFE HABITAT

Compiled and Edited by

Allen Y. Cooperrider, Raymond J. Boyd,
and Hanson R. Stuart

Design by

Shirley L. McCulloch

U.S. Department of the Interior • Bureau of Land Management • September 1986

TABLE OF CONTENTS

Preface a

Acknowledgments xi

Reviewers xni

Introduction XVH

I PLANNING

Chapter 1 Inventory and Monitoring Process 1

K Bruce Jones

Chapter 2 Data Types 1 1

K Bruce Jones

Chapter 3 Literature Review 29

Ora Wagoner

Chapter 4 Habitat Mapping 49

Richard M. Kerr

II MAJOR HABITATS

Chapter 5 Forests 73

Jack Ward Thomas and Jared Venter

Chapter 6 Rangelands 93

Henry L. Short

Chapter 7 Deserts 123

A] Bruce Jones

Chapter 8 Tundra 149

Peter C. Lent

Chapter 9 Riparian Habitats 169

Robert D. Ohmart and Bertin W. Anderson

Chapter 10 Marshes 201

Milton W. Welter

Chapter 1 1 Streams 225

Paul Cuplin

Chapter 12 Lakes 237

James F. LaBounty

III SPECIES GROUPS

Chapter 13 Fish 257

Paul Cuplin

Chapter 14 Amphibians and Reptiles 267

A.' Bruce Jones

Chapter 15 Songbirds 291

Ronald A Ryder

Chapter 16 Raptors 31 3

Michael Kochert

Chapter 17 Marsh and Shorebirds 35 1

Peter G. Connors

Chapter 18 Waterfowl 37 1

Robert L. Eng

Chapter 19 Colonial Waterbirds 387

Steven M. Speich

Chapter 20 Upland Game Birds 407

Robert L. Eng

Chapter 21 Rodents and Insectivores 429

Mayo Call

Chapter 22 Lagomorphs 453

Joseph A Chapman and Gale R Willner

Chapter 23 Carnivores 475

Richard A Spowart and Fred B. Samson

Chapter 24 Bats 497

Stephen P. Cross

Chapter 25 Ungulates 519

Raymond J. Boyd, Allen Y. Cooperrider,
Peter C Lent, and James A Bailey

IV HABITAT MEASUREMENTS

Chapter 26 Soils 567

James E. Stone

Chapter 27 Terrestrial Physical Features 587

Allen Y. Cooperrider

Chapter 28 Aquatic Physical Features 603

Paul Cuplin

Chapter 29 Hydrologic Properties 613

Bruce Van Haveren

Chapter 30 Water Quality 633

Paul Cuplin

Chapter 31 Vegetation 639

Bertin W. Anderson and Robert D. Ohmart

Chapter 32 Macroinvertebrates 661

Fred Mangum

V SPECIAL STUDIES

Chapter 33 Radiotelemetry 679

Paul L. Hegdal and Bruce A Colvin

Chapter 34 Food Habits 699

Allen Y. Cooperrider

Chapter 35 Weather and Climate 711

James A Bailey

VI ANALYSIS AND PRESENTATION

Chapter 36 Data Management 727

Larry Peterson and Iris Matney

Chapter 37 Statistical Analysis 741

William H. West

Chapter 38 Habitat Evaluation Systems 757

Allen Y. Cooperrider

Chapter 39 Evaluation and Interpretation
Allen Y Cooperrider

Chapter 40 Economic Analysis 785

John Loom is

Chapter 41 Written Communications 805

Donald E. Zimmerman

Chapter 42 Verbal Presentations 829

Eugene Decker

Glossary 841

Index 853

PREFACE

This book is intended to guide field biologists
and managers in planning, organizing, and adminis-
tering wildlife inventory and monitoring projects.
Although primarily designed for the professional
wildlife biologists working for the U.S. Bureau of
Land Management, we believe much of the material
will interest those working in other agencies and
institutions as well.

The book reviews current general procedures
and specific techniques. It is not a "cookbook," how-
ever, nor an attempt to standardize techniques
among biologists. We believe that providing a set of
rigid rules for inventorying and monitoring wildlife
habitat would be presumptuous since nature itself is
complex, diverse, and dynamic. We firmly believe
the professional judgment of trained biologists is
critical in land management. Their expertise cannot
be replaced by modern technology. Although mod-
ern technology can make some of their work easier,
only the biologist can —

• identify biological problems;

• ensure the correct problem has been identified;

• collect the information needed to address a
problem; and

• analyze, interpret, evaluate, and effectively
present information to influence management
decisions.

Management biologists have demanding, com-
plex, and challenging tasks ; few have the luxury of
specializing. Each must be knowledgeable of many
diverse wildlife groups, other natural resources, and
resource conflicts. Management biologists also need
good communication skills to effectively deal with
concerned citizens, ranchers, miners, administrators,
and scientists. This book is therefore intended to
help these biologists become more knowledgeable of
all facets of wildlife habitat management.

An overriding concern in designing this book
was to cover the process of inventory and monitor-
ing in its entirety, i.e., from initial problem identifica-
tion through presentation of results. We have
observed that many biologists do not allocate their
time and energy well among the various tasks neces-
sary to complete or conduct an inventory or moni-
toring project. Good data are often collected to
solve problems that are trivial compared to the real
priority concerns. Similarly, data are sometimes col-
lected at great expense but are not analyzed, or the

sampling design is so poor that the data cannot be
used. Another common problem is that good data
are collected and analyzed, but the results are never
effectively communicated to the public and the deci-
sionmakers. Therefore, the entire process of con-
ducting an inventory or monitoring a project is
presented. Data collection and analysis are obviously
important and central tasks of such projects, but
they tend to be tasks that biologists perform well.
Other aspects such as design, planning, and commu-
nication, are included to help biologists improve
their skills in these areas.

The book is modular; chapters can be read
either alone or in combination for general guidance
or as a reference source. As a reference source, a
chapter may contain a detailed description of a tech-
nique or refer the reader to another source contain-
ing a description. This depends on the amount of
work available on the subject, the detail involved,
and the availability of good descriptions in readily
accessible publications. For instance, field techniques
for lizard inventory are described in detail since
few good references are available on the subject.
However, a subject such as mark-recapture methods,
to which biostatisticians have devoted whole careers
and about which numerous books have been written,
has brief coverage supplemented by relevant
references.

The book is divided into six major sections.
Section I covers general procedures for planning,
designing, and organizing wildlife habitat inventory
and monitoring programs. These chapters cover such
topics as determining objectives; design; use of exist-
ing literature; habitat mapping; and choice of areas,
measurements, and techniques.

Section II provides guidelines for inventorying
and monitoring particular habitats. Each chapter
includes classification systems, species groups and
habitat features that are particularly important, major
impacts on such habitats, particular problems of
inventory or monitoring, and habitat analysis sys-
tems. The chapters in this section are designed for
biologists who are planning inventory or monitoring
studies in habitats they have not worked in
previously.

Section III provides guidance on inventorying
and monitoring habitat for particular animal groups.
The coverage of species groups is not organized
along taxonomic lines, but rather taxa for which sim-

ilar techniques are used are lumped together. For
each species group, the chapter includes a descrip-
tion of ( 1 ) habitat components, e.g., vegetation
structure, physical features, that correlate with spe-
cies in this group and (2) current techniques for
population inventories and monitoring.

Section IV describes techniques for measuring
habitat variables. These chapters include detailed
descriptions of actual measurement techniques and
identify advantages and limitations.

Section V covers special monitoring studies.
These include methods of food habits determination,
climatological studies, and studies of movement and
habitat use. Such information is generally used with

other habitat or population measurements to deter-
mine the cause of observed changes.

The final section, Section VI, covers techniques
and procedures for analysis, evaluation, interpreta-
tion, and presentation of data and results.

To try to obtain a publication that is as readable
and accessible as possible, we have at times edited
or rewritten chapter material. In some cases, in the
interest of time, we have not been able to return
materials to authors for review. We have made every
effort not to alter meaning or emphasis; nevertheless
some errors may have occurred. For these we apolo-
gize to authors and readers and hope that such cases
are few.

ACKNOWLEDGMENTS

A work such as this could not be completed
without the assistance and support of many individu-
als too numerous to mention. In addition to the staff
members listed and the authors themselves, other
reviewers have contributed substantially to the pub-
lication. We sincerely appreciate their efforts to help
make the publication more complete, accurate, and
readable. They are listed separately in the following
section.

Certain other individuals deserve particular rec-
ognition. Mayo Call, formerly of the U.S. Bureau of
Land Management (BLM), originally recognized the
need for such a publication and, before he retired,
convinced BLM management to support this effort.
The editors left with the task have alternately cursed
and thanked him for this effort. In the long run, he
deserves the credit for the idea, but in the short run,
we have often wished that he was here to share the
frustrations.

John Crawford, Chief of Wildlife for the BLM
from 1974 to 1985, was largely responsible for pro-
viding the moral and financial support for this effort.
During his tenure as Chief of Wildlife, the BLM wild-
life program developed from infancy to a substantial
force in multiple-use management. John's commit-
ment and effort was largely responsible for this de-
velopment. He recognized the need for sound
inventory and monitoring of wildlife habitat on pub-
lic lands and for providing technical support to field
biologists through publications of this sort.

The original format for the publication was de-
signed with the assistance of Paul Cuplin, Kniffy
Hamilton, Bruce Jones, Richard Kerr, and Chris
Maser.

The continued support of members of the BLM
wildlife staff in Washington, David Almand, Nancy
Green, Kniffy Hamilton, Alan Kesterke, Neal Middle-
brook, and William Radtkey, is greatly appreciated.
The ongoing support of members of the Denver
Service Center — Directors Delmar Vail, Neil Morck,
and Robert Moore, and Chief of the Office of Tech-
nology Transfer, Mary Gaylord — is appreciated. The
senior editors (Cooperrider and Boyd) would like to
express particular appreciation to their supervisors,
John Baker and Allan Strobel, for their consistent
support and encouragement throughout the entire
project, and also extend a special thank you to Rob-
ert Ader, who delayed several major projects so
some of his staff could help with this book.

Faye Leonard assisted with the contracting ef-
forts competently and cheerfully. Carl Anderson
provided much assistance with the typesetting and
printing. His knowledge of the technical aspects
of printing was invaluable and is much appreciated.

Finally, the editors gratefully acknowledge the
assistance and contributions of the staff of the BLM
library, particularly Sandra Bowers, Teresa Day, and
Judy Moisey.

xi

Use of trade or firm names in this publication is for the information and convenience of the reader. Such use does not
constitute an official endorsement or approval by the U.S. Department of the Interior of any product or service to the
exclusion of others which may be suitable. The U.S. Department of the Interior does not assume any liability or responsi-
bility for damages caused by reliance on the materials presented in this publication.

All scientific and common names of animals and plants used in this publication were verified according to the following
authorities:

Plants — Checklist of North American Plants for Wildlife Biologists. Prepared by T.G. Scott and C.H. Wasser. 1980. The
Wildlife Society, Washington, DC.

Macroinvertebrates — Fresh-water Invertebrates of the United States. Prepared by R.W. Pennak. 1953 The Ronald Press
Company, New York, NY.

Fish — A List of Common and Scientific Names of Fishes from the United States and Canada. Prepared by C.R. Robbins, R.M.
Bailey, C.E. Bond, JR. Brooker, E.A. Lachner, R.N. Lea, and W.B. Scott. Special Publication 12, 4th edition. 1980. Ameri-
can Fisheries Society, Bethesda, MD.

Amphibians and Reptiles — Standard Common and Current Scientific Names for North American Amphibians and Rep-
tiles. Prepared by J.T. Collins, R. Conant, J.E. Huheey, J.L. Knight, EM. Rundquist, and H.M. Smith. Herpetological Circu-
lar 12, 2nd edition. 1982. Society for the Study of Amphibians and Reptiles, Ohio University, Athens.

Birds — Checklist of North American Birds. Prepared by The Committee on Classification and Nomenclature of the Ameri-
can Ornithologists' Union. 6th edition. 1983- Allen Press, Inc., Lawrence, KS.

Mammals — Revised Checklist of North American Mammals North of Mexico. 1982. Prepared by J.K. Jones, Jr., D.C. Carter,
H.H. Genoways, R.S. Hoffman, and D.W. Rice. Occasional Papers, the Museum, Texas Tech University. 80. Texas Tech
Press, Lubbock.

Unless otherwise credited, all photographs were obtained from BLM photo libraries.
xu

REVIEWERS

The accuracy, thoroughness, and readability of the entire book has been enhanced greatly by the follow-
ing reviewers who have read and critiqued one or more chapters. Their assistance is greatly appreciated.

Bruce B. Ackerman

Idaho Cooperative Wildlife Research Unit
University of Idaho
Moscow, ID 83843

Layne Adams, Wildlife Biologist
U.S. Bureau of Land Management
Fairbanks District Office
Fairbanks, AK 99797

William F. Andelt, Wildlife Extension Agent
Kansas State University
Cooperative Extension Service
Garden City, KS 67846

Stanley Anderson, Wildlife Biologist

U.S. Fish and Wildlife Service

Wyoming Cooperative Fish and Wildlife Research Unit

Laramie, WY 82071

Gary N. Back, Research Associate
University of Nevada — Reno
Range, Wildlife, and Forestry
Reno, NY 89512

James A. Bailey, Professor of Wildlife Biology
Colorado State University
Department of Fishery and Wildlife Biology
Fort Collins, CO 80523

Robert G. Bailey, Geographer
U.S. Forest Service
Land Management Planning Systems
Fort Collins, CO 80524

Gary Bateman

Northern Arizona University
Department of Biological Sciences
Flagstaff, AZ 86011

Fred T. Batson, Natural Resource Specialist

U.S. Bureau of Land Management

Service Center

Denver, CO 80225-0047

Richard G. Beidleman, Professor of Biology
The Colorado College
Department of Biology
Colorado Springs, CO 80903

John Bosworth, Environmental Coordinator
U.S. Bureau of Land Management
Susanville District Office
Susanville, CA 96130-3730

Sandra L. Bowers, Librarian
U.S. Bureau of Land Management
Service Center Library
Denver, CO 80225-0047

R. Bruce Bury, Zoologist
U.S. Fish and Wildlife Service
Denver Wildlife Research Center
Fort Collins, CO 80524

Jack Chugg, Soil Scientist
U.S. Bureau of Land Management
Washington Office Range Staff
Washington, DC 20240

Lewis M. Cowardin, Wildlife Biologist
U.S. Fish and Wildlife Service
Northern Prairie Wildlife Research Center
Jamestown, ND 58401

Paul E. Cuplin, Fishery Biologist
U.S. Bureau of Land Management
Service Center
Denver, CO 80225-0047

Robert K. Davis, Wildlife Economist
U.S. Bureau of Land Management
Washington Office Wildlife Staff
Washington, DC 20006

Russell Davis

The University of Arizona

Department of Ecology and Evolutionary Biology

Tucson, AZ 85721

Stephen Destefano

Idaho Cooperative Wildlife Research Unit
University of Idaho
Moscow, ID 83843

Richard W. Dierking, Soil Scientist
U.S. Soil Conservation Service
West National Technical Center
Portland, OR 97209

Mark A. Dimmitt, Curator of Plants
The Arizona-Sonora Desert Museum
Tucson, AZ 85704

R. Michael Erwin, Wildlife Biologist
U.S. Fish and Wildlife Service
Patuxent Wildlife Research Center
Laurel, MD 20811

Claire Farrell

Southern Oregon State College
Department of Biology
Ashland, OR 97520

Stephen A. Flickinger, Professor of Fishery Biology

Colorado State University

Department of Fishery and Wildlife Biology

Fort Collins, CO 80523

Michael Garratt, Statistician
U.S. Bureau of Land Management
Service Center
Denver, CO 80225

Kenneth M. Giesen, Wildlife Researcher
Colorado Division of Wildlife
Wildlife Research Center
Fort Collins, CO 80526

Nancy F. Green, Wildlife Biologist
U.S. Bureau of Land Management
Washington Office Wildlife Staff
Washington, DC 20006

Carole K. Hamilton, Wildlife Biologist
U.S. Bureau of Land Management
Washington Office Wildlife Staff
Washington, DC 20006

John Haugh, Vertebrate Ecologist
U.S. Bureau of Land Management
Washington Office Resource Sciences Staff
Washington, DC 20006

Steve Hawks, Wildlife Biologist
U.S. Bureau of Land Management
Susanville District Office
SusanvUle, CA 96130-3730

Dale Hein, Professor of Wildlife Biology
Colorado State University
Department of Fishery and Wildlife Biology
Fort Collins, CO 80523

Marian Hershcopf, Librarian
Colorado Division of Wildlife
Research Center Library
Fort Collins, CO 80526

Dale Hoffman, Limnologist
U.S. Fish and Wildlife Service
Region 6
Denver, CO 80525

Rick Hoffman, Wildlife Researcher
Colorado Division of Wildlife
Wildlife Research Center
Fort Collins, CO 80526

Jon K. Hooper, Coordinator, Parks and Natural Resources

Management Option
California State University, Chico
Department of Recreation and Parks Management
Chico, CA 95929-0560

Marshall A. Howe, Section Leader, Migratory Nongame

Birds
U.S. Fish and Wildlife Service
Patuxent Wildlife Research Center
Laurel, MD 20811

Jack D. Jones, Wildlife Biologist
U.S. Bureau of Land Management
Butte District Office
Butte, MT 59702-3388

K. Bruce Jones, Wildlife Biologist
U.S. Bureau of Land Management
Phoenix Training Center
Phoenix, AZ 85016

William Kepner, Fishery Biologist
U.S. Bureau of Land Management
Phoenix District Office
Phoenix, AZ 85017

William L. Kirk, Habitat Specialist
U.S. Fish and Wildlife Service
Anchorage, AK 99503

James F. LaBounty, Fishery Biologist
U.S. Bureau of Reclamation
Denver, CO 80225

Joe Lint, Wildlife Biologist
U.S. Bureau of Land Management
Roseburg District Office
Roseburg, OR 97470

CD. 1 ittk field
Malheur Field Station
Princeton, OR 97721

Larry S. Mangan, Wildlife Biologist
U.S. Bureau of Land Management
Shoshone District Office
Shoshone, ID 83352

Chris Maser, Research Wildlife Biologist
U.S. Bureau of Land Management
Corvallis, OR 97331

Ken McGinty, Technical Editor
U.S. Bureau of Land Management
Service Center
Denver, CO 80225-0047

Gary McVicker, Natural Resource Specialist
U.S. Bureau of Land Management
Washington Office Renewable Resources Staff
Washington, DC 20240

Brian A. Millsap, Raptor Biologist
National Wildlife Federation
Raptor Information Center
Washington, DC 20036

Sam Montgomery, Wildlife Biologist
U.S. Bureau of Land Management
Colorado State Office
Denver, CO 80205

James A. Mosher, Consultant
Savage River Consulting
Frostburg, MD 21532

Lewis Nelson, Jr., Professor

University of Idaho

College of Forestry, Wildlife, and Range Sciences

Moscow, ID 83843

Nick Nydegger, Wildlife Biologist
U.S. Bureau of Land Management
Boise District Office
Boise, ID 83705

Richard R. Olendorflf, Wildlife Biologist
U.S. Bureau of Land Management
California State Office
Sacramento, CA 85825

Lewis W. Oring, Professor
The University of North Dakota
Department of Biology
Grand Forks, ND 58202

Ed Parsons, Natural Resource Specialist
U.S. Bureau of Land Management
Washington Office Range Staff
Washington, DC 20006

Howard Quigley

Idaho Cooperative Wildlife Research Unit
University of Idaho
Moscow, ID 83843

William Radtkey, Wildlife Biologist
U.S. Bureau of Land Management
Washington Office Wildlife Staff
Washington, DC 20006

Harry V. Reynolds, Game Biologist
Alaska Department of Fish and Game
Division of Game
Fairbanks, AK 99701

Richard T. Reynolds, Research Wildlife Biologist
U.S. Forest Service

Rocky Mountain Forest and Range Experiment Station
Fort Collins, CO 80526-2098

Chandler S. Robbins, Research Biologist
U.S. Fish and Wildlife Service
Patuxent Wildlife Research Center
Laurel, MD 20708

John T. Rotenberry, Assistant Professor
Bowling Green State University
Department of Biological Sciences
Bowling Green, OH 43403

Robert Roudabush, Soil Scientist
U.S. Bureau of Land Management
Arizona State Office
Phoenix, AZ 85011

Hal Salwasser, National Wildlife Ecologist
U.S. Forest Service

Wildlife and Fish Ecology Unit
Fort Collins, CO 80524

Henry L. Short, Terrestrial Ecologist
U.S. Fish and Wildlife Service
Western Energy and Land Use Team
Fort Collins, CO 80526-2899

Ronnie Sidner

The University of Arizona

Department of Ecology and Evolutionary Biology

Tucson, AZ 8572 1

Karen Steenhof, Wildlife Biologist
U.S. Bureau of Land Management
Boise District Office
Boise, ID 83705

George A. Swanson, Wildlife Biologist
U.S. Fish and Wildlife Service
Northern Prairie Wildlife Research Center
Jamestown, ND 58401

Charlene Swibas, Technical Information Specialist
U.S. Environmental Protection Agency
National Enforcement Investigations Center
Denver, CO 80225

Michael Tewes

Idaho Cooperative Wildlife Research Unit
University of Idaho
Moscow, ID 83843

S.P. Thompson

U.S. Fish and Wildlife Service
Stillwater Habitat Management Area
Fallon, NV 89406

Roy E. Tomlinson, Southwest Dove Coordinator
U.S. Fish and Wildlife Service
Office of Migratory Bird Management
Albuquerque, NM 87103

Jared Verner, Research Ecologist
U.S. Forest Service
Forestry Sciences Laboratory
Fresno, CA 93710

Lee F. Werth, Natural Resource Specialist
U.S. Bureau of Land Management
Service Center
Denver, CO 80225-0047

U.W. Wilson

U.S. Fish and Wildlife Service
Olympia, WA 98504

Jerry O. Wolff, Assistant Professor of Biology
University of Virginia
Department of Biology
Charlottesville VA 22901

Leonard S. Young, Wildlife Biologist
U.S. Bureau of Land Management
Boise District Office
Boise, ID 83705

INTRODUCTION

Inventory and monitoring of wildlife and wild-
life habitat is often considered mundane and uninter-
esting work, yet many great discoveries in both
general biology and more specifically wildlife biology
were made while biologists and naturalists per-
formed these "mundane" tasks. Charles Darwin and
Alfred Wallace were both conducting "inventories"
when they developed their concepts of evolution
and natural selection. In more recent times, mem-
bers of the U.S. Biological Survey, the forerunner of
the U.S. Fish and Wildlife Service, inventoried wild-
life and wildlife habitat of western North America,
contributing significantly to our understanding of the
ecology and zoogeography of these regions. Al-
though much of their work involved collecting and
documenting evidence of fauna in various regions,
their contributions went far beyond merely listing
the animals present.

Consider the contributions of C. Hart Merriam
(Merriam's Life Zones), Olaus Murie (The Elk
of North America and Alaska- Yukon Caribou ), or

Aldo Leopold, who developed many revolutionary
ideas about such subjects as predation and wildlife
habitat management while performing duties as a
forester for the U.S. Forest Service in the Southwest
and later directing game surveys in the Midwest.
One can, of course, argue that these men were doing
work more akin to research, particularly later in
their careers. This may be partly true, but the dis-
tinction between research and the activities of inven-
tory and monitoring can be nebulous. In any case,
the skills required for successfully conducting an
inventory or monitoring study are similar to those
needed for successful research. Both require an abil-
ity to systematically make and record observations,
conceptually and quantitatively synthesize large
amounts of data, and effectively communicate, both
verbally and in writing.

Inventory and monitoring of wildlife habitat is a
statutory responsibility of many land management
agencies including the U.S. Bureau of Land Manage-
ment and the U.S. Forest Service.

Inventory and monitoring of wildlife habitat is
based on the assumption that measurements of a set
of habitat attributes can be used to predict presence
or abundance of wildlife species. The same relation-
ship between habitat variables and wildlife popula-
tions is used for both habitat inventory and
monitoring. This relationship has long been recog-
nized by naturalists and is the basis of several recent
habitat evaluation systems. Habitat variables may

consist of vegetative structure, plant species compo-
sition, presence of physical features, or other factors.
Predictions resulting from such measurements need
to be verified by observing presence or measuring
relative abundance of the wildlife species of interest.
In some cases these measurements need to be taken
concurrently to determine the most relevant habitat
variables. Thus species or population measurements
are a necessary part of development and use of any
system for inventory and monitoring of wildlife
habitat.

Wildlife habitat inventory consists of measuring
selected habitat variables on a piece of land to infer
presence or abundance of wildlife species. The pur-
pose is to determine the wildlife resources currently
supported by the area.

Wildlife habitat monitoring consists of repeat-
edly measuring habitat or population variables to
infer changes in the capability of the land to support
wildlife. The purpose is usually issue-oriented, i.e.,
it determines how some activity such as mining,
livestock grazing, or recreational activity is affecting
wildlife habitat and ultimately wildlife populations.
Monitoring is also used to determine the effective-
ness of habitat management practices such as seed-
ing, burning, or water development. Although
monitoring is designed to detect changes from hu-
man activities, it may also detect changes caused by
climatic conditions or other factors not under hu-
man control. Therefore, the purpose of monitoring is
not only to measure change but also to determine
the cause(s) of change. In particular, programs for
monitoring wildlife habitat must be designed to sepa-
rate the effects of human activity from those caused
by variations in weather.

Typically, a monitoring program consists of
measuring habitat variables that are required by key
wildlife species or correlate with presence or abun-
dance of key wildlife species. For example, the num-
ber of pileated woodpeckers highly correlates with
the number of suitable nesting snags. A reduction in
the number of such snags can thus be assumed to
cause a reduction in the number of breeding
woodpeckers.

Habitat monitoring, like inventory, may be ac-
companied by population measurements to confirm
the habitat relationships. If the purpose is to measure
the effectiveness of a habitat management practice,
then species or population measurements must be
taken. For example, if water is being developed to

improve habitat for Gambel's quail, then monitor-
ing should be designed to measure changes in the
number of quail, not changes in the number of
"guzzlers."

Over 50 years ago, Aldo Leopold, in a chapter
on "Game Range" in his classic volume on Game
Management wrote:

"What is Game Range? When the game man-
ager asks himself whether a given piece of land
is suitable for a given species of game, he must
realize that he is asking no simple question,
but rather he is facing one of the great enigmas
of animate nature. An answer good enough for
practical purposes is usually easy to get by
the simple process of noting whether the spe-
cies is there and ready, or whether it occurs on
'similar' range nearby. But let him not be cock-
sure about what is similar,' for this involves the
deeper question of why a species occurs in
one place and not in another, which is proba-
bly the same as why it persists at all. No living
man can answer that question fully in even one
single instance."

When Leopold referred to a biologist determin-
ing "similar" ranges, he introduced the concept of
a "habitat model" and described the reasons moni-
toring is necessary.

Habitat models can of course be of many types.
Hunters often use conceptual models for determin-
ing, for example, high density autumn pheasant habi-
tat when passing some fields and hunting in others.
Hunter perceptions of good habitat may be accurate
even though they are unable to describe specific
qualities when choosing sites. After a day in the field,
a hunter may reflect on the day afield and suggest
that "millet fields adjacent to uncultivated swales and
breaks are the best places to hunt." This is a verbal
model. A biologist may go further and describe opti-
mum habitat as "grain fields of 40 acres or more
within Vz mile of water and within 'A mile of uncul-
tivated areas, having brushy cover at least 4 feet
high." This is a quantitative model which can be sys-
tematically tested and evaluated.

Two points deserve noting about habitat models
and their use by biologists and others. First, we all
use models in our everyday life, including habitat
models. Many are good, although based only on
"professional judgment," i.e., conceptual models.
Second, with the controversy surrounding many cur-
rent biological assessments, biologists are being
asked, encouraged, or even required to make their
models more explicit by various parties, interest
groups, publics, and agencies involved with or af-
fected by project decisions. Explicit and quantitative
models not only provide a mechanism for improving
predictions, but also provide an important communi-
cation tool by taking some of the mystery out of
"professional judgment."

But models have their limits. Models are approx-
imations at best. As Aldo Leopold suggested, no liv-
ing person can develop a "perfect" habitat model.
However, monitoring can allow us to test and refine
our models in addition to minimizing the risk of
resource damage from poor predictions. We can thus
have a cycle of prediction, decision, action, and
monitoring.

This approach to natural resource problems has
received a lot of attention in the last 10 years and
has been variously described as adaptive environ-
mental assessment, cyclic-incrementalism, "muddling
through," or even "common-sense management," as
discussed in several chapters in this book. It seems
to be a logical and common-sense approach to solv-
ing a wide variety of natural resource problems.
With such an approach, the separate but interrelated
tasks of problem identification, data collection, pre-
diction, decision, and monitoring can also be linked
in a cyclical process. This process can result in bet-
ter management of wildlife resources and also devel-
opment of better tools (models) for understanding
wildlife habitat relationships. Although many of the
following chapters describe isolated parts of this
process, they should be viewed in context, not as
ends in themselves, but as necessary steps in a con-
tinuing process toward better management of our
natural resources.

Allen Y. Cooperrider

xviii

I PLANNING

1 The Inventory and Monitoring Process

2 Data Types

3 Literature Review

4 Habitat Mapping

TFIF

INVENTORY
AND

MONITORING
PROCESS

Editor's Note: This chapter provides an overview of
the entire inventory and monitoring process, with
emphasis on the planning stage. Biologists in their
hurry to get to the field to gather the all-important
data often neglect to spend the necessary time in
planning their work. They may also neglect to al-
low time for the necessary follow-up procedures,
such as data analysis. Such actions result in much
wasted time and money. This chapter not only pro-
vides an overview of the entire process, but also
directs the reader to other sections and chapters in
the book that cover specific topics in more depth.

K. Bruce Jones

U.S. Bureau of Land Management
Phoenix Training Center
Phoenix, AZ 85015

"The amount of time spent in solving a problem is
inversely proportional to the amount of time spent
in characterizing the problem."

— Alan Speigel, Speigel's Laws of Management

"Planning without action is futile. Action without
planning is fatal."

— K. Hamilton and E.P. Bergersen, Methods to
Estimate Aquatic Habitat Variables

INTRODUCTION

The most critical stage of implementing and
completing an inventory or monitoring study is not
data collection, presentation, or interpretation, but
rather design. Years of data can be useless if a study
is poorly designed. This chapter outlines a procedure
for any inventory or monitoring project. By follow-
ing these steps, biologists will establish a flexible,
systematic, and logical approach toward solving wild-
life habitat management problems. Use of this proce-
dure should also increase the effectiveness of
inventory and monitoring studies, reduce costs, and
decrease failures caused by poor design or imple-
mentation. The process will also lead to improved
interpretation and presentation of results. By focus-
ing inventories and monitoring studies on specific
issues, wildlife resources should receive greater con-
sideration in the decisionmaking process and realize
greater benefits. Therefore, issue-driven inventories
and studies, such as resolving conflicting land-use
opportunities, are emphasized.

The inventory and monitoring process consists
of a series of events:

(1) Problem definition (scoping) —

• identifying issues, concerns, or opportuni-
ties;

• reducing general problems to specific ones;

• predicting and analyzing extent of potential
problems;

• identifying specific inventory and monitor-
ing study objectives;

• prioritizing objectives; and

• deciding types and levels of data needed.

Current address: Office of Endangered Species, U.S. Fish and
Wildlife Service, Washington, DC. 20240.

Inventory and Monitoring

(2) Data collection (inventory, studies);

(3) Data analysis, interpretation, evaluation, pres-
entation, and storage;

(4) Management decision;

(5) Monitoring studies; and

(6) Management review.

Although the stages may be called different
names in different organizations, this approach to
problem solving is quite basic. This chapter covers
this process with emphasis on steps 1, 2, 3, and 5
which are major concerns of staff biologists and
specialists.

PROBLEM DEFINITION (SCOPING)

Problem definition, often termed scoping, deter-
mines the detail of inventory or monitoring studies
and is critical in conducting these efforts success-
fully. It is the process of determining specific prob-
lems and issues that need to be addressed, data
needed to address these issues, and priority issues
and data needs. It consists of six sequential steps as
listed in the Introduction and described here.

Identifying Issues, Concerns, or
Opportunities

Wildlife- and habitat-related problems originate
from the following six general sources:

( 1 ) Need for baseline data for land-use planning;

(2) Proposals from wildlife biologists;

(3) Management proposals from persons outside
the wildlife resource;

(4) Major land-use planning guidance;

(5) Ongoing impacts from resource development,
originating within the agency; and

(6) Ongoing impacts from resource development,
originating outside the agency.

Baseline data for land-use planning are the basic
driving force of wildlife inventories because they
show conditions as they currently exist. Baseline
data are useful in determining potential impacts and
in comparing actual impacts after development.

Proposals from wildlife biologists to improve
habitat for one species may also produce potential

problems for other species. Consider, for example, a
habitat management plan that proposes to convert
pinyon-juniper (Pinus-Juniperus sp.) and chaparral
habitats into grassland for enhancing pronghorn (An-
tilocapra americana) populations. Although this
proposal would benefit pronghorn, it would also de-
crease habitat available to raptors, upland game, and
nongame species.

Proposals by managers outside the wildlife re-
source could cause many potential impacts. These
impacts are identified by biologists, other resource
specialists, and managers. For example, a biologist
identifies a general problem between a mining plan
and wildlife habitat; a proposed open-pit mineral
operation would greatly reduce the amount and
quality of wildlife habitat in the area.

Some of the more general, broad-scale problems
that require inventories and monitoring studies are
identified through land-use planning. For example, an
areawide grazing proposal, if implemented, could
negatively affect several million acres of wildlife
habitat. This type of problem could be identified by
a wide variety of resource specialists within an
agency during the scoping process. Ongoing impacts
or problems warranting inventories and monitoring
can also be identified by resource specialists inside
an agency (e.g., range conservationists) or outside.
For example, biologists within an agency could iden-
tify that current livestock grazing practices are re-
ducing the quality and quantity of forage available to
big game. Similarly, these kinds of problems can be
identified by sources outside an agency, including
other federal agencies, state game and fish depart-
ments, private wildlife organizations (e.g., The Wild-
life Society), and local citizens.

Reducing General Problems to Specific
Ones

After general issues have been identified, the
next step is to break down general problems (as pre-
viously discussed) into specific ones. For example, a
biologist finds a correlation between reduced wild-
life habitat and habitat quality and increased live-
stock grazing. The next step is to identify, more
specifically, factors that might account for relation-
ships between grazing pressure and loss of wildlife
habitat and quality. In this example, the agency biol-
ogist might identify three specific problems:

( 1 ) Increased stocking of livestock may reduce
the amount and number of perennial plant
species.

( 2 ) Long-term overstocking of livestock may re-
duce the quality of the "A horizon" in the
soil, resulting in large-scale erosion and a
drop in vegetation productivity.

Inventory and Monitoring

Perennial plant species may be reduced by increased live-
stock grazing.

(3) Overstocking and poor livestock distribution
may eliminate cover around waterfowl nest-
ing ponds.

Sections II and III of this book can provide as-
sistance in identifying habitat and species problems.
For example, a new biologist initiating an inventory
in a grassland area with known populations of prairie
chickens (Tympanuchus cupido) may wish to con-
sult Chapter 6, Rangelands, and Chapter 20, Upland
Game Birds, for guidance on habitat and featured
species.

In identifying specific problems from general
ones, biologists should seek expertise from other re-
source specialists. In the example given above, the
soil scientist or soil conservationist plays a major
role in making general problems specific. In this
example, range conservationists, hydrologists, and
surface protection specialists would also play major
roles in identifying specific problems.

Predicting and Analyzing Extent of
Potential Problems

After specific problems have been identified, the
biologist should attempt to make some preliminary
predictions as to the extent of these specific prob-
lems. General predictions can be formulated from a
large variety of information sources:

( 1 ) Existing agency data (e.g., range survey files);

(2) Agency expertise in resources other than
wildlife (e.g., soil sciences);

(3) Review of literature containing data and re-
sults on similar problems (e.g., manuscripts

on impacts to wetland vegetation from live-
stock grazing);

(4) Public input (e.g., observations of citizens
familiar with the problem and area);

(5) Interest group input (e.g., local wildlife advo-
cacy groups familiar with the area and/or
problem);

(6) Experts on the habitat or species in question.

The biologist should make sure he or she has
gathered available information before collecting addi-
tional data. Chapter 3, Literature Review, provides
guidance on searching existing data sources. In addi-
tion to sources listed above, some recently described
procedures may allow biologists to predict specific
problems that could exist between resources and the
extent. Thomas (1979) developed life-form tables
that group species according to components of habi-
tat used for feeding and breeding. Similarly, Short
(1982) developed a procedure that groups species
into guilds based on habitat layers used for feeding
and breeding. Both procedures allow biologists to
use varying degrees of existing information to assess
the complexity of wildlife species arranged in habitat
space. Biologists can then predict the extent of im-
pacts to potential species that would occur if certain
habitat components were lost. Although these proce-
dures neither predict species occurrence in an area
nor generate species-specific data, they do provide
an excellent preliminary analysis of potential impacts
to most wildlife species within a given area.

The U.S. Fish and Wildlife Service and U.S. For-
est Service have developed models that can be used
for analyzing potential impacts to wildlife. Generally,
these models provide an index of habitat quality for
various vertebrates. By knowing the amount and
quality of habitat that may be affected by a proposed
action, the biologist can use these models to deter-
mine corresponding impacts on specific vertebrates.
The use of these models are discussed in greater
detail in Chapter 2, Data Types, and Chapter 38,
Habitat Evaluation Systems.

After compiling information from existing data
sources and making some preliminary predictions,
the biologist will know whether an on-the-ground
inventory or study is needed to adequately assess
or rectify specific problems. Existing data and pre-
liminary analyses may be enough that a biologist
needs only to collect data on one or a few specific
problems.

Based on existing data and preliminary predic-
tions, the agency biologist can provide decision-
makers with information that may affect a proposed

Inventory and Monitoring

action. The impact of the biologist's recommenda-
tions on the proposed action will depend on the
adequacy of existing biological data and economical
and political factors. Preliminary predictions influ-
ence managers to make the following decisions rela-
tive to the proposal:

( 1 ) Drop the proposal entirely;

(2) Leave the proposal as is; or

( 3 ) Modify the proposal so adverse impacts are
reduced or eliminated.

Providing good information to managers at an
early stage can be quite effective in mitigating im-
pacts. Proposals can be dropped or modified before a
project has progressed too far in planning.

Modifying a proposed action often involves pro-
posing new alternatives to mitigate predicted ad-
verse impacts. The degree of mitigation will depend
on the accuracy of existing data and techniques used
to make preliminary predictions. Broad, incomplete
wildlife data used to predict impacts of a site-specific
proposal would not produce accurate assessments.

Identifying Specific Inventory and
Monitoring Study Objectives

At this stage in the process, the agency biologist
has identified specific problems, determined the ade-
quacy of existing data, and made some preliminary
predictions. The biologist will, therefore, have a list
of specific problems requiring on-the-ground data.

To develop a series of procedures and analyses
needed to address the identified problems, the biolo-
gist should transform each specific problem into
specific objectives. For example, a specific problem
was identified when the biologist suspected in-
creased livestock grazing was reducing habitat struc-
ture. He or she must then ask the following types
of questions:

( 1 ) Is livestock grazing reducing habitat struc-
ture?

(2) If so—

• Are stocking rates related to reduced habi-
tat structure?

• How much livestock grazing causes losses
in habitat structure?

• Which plants are structurally affected by
livestock?

These are examples of a few specific questions
that can be asked about a specific problem. The
questions then should be transformed into specific
objectives. For example, determine the plant species
that are structurally reduced by livestock. This infor-
mation can then be used for prioritizing objectives.

Agency biologists must follow this logical proce-
dure in formulating specific objectives, because the
effectiveness of the inventory or study will only be
as good as the biologist's understanding of the prob-
lem(s) and his or her ability to formulate specific
objectives that will answer specific questions.

Prioritizing Objectives

Typically, funding and personnel are inadequate
to collect on-the-ground data for all identified objec-
tives. Therefore, the agency biologist needs to priori-
tize. In many instances, agencies have identified
priority species and habitat; however, the biologist
may need to consider other factors, such as —

• policies, laws, and regulations;

• agency priorities;

• potential degree of impact;

• feasibility (i.e., whether the problem can be
logically, biologically, or economically re-
solved in the time given);

• public interests.

Deciding Types and Levels of Data Needed

Once a decision is made to collect data on
priority objectives, the biologist must decide the
types and levels of data needed. Chapter 2, Data
Types, provides guidance on types and levels of data
needed for inventory and monitoring studies.

DATA COLLECTION (INVENTORY, STUDIES)

After all objectives and methods for an inven-
tory or study are determined, the biologist should
develop a schedule. This schedule should allot
time for —

• hiring or contracting;

• obtaining necessary equipment;

• holding personnel meetings;

• doing preliminary field work;

• selecting specific methods;

• habitat mapping and field sampling; and

• analyzing data and interpreting, evaluating,
and presenting results.

Inventory and Monitoring

Although data analysis, interpretation, evalua-
tion, and presentation are covered separately, any
schedule should include enough time for these criti-
cal tasks. Some efforts require more than one person,
but even a single biologist should follow the same
sequence of events when conducting an inventory or
monitoring study. Section IV in this book, Habitat
Measurements, provides guidance on field data col-
lection methods.

Hiring or Contracting

Competent field personnel are essential in ob-
taining inventory and monitoring study objectives.
The degree of expertise needed will vary greatly,
depending on the type of animals and habitats to be
sampled. Vegetation and topography can be sampled
by a general biologist having a few days of training.
Animal population sampling, however, is specialized
and usually requires someone with formal education
and field experience with the taxa to be sampled.
Contracting may be a way of obtaining the expertise
needed to conduct a highly specialized inventory
or monitoring study. However, contractors generally
cost more than in-house personnel.

Whether an inventory or monitoring study is
conducted in-house or by contract, a biologist must
plan specific hiring needs at least 3 months before
starting. The best possible study design and funding
level could fail if a hiring or contracting plan was
late or poorly organized.

Obtaining Necessary Equipment

The biologist must determine the types of
equipment needed to meet inventory or monitoring
study objectives. The biologist must order any ex-
pensive or hard-to-get equipment at least 3 months
before the inventory or monitoring study.

Holding Personnel Meetings

Preliminary meetings of all involved personnel
are perhaps the most important and essential part of
a successful inventory or monitoring study. At the
first meeting, the lead biologist should introduce
everyone, assure objectives are clear, and outline
agency procedures. The lead biologist should then
request the specialists to review existing data and lit-
erature pertinent to their respective inventory or
monitoring study objectives. For example, a raptor
specialist could gather locality records from all mu-
seums that have made collections in the study area.
He or she should then contact regional and national
experts who have worked with the species in ques-
tion. Literature specific to the species and habitats
should also be reviewed. Team members can learn a
great deal about the general ecology, habitat require-
ments, and population dynamics of species through

literature reviews. This knowledge will also aid them
in selecting the best sampling methods and analyses
for the study.

Doing Preliminary Field Work

While team members are accomplishing all the
tasks outlined in the first meeting, a field trip into
the study area should be conducted to familiarize all
personnel with habitat diversity and size, overall
topography, soil types, elevations and, most impor-
tantly, access. Maps should be supplied to all person-
nel before this initial field trip. Biologists should
also meet with appropriate supervisors and land
users during this period to avoid complications dur-
ing active sampling.

Selecting Specific Methods

After all literature and museum reviews have
been completed, a series of meetings should be con-
ducted to determine the methods that should be
used, the time the monitoring study or inventory
will begin, and the length of the study. At this point,
the team leader should make sure that all team mem-
bers know funding and schedule limitations and reit-
erate the objectives. The following are the types of
questions that should be addressed during this type
of meeting:

( 1 ) What methods will be used for each habitat
and animal type?

(2) What sample size will be needed to meet
each objective?

(3) Which groups of animals and habitats will
be sampled during each season?

(4) What additional equipment is needed (e.g.,
aircraft or four-wheel drive vehicles)?

( 5 ) Can all habitat data be collected by using
one or two methods; will these data be
sufficient for all wildlife groups?

(6) Can all wildlife groups be sampled at the
same location for all habitat types?

(7) What are the major sampling deficiencies?

(8) Is access to sample sites a problem?

(9) Has everyone contacted the appropriate
specialists in other federal and state agen-
cies?

(10) Can other ecological questions be answered
without increasing funding or time?

(11) How will the data be analyzed using the
chosen sampling methods?

(12) How will the data be interpreted and pre-
sented?

Inventory and Monitoring

(13) How and where will raw data, summary
analyses, and interpretations be stored?

(14) How much preliminary sampling is needed?

All of these are extremely important questions
that must be answered before an inventory or moni-
toring study is conducted. Once these questions
are answered, the biologist or team leader should
write an inventory or monitoring plan covering ob-
jectives, methods, data analysis, and storage.

Habitat Mapping and Field Sampling

Habitat Mapping. Following this meeting, a field
trip should be conducted to map habitat types and
locate sample sites. Sample sites should be randomly
located within the area in question. If an inventory
or study is conducted by habitat type, the entire
sample area should be stratified into subareas. Each
subarea would represent a habitat type, and the
number of samples in each would be proportional to
the amount occupied by that habitat type in the
sample area. Subareas may also represent controls
and different intensities of land use. For example, a
biologist may want to develop a study to determine
the effects of fire on vegetation and vertebrate
species. The biologist would randomly sample
vegetation and wildlife within stratified subgroups
such as unburned, burned, and partially burned
areas. Chapter 4, Habitat Mapping, provides guidance
on this important step.

Field Sampling. Once sample sites are located, the
biologist should develop a sampling schedule with
deadlines for completing each sample. At this point,
preliminary samples should be taken. This allows
each biologist to test the methods he or she has
selected and to fine-tune, method-specific skills (e.g.,
bird transects).

During field work, the lead biologist should
conduct weekly meetings to identify problems and
ensure quality control over the specific, planned
objectives.

DATA ANALYSIS, INTERPRETATION,
EVALUATION, PRESENTATION, AND
STORAGE

Once data have been collected, they need to be
analyzed and prepared in a way that provides useful
information to a decisionmaker. These tasks are just
as important as collecting data; unfortunately, many
biologists do not allow or spend enough time on this
follow-up. Section VI in this book provides guidance
on these efforts.

Data Analysis, Interpretation, and
Evaluation

When field work is completed, the team should
meet to analyze the data. The type of analyses used
should have been decided when specific methods
were selected, before the inventory or study.

Once analyzed, the data should be interpreted
and evaluated. The opinions of those biologists who
conducted the inventory or study should be pre-
served because the biologists are most familiar with
the data and circumstances of collection. Chapter 37,
Statistical Analysis, provides guidance on using statis-
tics; Chapter 38, Habitat Evaluation Systems, pro-
vides guidance on systems developed especially for
evaluating habitat; Chapter 39, Evaluation and Inter-
pretation, provides guidance on the more subjective
analysis procedures; and Chapter 40, Economic Anal-
ysis, describes how changes in wildlife productivity
can be used to predict changes in economic value.

Presentation

A biologist must effectively communicate results
to managers and others. For example, detailed statis-
tical analyses and complicated tables and figures
should not be used to convey information to man-
agers. These types of data analyses and displays are
best suited to a specialized technical audience. Sim-
ple, easily accessible tables, figures, and interpreta-
tions should be used for more general audiences.

Generally, documents or manuscripts containing
inventory or study data and interpretations should
be structured in the following format:

( 1 ) Introduction, describing the problem and
the reasons the biologist conducted the in-
ventory or study;

( 2 ) Methods Section, describing procedures
used to collect data; and

(3) Results/Discussion/Summary, describing
results of the inventory or study and an inter-
pretation of those results, followed by a sum-
mary of the results and conclusions (one to
three sections).

A complete manuscript contains all these compo-
nents. This is the best way to ensure all elements of
the inventory or study are preserved for future con-
siderations, especially the biologist's interpretations
of the data. Chapter 41, Written Communications,
provides guidance on preparing written material.

In addition to writing a separate manuscript, the
biologist may be required to orally present results
to managers, the public, and others in briefings, staff

Inventory and Monitoring

meetings, public forums, and other gatherings. Per-
sonal communications are often the most effective,
and the biologist should not overlook their impor-
tance. Chapter 42, Verbal Presentations, provides
guidance on oral presentations.

Data Storage

Data stored in files and on computers preserve
results of inventories and studies. Computers also
allow biologists to continually add information to a
file in a standard format and retrieve it easily. This is
especially important in monitoring studies where
data were collected infrequently or over several
years. Chapter 36, Data Management, provides guid-
ance on managing and storing data.

MONITORING STUDIES

After an inventory or study has been conducted
and the results presented to a manager, a decision
will be made on some action. Such a decision should
be based on particular objectives for wildlife and
other resources, and on the predicted effects of the
chosen wildlife resource alternative. Predicted
effects are normally based on limited data; thus, the
biologist needs to monitor the wildlife resources
to determine if the objectives are being met. This
requires additional data collection and may result in
changes in management. Hence, monitoring is a
cyclic phenomena (Figure 1 ) in which periodic data
collection is followed by re-evaluation of earlier
management decisions.

The design and implementation of monitoring
studies follow the procedure listed earlier in this
chapter. In general, monitoring studies are more site-
and problem-specific than inventories. Inventories
usually involve determining the species and habitats
in an area and, less frequently, habitats used by cer-
tain species. The accuracy of wildlife uses of habitats,
taken on a one-time basis, is usually low because of
limited sample periods. Longer term monitoring
studies better estimate fluctuations in wildlife uses of
habitat. To determine how certain land-use practices
affect wildlife and habitat, the biologist should con-
duct a long-term monitoring study rather than a one-
time inventory. Several inventories should be con-
ducted to better understand the wildlife species and
habitats occurring in an area.

Although monitoring may occur independently
of an inventory (e.g., bald eagle nesting success re-
lated to dewatering of an aquatic habitat ), many
monitoring studies resulting from land-use plans will
be based on inventory results. Therefore, biologists
may need to sample after an inventory has been
conducted in an area.

Generally, monitoring studies involve collecting
wildlife and habitat information over time to deter-

mine-

(1)

(2)

(3)

(4)

Wildlife use of habitat components (e.g., pre-
ferred height of sagebrush for sage grouse
nesting);

Effects of certain land uses on certain wildlife
and habitats:

• individual wildlife species (e.g., how off-
road vehicle [ORV] use affects desert tor-
toise [Gopherus agassizi] density);

• individual plant species (e.g., how intensive
livestock grazing affects key mule deer
browse);

• individual habitat components (e.g., how a
mining operation increases the sediment
load in a trout stream);

• portions of the wildlife community (e.g.,
how a powerline affects migratory raptors);

• entire wildlife communities (e.g., how
water impoundment affects vertebrates in a
riparian community).

Species or habitat changes caused by certain
natural environmental conditions (e.g., effects
of flooding on desert fishes);

Accuracy of predictive models (e.g., compari-
son of white-winged dove [Zenaida asiatica]
responses to habitat changes with those pre-
dicted by a U.S. Fish and Wildlife Service
Habitat Suitability Index [HSI] model);

Off-road vehicle use can damage wildlife habitats.

Inventory and Monitoring

Broad-scale Issue & Problem
Identification

I

Reduction of General Problems
to Specific Ones

Preliminary Predictions & Analysis

Decision Made

with Existing Data

i

Predictions Made
Based on Decis

1

Monitoring Study

Adjust

i

Predictions

No Decision, No New
Data Collection

Decision Adjusted

Based on Monitoring
Data

/

No New Data
Collection

Specific Problem
Identification

n

Decision not Made with

Existing Data-Request -4-
Additional Data (Monitoring
Study or Inventory)

t
oritize Field

Objectives

Collection

1

Decision on Levels and
Types of Data Needed

1

Data Collection (Monitori
Study or Inventory)

1

Data Analysis and
Interpretation/
Recommendation

Figure 1. Inventory and monitoring process.
8 Inventory and Monitoring

( 5 ) Improving the accuracy of predictive models
(e.g., possibly modifying the HSI model for
white-winged dove, by using on-the-ground
data collection);

(6) Additional mitigation to protect wildlife and
habitat in an area (e.g., monitoring data may
have indicated a mitigation plan was not pro-
tecting nesting habitat for black hawks [Bit-
teogallus anthracinus]. From new data, new
stipulations may be recommended to protect
the raptor's nesting habitat.);

(7) Additional habitat improvement to benefit a
species or habitat of concern (e.g., a water
development program may not be increasing
pronghorn herds in an area as predicted.
From the new monitoring data, the plan will
probably need to be modified.).

Some of the above objectives related to improv-
ing a biologist's predictive ability or the models
used. Thus, monitoring can serve dual purposes: ( 1 )

determining whether management objectives are
being met and ( 2 ) improving a biologist's predictive
capabilities. This approach to management and moni-
toring is variously termed "cyclic-incrementalism"
(Bailey 1982), "adaptive-environmental assessment"
(Holling 1978), "adaptive-management" (Barrett and
Salwasser 1982) or, more simply, "common-sense
management" (Barrett and Salwasser 1982) or "mud-
dling through" (Bailey 1982). These approaches are
fundamental to monitoring and are discussed in
more detail in many chapters throughout this book.

SUMMARY

In summary, biologists are strongly advised to
follow the logical thought processes developed in
this chapter. Successful inventories and monitoring
studies do not come easily. Only through careful
planning, precise problem and objective identifica-
tion, and complete compilation of results, will an
inventory or monitoring study be successful.

Inventory and Monitoring

LITERATURE CITED HOLLING, C.S., ed. 1978. Adaptive environmental assess-

ment and management. Vol. 3, International Series on
Applied Systems Analysis. John Wiley and Sons, New
York, NY. 377pp.
SHORT, H.L. 1982. Techniques for structuring wildlife
guilds to evaluate impacts on wildlife communities.
BAILEY, J.A 1982. Implications of "muddling through" for U.S. Dep. Inter., Fish and Wildl. Serv., Spec. Rep. 244.

wildlife management. Wildl. Soc. Bull. 10:363-369. 34pp.

BARRETT, R.H. and H. SALWASSER. 1982. Adaptive man- THOMAS, J W, ed. 1979. Wildlife habitats in managed

agement of timber and wildlife habitat using dynast forests — the Blue Mountains of Oregon and Washing-

and wildlife-habitat relationship models. Western ton. U.S. Dep. Agric, For. Serv. Agric. Handbook 553-

Assoc, of Fish and Wildl. Agencies, Proc. 62:350-365. 512pp.

10 Inventory and Monitoring

DATA TYPES

K. Bruce Jones

Editor's Note: A good, well-planned and organized
inventory or monitoring program cannot be based
on cookbook methodology1. The biologist must use
considerable judgment in identifying problems,
specifying objectives, setting priorities, and choos-
ing methodology. The choice of data to be collected
is critical. Yet the biologist can collect an almost
unlimited number of types of data This chapter
provides an overview of data types and appropriate
situations for collecting a particular type.

U.S. Bureau of Land Management
Phoenix Training Center
Phoenix, AZ 8S015

INTRODUCTION

After identifying specific objectives and ques-
tions for an inventory or monitoring study, a biolo-
gist must decide on the types and levels of data
needed. This decision is often a confusing part of the
inventory and monitoring process. Many different
types of data can be collected to answer specific
questions. In addition, the intensity of data collection
must be determined. For example, an objective is
established to determine the influence of a new
water development on an area's pronghorn (Antilo-
carpa americana) population. Do we need to deter-
mine the density of pronghorns in the area before
and after the development? Do we need a control?
Would relative abundance rather than density an-
swer the question? Decisions made on sampling de-
sign and data types are extremely important. In the
previous example, why spend a great deal of time
and money collecting density data when abundance
information will suffice? If a control was not used,
how will we know if the observed changes in
pronghorn abundance result from our habitat
improvement?

This chapter discusses different types of data
and sampling considerations, and provides some
basic guidelines for making decisions on types of
data and intensity of data collection to use. It also
discusses some procedures that are used to analyze
existing data.

SAMPLING CONSIDERATIONS

Throughout this chapter, I highlight some basic
data types and sampling intensities. Some basic fac-
tors related to sampling design and intensity apply to
all types of data collection.

Before a biologist starts collecting data, a deci-
sion must be made on sampling intensity. The num-
ber of samples required will depend on the
variability of animals and habitats to be measured
and limitations of sampling techniques used. For

Current Address: Office of Endangered Species, U.S. Fish and
Wildlife Service, Washington, DC 20240

Data Types

11

example, homogeneous habitats require fewer sam-
ples because they are less variable. Conversely, het-
erogeneous habitats require more samples because
they are more variable. Sampling techniques and
habits of animals also affect sample size. For exam-
ple, bird transects detect far fewer birds in closed
habitats than in open ones. Therefore, to make the
probability of detection equal in the two habitats,
sample size must be increased. Also, certain bird
species are secretive, requiring more intensive sam-
pling to adequately cover such species.

A number of minimum sample size equations
have been developed for various vertebrates. Most of
these are based on population variability and animal
detectability. Seber (1973), Smith et al. (1975), and
Gysel and Lyon (1980) provide formulas and discuss
how minimum sample size is determined.

Because most inventories and monitoring stud-
ies are restricted by funds and personnel, biologists
are often confronted with deciding on where to
place samples. Generally, samples should be placed
randomly within habitats or areas. Cause-and-effect
studies require biologists to place samples within
preselected areas (stratified sampling). However, bi-
ologists should select random sites within these pre-
selected areas, assuring that area-related variability
is considered. Random sites can be selected by as-
signing numbers to grids placed over a map of the
area in question. Numbers can be selected from ran-
dom numbers tables. In monitoring studies where
the biologist wants to determine animal or habitat
changes before, during, and after a specific land use
(e.g., prescribed burn), controls should be estab-
lished. Without control, there is no way to deter-
mine whether specific land use has affected habitat
structure. For example, improvements in vegetation
structure on an allotment with a three-pasture, rest-
rotation system may result from higher-than-normal
precipitation over the past 5 years rather than from
the system. In this example, control samples taken
outside the grazing system but where similar physi-
cal conditions (vegetation, soils, to name two) exist
would allow the biologist to account for other possi-
ble causes of observed vegetation structure changes.

In cases where money and access restrict sam-
pling, biologists may have to cluster samples rather
than spread them over the entire habitat. This will
reduce the probability that habitat variability was
adequately sampled. For example, a biologist wants
to compare vegetation structure between pinyon-
juniper (Pinus-Juniperus sp.) and chaparral, but is
limited to two samples in each habitat type and one
habitat is five times larger than the other. Because
habitat size is not considered in allocating samples,
the biologist may not be able to determine which
habitat has greater vegetation structure, even if the
data suggest large differences. If samples of habitat
and animal variables are taken from the same site,

then the biologist can determine relationships be-
tween animals and their habitats.

The types and numbers of animals and habitat
characteristics sampled directly depend on objec-
tives established for an inventory or study. This will
determine the complexity and intensity of sampling.
Intensity can range from simple verification of one
animal in one habitat to large-scale vertebrate sam-
ples taken in many habitats over several years. (Ta-
ble 1 ) provides some examples of different sampling
intensities, given some specific objectives. In addi-
tion, Green (1979) provides an excellent discussion
of sampling design and data intensity.

Generally, biologists are confronted with three
types of data: animal, habitat, and animal/habitat.

ANIMAL DATA COLLECTION

Animal data are characterized by two basic
types: ( 1 ) individual species and ( 2 ) groups of spe-
cies. Each of these are detailed below.

Individual Species Measurements

There are many types of individual species data,
which are summarized in (Table 2). Although biolo-
gists will not likely collect all these types of data,
they will need some knowledge of them to interpret
results of contracts and other existing studies.

Presence. The simplest measurement of individual
species populations is presence/absence. In many
cases, the only objective is to verify use of a habitat
or site-specific area by a species (e.g., bald eagles
[Haliaeetus leucocephalns] along a riparian stretch).
These types of data help determine potential impacts
to a species and its habitat from development or
natural factors such as fire. These data also help
focus the biologist's efforts on site-specific areas,
especially those of threatened or endangered species.

One-time verification of a species in an area
may only be accurate for a short time, especially
when a species uses specific habitats only during cer-
tain seasons. For example, black hawks (Buteogallus
anthracinus) may nest in a riparian area during a
wet year when aquatic prey is abundant, but fail to
nest in the same area in the following year because
of drought-reduced prey abundance. Another con-
cern in using presence/absence data involves one-
time failure to identify a species in an area. Several
factors may account for species not being verified in
an area: unskilled observers, poor sampling tech-
niques, weather conditions during sampling, and ab-
sence of that species in the area during the sampling
period. The biologist should be very cautious in
using one-time, presence/absence data to assess an
area's fauna. Multiple-year verification of species
in habitats is preferred.

12

Data Types

Table 1. Examples of specific objectives carried through the design process.

Specific Objective/Question

Sampling Questions

Solution

Determine how a new water
development will affect mule
deer use of key browse in the
area.

What are the key browse in the
area?

How does water development
affect use?

How did mule deer use in the
area compare before installing
the new development versus
after?

Did other changes occur that
might account for differences
before and after development?

Since I did not know the size of the
area (several thousand acres of contin-
uous sagebrush-grass), I systematically
located my sample points along four
lines, out from the water development at
right angles (cross shaped). Samples
were located at 100 m (300 ft) intervals
along these lines (10 samples/line).
Angle of the first line was randomly se-
lected. At each point, I measured key
browse by a modified Cole technique.

Since cows graze the area year-round, I
could not determine whether browse
were being used by livestock or deer.
To assess relative food consumption, I
took deer and cow fecal samples at
each point. To detemine deer use,
I established 20 points along each line
(50 m [165 ft] apart, every other point at
a browse utilization station), where I
read 0.025 ha (0.01 a.) deer pellet
plots. I sampled both deer and browse
utilization twice before construction of
the development, 30 days after devel-
opment, and then once each session
for 3 years.

Determine if a riparian improve-
ment plan will increase the
number of nesting black hawks
and nesting success on Yankee
Creek.

What was the black hawk nest-
ing density like before imple-
menting the plan?

What was the average number
of young produced at nests
before implementing the plan?

What was the black hawk nest-
ing density 1, 3, 5, 10, and 15
years after implementation of
the plan?

How many new nests were built
1, 3, 5, 10, and 15 years after
implementation of the plan?

What was the impact of the plan
on primary black hawk prey?

Did factors other than those
provided by the plan affect the
results?

In order to determine the number of
black hawk nests before and after im-
plementation of the plan, I counted
all nests within cottonwood-willow
stands along the creek. To estimate
increases in the amount of tree canopy
in periods after implementation, I ob-
tained infrared, low-level photos of the
area every other year for 15 years

To determine the effects of the plan on
the fish prey base, I established perma-
nent sampling stations where I sampled
fish before and 1, 3, 5, 10, and 15
years after the plan was implemented.
This included sampling riffles, runs and
pools, and sites near known nests.
Samples were taken in the spring.

Since there were over 15 nests, nest
success at all of them could not be
monitored. I, therefore, randomly se-
lected three nests to monitor at 1 , 3, 5,
10, and 15 years after implementation
to determine nest success and habitat
uses. New nests were plotted against
changes in vegetation. Before and 1, 3,
5, 10, and 15 years after implementa-
tion, I determined the total number of
active nests in the creek. I also moni-
tored weather and flow conditions in the
creek for years previously listed via a
hygrothermograph and a U.S. Geologi-
cal Survey gauging station, respectively.

Data Types

13

Table 2. Summary of data types, some procedures and limitations.

Data Type

Method/Expression

Application/Limitation

1. Individual Species

A. Presence/absence

1. Searches (visual)

List of species in an area or

Used to verify key species in certain

habitat.

areas or habitats. Multi-year samples
needed for species that migrate or use
habitats seasonally. To compare two
or more areas, sampling intensity must
be equal. Random or systematic
sampling.

2. Capture and harvest

Same as above.

As above, along with sampling tech-
niques; must be similar to compare two
or more areas or habitats (e.g., similar
depth traps). Indirect sampling
methods.

3. Sign

Same as above.

Feces, nests, and other physical evi-
dence that a species occupies an area
or habitat. Indirect method. Sampling
limitations as above.

4. Auditory

Same as above.

Direct and indirect methods of verifying
species in a habitat or area via sound.
Sampling limitations as above.

5. Museum or other exist-

Animal and plant records by

Indirect method of verifying species in

ing records

county, township and range,

habitats or areas. Data highly variable.

sections, and habitat type with

Many records only at county level.

general notes. From museums,

Some animals/plants misidentified.

literature, and individual re-

cords. Data same as above.

B. Abundance (indexes

of population size)

1. Counts/searches

Number of animals/unit effort

Direct method of determining abun-

(usually time).

dance of animals in certain areas or
habitats. Some methods such as road
riding are inexpensive. Data good for
comparing different sites and impacts
of land use on wildlife and habitats if
controls are used. Good techniques for
monitoring if permanent sites are estab-
lished. Subject to observer biases,
and good only for readily observable
species.

2. Capture/trapping

Number of animals caught/unit

A series of individual methods for deter-

time.

mining species abundance. Methods
usually specific to certain species
groups (e.g., pit-fall traps and lizards).
Less observer bias than direct
searches. Some procedures excel in
verifying secretive species. Relatively
cost-effective. Excellent procedures for
monitoring, provided sampling is
systematic.

14

Data Types

Table 2. Summary of data types, some procedures and limitations (continued).

Data Type

Method/Expression

Application/Limitation

3. Line searches

4. Sign

5. Auditory

Similar to searches within an
area or habitat, but generally
expressed as number of ani-
mals seen/line distance (e.g.,
birds per mile).

Includes scent stations, nests,
number of tracks or pellet
groups. Data are expressed in
number of sign/unit area or
effort.

Generally the recording of dif-
ferent animal calls/unit time.

Direct methods of determining abun-
dance of animals in certain habitats or
an area as a function of distance and
time spent searching. When lines are
permanently located, these methods
are excellent for long-term monitoring.
Similar to searches, these have ob-
server-associated biases.

Commonly used as indexes of certain
big game abundance. Also used for
lizards, snakes, tortoises, and small
mammals. Most of these indexes pro-
duce highly variable data. Recommend
these procedures only when technigues
cannot be used.

Commonly used for small birds, raptors,
frogs, toads, and predators. Often diffi-
cult to count or estimate the number
of different individuals. Generally used
for verifying a species in an area.

C. Density/spatial
1. Counts/searches

2. Line searches

3. Mark and recapture

4. Depletion

5. Capture/trapping

Number of animals/unit area.

Animals/unit area as a function
of expected versus observed
(Chi-sguared test).

Population size/area as a func-
tion of marked versus unmarked
animals.

As above, marked animals are
considered removed.

Number of animals/unit area.

Common procedure used for key spe-
cies including big game and threatened
or endangered species. Labor-intensive
and effective only on easily verified
species. If combined with point proce-
dure (e.g., variable circle used for
birds), procedure is less labor-intensive.

Includes methods with horizontal belts
along a line (e.g., Emlen's bird tran-
sects). Used for birds and other spe-
cies easily observed. Generally
inexpensive to run, but procedures
reguire expert observers. Not effective
in small discontinuous habitats (e.g.,
riparian).

Several different procedures. Some
one-time sampling; others multiple sam-
ples. Commonly used to monitor terres-
trial and aguatic populations. Procedure
allows monitoring of individual animals.
Generally labor-intensive.

Similar to above. Some procedures
involve physical removal of animals. Re-
moval usually on small, easily captured
animals (e.g., fish). Generally labor-
intensive. Best results with closed
populations.

Live and kill traps. Used in intensive
studies where population size is
needed. Most effective on readily
trapped animals. Also used in conjunc-
tion with mark-and-recapture proce-
dures. Time- and labor-intensive.

Data Types

15

Table 2. Summary of data types, some procedures and limitations (continued).

Data Type

Method/Expression

Application/Limitation

6. Telemetry

Animal movement/unit time or
percentage ot time spent in an
area or habitat, performing
certain behaviors.

Effective method for determining behav-
ior and movement patterns. Initially
labor-intensive. Generally limited to
priority species.

D. Population structure

1 . Sex ratios

2. Age ratios

Ratio males to females.

Ratio of animals in certain age
groups.

Some animals have a sex ratio range
that represents a fit population. Num-
bers outside these ranges can indicate
a stressed population. Generally used
for big game. Less accurate for lower
vertebrates because their sex ratios
vary with different, natural environmen-
tal conditions.

Similar to sex ratios. Other-than-normal
ratios of animals in certain age groups
can indicate a stressed population.
Generally used for animals whose age
structure does not vary with environ-
mental fluctuations. Be cautious in us-
ing these in lower vertebrates for
reasons listed for sex ratios.

E. Productivity

(recruitment rates)

Generally, reproductive effort
(number of recruits).

Commonly used for big game and other
species whose birth or recruitment
rates can be measured. Generally, la-
bor-intensive. Many populations cyclic.

F. Population condition

Generally, percentage of ani-
mals in certain condition
classes (e.g., rankings of fat
storage).

Animals conditions used as indicators
of population fitness and stress. Limited
to priority species. Labor-intensive.

II. Wildlife Communities

A. Species richness

1. Presence/absence
(habitat type)

2. Presence/absence
(habitat components)

List of species in a habitat type
or total number species/habitat
type.

Similar to above but expressed
by habitat component (e.g.,
number of species using rock
for feeding substrate).

Lists complied from on-site data collec-
tion and existing information. Can be
used to compare habitat species rich-
ness. Commonly used for land-use
planning, with limited utility on site-spe-
cific assessments.

Used to identify critical components of
habitat based on species' uses. Also
used to assess impact of habitat com-
ponent loss on groups of wildlife. Guild-
ing is a process for grouping animals
(see text). Use for land-use planning,
but cautiously for site-specific analysis.

B. Community abundance

Overall abundance of a taxo-
nomic group (e.g., small mam-
mals) in habitats. Data
expression varies with sampling
methods (see Individual Spe-
cies section).

Used to assess value of habitats (e.g.,
pinyon-juniper versus chaparral) for
certain groups of wildlife. More time-
and labor-intensive than presence/
absence. Problems similar to those for
sampling individual species abundance
(see Individual Species section).

16

Data Types

Table 2. Summary of data types, some procedures and limitations (continued).

Data Type

C. Community density

Method/Expression

As above but expressed as
animals/unit area.

Application/Limitation

As above. Subject to problems associ-
ated with individual species densities.

D. Biomass

Weight or mass (gms) a spe-
cies contributes to the overall
community. Often expressed in
ratios.

Used to estimate species contributions
to a community based on mass. Also
estimates trophic levels and food avail-
able to certain predators. Limited utility
in most management-related studies.

E. Species similarities
(similarity coefficients)

Index generally between 0 (no
similarity) and 1 (identical).
Single value.

Used to compare species assemblages
on two different sites. Calculations are
simple and do not require abundance
(presence/absence only). Some formu-
las emphasize similarity; others, differ-
ences. Excellent for comparing wildlife
communities.

F. Species diversity and
habitat use diversity

Index generally ranging from 0
(no diversity) to 10 (high diver-
sity). Certain formulas empha-
size evenness of abundance
shared by species on a site
(log10); others, richness of spe-
cies (loge).

Commonly used to compare species
diversity between sites or the diversity
of habitats used by species (habitat use
diversity or niche breadth). Abundance
or density data needed for calculations.
Subject to biases associated with abun-
dance and density. When sampling is
conducted over long periods, data
provide insight into changes in species
dominance relative to a management
prescription.

G. Species overlap

Like similarity coefficients, but
value range of 0 to 1 is based
on shared species abundance
versus total abundance. Horn-
Overlap is the most commonly
used procedure.

Used to assess differences in animal
community structure between habitats.
Abundance or density data required.
Also assesses species contributions on
one site versus those on other sites.
Biases similar to those of abundance
and density sampling.

Habitat (characteristics, com-
position, and structural condi-
tions)

A. Point

Percentage of plant cover, sub-
strate type, plant frequency,
plant and other horizontal and
vertical attributes of habitat.

Used for gross characterizations of
habitats and areas. When points are
permanently located, changes in habitat
can be determined. Often combined
with a line to determine percentages.
Generally poor for grasses and forbs.

B. Line

As above, but as a function of
distance along a line.

Extremely useful for characterizing plant
composition and frequency and habitat
components for a habitat or area. Gen-
erally quick and easy to read. Excellent
for monitoring changes if permanently
marked. Most effective on habitats
dominated by shrubs.

C. Plot

Number of a habitat compo-
nent/unit area. Also, weight
contributed by a plant species/
unit area.

Commonly used for grass and forb
density, composition, and biomass on
an area or in a habitat. Labor-intensive
Not generally effective for woodland
and forest habitats. Good for monitoring
rangeland if permanently located.

Data Types

17

Table 2. Summary of data types, some procedures and limitations (concluded).

Data Type

Method/Expression

Application/Limitation

D.

Plotless

Number of plants or a certain

Fast and effective way of determining

habitat component/unit area as

absolute densities of trees, shrubs, and

a function of distance from a

other habitat components such as

center point.

snags. Structural characteristics of hab-
itat components are also usually mea-
sured. Point-center-quarter method is a
commonly used plotless method. Excel-
lent for monitoring if points are perma-
nently located.

IV. Animal '/Habitat

A.

Animal/habitat indexes

Usually expressed as a quality

Indexes are suited for broad-scale land-

index value ranging from 0

use planning, especially when using a

(poor quality habitat) to 1 (best

generic model. Can be used for more

possible habitat). Data also

site-specific application if area-specific

expressed as quality (suitability)

data are used to develop the model.

index x the area = habitat

Habitat Suitability Index models use a

value.

series of ecological relationships to
develop models for individual species.
Habitat Relation models rank serai
stages of habitats for individual species
relative to performing important ecologi-
cal functions.

B.

Animal/habitat correlations

Data expression varies with

Procedures that use empirical data to

procedures (see below).

determine relationships between ani-
mals and habitats.

1. Correlation and regres-

Data expressed as a linear or

If animal and habitat data are collected

sion

nonlinear (e.g., polynomial)

on similar sites, these equations are

relationship, usually between

ideal for area-specific monitoring and

animal population and one

development of predictive equations for

or several habitat variables.

management assessment. SPSS and
BMDP software provide these types of
analysis. With practice, these are easily
developed.

2. Principle components

Data generally expressed as

Valuable tool for illustrating differences

location along axes (e.g.,

in animal uses of habitat. Allows the

X,Y,Z). Locations on axes repre-

biologists to compress several habitat

sent combined habitat

variables into two or three axes which

variables.

then are easily illustrated.

18

Data Types

Abundance and Density. Abundance and density
measurements have been widely used in wildlife
inventories and monitoring studies. Generally, there
are three types of abundance measurements: total
number of animals in a population; number of
animals per unit area (density); and abundance and
density of one population relative to another
(Caughley 1977). Rarely will biologists measure the
absolute number of animals in a population or area
and need actual numbers or density. Therefore, most
wildlife inventories and monitoring studies use
estimates of population size and density. Estimates
are usually accomplished by indexes that are
correlated with population size and density
(Caughley 1977; Eberhardt 1978).

Estimates of population size or absolute abun-
dance are commonly used in wildlife inventories.
Because these estimates are indexes of populations
and not ascribed to units of areas, they are usually
termed abundance. Abundances include animal num-
bers obtained from direct observation, sign (tracks,
etc.), and captures. Certain biases are associated with
each of these methods. Caughley (1977) and Eber-
hardt ( 1978) provide excellent summaries of these
biases; biologists should consult these and other
sources before deciding a method. Estimates of abun-
dance include, but are not limited to —

• number of animals seen per hour of observation
(also termed time-restraint);

• number of animals seen per linear distance (e.g.,
raptors seen per mile of powerline );

• number of animals trapped per 24 hours;

• number of tracks counted per hour of observa-
tion;

• number of calls heard per hour (e.g., frogs and
owls).

All these can be measured without regard to area.
Therefore, they are estimates that presumably are
correlated with population size.

Relative abundance generally refers to the con-
tribution a species makes to the total abundance
of that wildlife community. For example, if three ob-
servations of species A and one observation of spe-
cies B were made, then species A makes up 75%
of the abundance and species B, 25%. These relative
abundances are expressed as 0.75 and 0.25, respec-
tively (see Hendrickson et al. 1980 for excellent fish
examples).

Density refers to animal numbers per unit area
(Caughley 1977). Estimates of density are obtained
in ways similar to those acquired for abundance
estimates, but they are applied to a naturally or artifi-
cially ascribed area (e.g., habitat boundary or marked

hectare or acre plot, respectively). For example, a
biologist estimates lizard density by searching a pre-
marked area and counting all lizards seen (Bury
1982). Similarly, bird density can be estimated by
using a variable plot method where the number of
birds are counted within 1/100 ha (0.025 a.) circular
plots (see Ralph and Scott 1981, for examples).

Estimates of density are also obtained by count-
ing the number of animals seen within prescribed
belts established along a fixed line. By using proba-
bility statistics, animal densities can be predicted
in prescribed belts outside areas adjacent to the
transect line. Perhaps the most commonly used line/
area transect is that developed by Emlen ( 1971 )
for birds. Emlen's bird transect has been modified to
reduce sampling biases, including statistical applica-
tions available on computer software. Ralph and
Scott ( 1981 ) provide a series of papers about bird
sampling.

When comparing estimates of population abun-
dance and density, the biologist should consider
the number of samples to be allocated to each area.
This is especially important when comparing densi-
ties. To compare two sites, biologists should com-
pensate for area-related variability by allocating
sample size relative to proportions of area. For exam-
ple, if area A is 1,000 ha ( 2,500 a. ); area B is 5,000
ha (2,500 a.); and area C is 10,000 ha (250,000 a.),
then area A would have 1/10 the number of samples
of area C, and area B, 1/2 the number of area C.

Depletion sampling and mark-and-recapture are
other methods commonly used to estimate popula-
tion size and density. Generally, removal methods
involve removing animals during two or more sam-
ple periods, where only two factors are used: popula-
tion size N and capture probability. Population data
are obtained by comparing expected captures to
actual captures, generally by a goodness-of-fit test
(Chi-squared test). One problem with this method is
that it assumes all animals have an equal probability
of capture. In many cases, this is not true. For exam-
ple, the probability of capturing large fish is greater
than small fish when electroshocking is used. Otis et
al. ( 1978) and White et al. (1982) have developed
more generalized removal models that can account
for unequal capture probability. Removal procedures
are common in estimating the number of fish in a
stream, primarily because the biologist is sampling a
relatively closed system. White et al. (1982) provide
a good review of a number of removal methods.

Mark-and-recapture techniques involve marking
animals and then sampling at a later date to deter-
mine the ratio of marked to unmarked animals. Pop-
ulation size is determined from the ratio of marked
to unmarked animals in one or more successive sam-
ples. The Petersen-Lincoln estimator is perhaps the
most commonly used mark-and-recapture method.

Data Types

19

Mark-and-recapture methods require several assump-
tions: ( 1 ) no loss or gain of marked animals, ( 2 ) no
recruitment, (3) equal catchability in marked and
unmarked animals, and (4) equal mortality rates in
marked versus unmarked animals (Tipton 1980).
These criteria are more easily met in samples taken
within short periods (e.g., consecutive days) than
over longer periods (e.g., several months). Data ob-
tained by these methods often vary (Tipton 1980),
primarily because multiple samples have been taken
under different climatic conditions, affecting animal
activity and behavior. Otis et al. (1978) and White et
al. ( 1982) provide detailed discussions of these
biases.

Spatial Arrangement and Movement. In addition
to estimating abundance and density, determining
spatial arrangement and movement of individual
species in habitat types may be important. For
species that require different habitats for different
functions, such as breeding or feeding, biologists
should collect data on species movements and
different uses of habitats. Individual movements can
be obtained from direct observations of marked or
unmarked animals or from radiotelemetry. Chapter
33, Radiotelemetry, provides guidance on such
efforts. These types of data help biologists determine
habitat needs of individual species, which aid in
various types of environmental assessment.
Limitations of mark-and-recapture (resighting)
animals (see previous section) and of radiotelemetry
are generally related to cost, the former being labor-
intensive and the latter requiring expensive
equipment.

Population Structure. Sex and age ratios are used
to determine conditions of natural populations,
especially big game. Sex ratios of healthy populations
are compared with those of populations being
studied, thus giving the biologist an indicator of
population health. Although sex ratios may indicate
population problems, biologists should take other
population data to support conclusions derived from
sex ratios.

Generally, age ratios also reflect population con-
dition. Lack of certain age classes such as juveniles
and breeding adults can indicate a stressed popula-
tion. For example, biologists are concerned about
large numbers of old desert tortoises (Gopherus
agasstzii) and lack of young in Arizona populations.
As with sex ratios, corroborative data such as abun-
dance, density, and productivity should be collected
along with age ratios.

Biologists should be cautious when using and
interpreting population structure data. Many species'
age and sex ratios fluctuate naturally with changes
in environmental conditions. Therefore, it is impor-
tant to obtain multiple-year samples with controls to
determine human changes in population structure.

Productivity. Like population structure, productiv-
ity can indicate population fitness. Productivity is
usually considered the number of young (as defined)
recruited into the population per adults, adult fe-
males or adult pairs. Examples are the number of
young fledged per nest and fawns per doe. Although
often a better indicator of species' fitness than popu-
lation structure, this measurement may be hard to
obtain with some species.

Condition. Condition of individual animals can
indicate fitness of a population, especially population
stress. For example, if 90% of the animals in a
bighorn sheep (Ovis canadensis) population are in
poor condition (e.g., emaciated, having thin coats or
high incidences of sinusitus ), then a biologist might
predict that this population is in trouble. Although
often good indicators of population condition, these
measurements are generally hard to obtain,
expensive, and labor-intensive.

General Considerations. Biologists should be
careful in selecting, making, and interpreting
population measurements, particularly estimates of
abundance, density, population structure, or
productivity. Population numbers, sex and age ratios,
and productivity all vary naturally. If a biologist
needs to collect such data on a species whose
abundance fluctuates naturally (e.g., rabbit
populations), then samples should be taken over
several years.

Knowledge of natural population fluctuations is
extremely important in interpreting data from cause -
and-effect studies. For example, observed declines
in a population on a mined area may result from low
precipitation during sampling rather than from mine-
caused habitat alteration.

Community Measurements

Community measurements are data collected on
species groups. The species groups may be taxo-
nomic (e.g., songbirds) groups that use habitat in
similar ways (guilds) or other logical groupings. For
example, a biologist may want to compare species
richness (total number of species) between two
habitat types to help decide where to place a
powerline.

Community data may also involve only a specific
taxonomic group. For example, because of limited
funding, a biologist may only assess species diversity
of bird communities on habitat types. Data collection
limited to a taxonomic group or community may
also be caused by objectives established during pre-
planning. For example, data may be collected only
on raptors because of priorities established during
preliminary planning. Community measurements are
summarized in (Table 2).

20

Data Types

The simplest and most widely used community
measurement is the presence/absence of the wildlife
groups. Presence/absence data allow biologists to
determine wildlife community composition in differ-
ent habitat types. These data also allow biologists
to compare faunal richness of habitat types. These
comparisons may be useful in land-use planning. For
example, an inventory might show that pinyon-juni-
per habitats are more faunally rich than adjacent
desert grassland habitats and possess a greater num-
ber of threatened or endangered species. These data
are then used to support a management alternative
of placing a mining road through desert grassland
habitat rather than through pinyon-juniper.

While primarily used at a habitat level (e.g.,
pinyon-juniper), presence/absence data can also be
used to determine species groups that depend on
certain habitat components or spatial arrangement
(e.g., vertical vegetation arrangement). For example,
bird species can be verified in predetermined incre-
ments of vertical vegetation layers, thus indicating
bird species arrangements within a habitat type. This
type of data can also be useful in land-use planning.
A biologist may find that shrub and old growth com-
ponents of pinyon-juniper contain the greatest vari-
ety of birds. These data are then used to support a
management alternative that would limit wood cut-
ting to middle-aged trees.

Presence/absence data collection is far less ex-
pensive than abundance or density data collection,
especially when censusing an entire wildlife commu-
nity. Collecting presence data for wildlife groups is
similar to that for single species; however, the analy-
sis, interpretation, and use of such data are often
quite different.

Several types of statistical procedures have been
used to analyze presence/absence data of wildlife
communities. Similarity coefficients are the simplest
of these analyses. A summary of these coefficients
is listed in (Table 2). Generally, similarity coeffi-
cients compare the number of species common to
two habitats with the total number of species. Gen-
erally, these values range between 0 (no similarity)
and 1 (identical communities). Some similarity coef-
ficients such as Simpson's emphasize similarities be-
tween two sites, and others such as Jaccards
emphasize differences between sites. See Pimentel
( 1979) for a full discussion of these methods.

Cluster analysis is another method to group
species that use similar habitat or habitat compo-
nents. Cluster analysis also provides a visual perspec-
tive of degrees of similarity between groups and
variability within a cluster of species. Clusters are
depicted in two-dimensional graphs along X and Y
axes or in histograms based on separation between
groups. Separation or distance between clustered
species groups is determined several ways depending
on the cluster statistic used. For example, one clus-
tering statistic determines group distances by calcu-
lating a central location in the X and Y axes of each
group. Pimentel ( 1979) provides a detailed explana-
tion of this application of cluster analysis, and Short
and Burnham ( 1982) provide an example of this
analysis, demonstrating how bird species cluster in
terms of habitat uses.

Another approach to determining groupings of
species is to determine guilds, which are groups
of species that use similar habitat components. Short
(1983, 1984) also describes a simple procedure for
determining species guilds based on where they
breed and feed. For example, one guild breeds in
vegetation canopy (above 4.5 m [15 ft]) and feeds
on the surface, whereas another guild could feed in
the same location but also in the canopy. These
groups are different in their use of habitat for feed-
ing and breeding. Other important ecological func-
tions can be used to group species into guilds,
although these are not described by Short. For exam-
ple, thermoregulation is important in reptile commu-
nities; therefore, for this community, it is important
to describe where and how they thermoregulate
in addition to where they breed and feed.

By grouping species that use habitats in similar
ways (guilds), the biologist can predict the species
groups that will be affected when a habitat compo-
nent is altered or lost. The detail or breakdown of
habitat components with which species are associ-
ated can vary. For example, whereas one biologist
identifies 3 vertical layers of vegetation, another may
identify 10. The biologist who identifies three layers
may be more concerned with horizontal arrange-
ment of vegetation for determining habitat require-
ments of small mammals and, therefore, will only use
three vertical layers, whereas another biologist,
studying bird communities, may need greater deline-
ation of vertical space.

In interpreting presence/absence data, biologists
should consider the intensity of the sample. In short-
term inventories, a biologist would not likely ob-
serve all species in a given area, especially rare spe-
cies and those with restricted activity. The methods
used to verify species in an area may also be incom-
plete, and this will affect the total number of species
observed.

Simply stated, guilding is a method of sorting
wildlife into groupings based on the use of habitat
components for important ecological functions. The
primary function of placing species into guilds is
to provide a way of looking at how species groups
might be affected by changes in habitat structure,
rather than determining requirements of each spe-
cies and then projecting how each might be affected

Data Types

21

by a proposed action. For example, a group of spe-
cies dependent on downed logs for breeding, feed-
ing, and thermal cover, would be negatively affected
by removal of downed and dead material from the
area.

Guild data can be effectively used during early
land management planning stages, especially to de-
termine potential conflicts with land-use proposals.
Guilding also helps focus attention on heavily used
components of habitat. Biologists, however, should
be cautious when using guild data. For example,
guild data do not account for the ability of a species
to shift to secondary habitat structures when pre-
ferred habitat is lost. In addition, guilds are artifi-
cially defined groups described by biologists. Mannin
et al. (1984) found that 67% of birds within guilds
in Oregon forest habitats responded differently to
habitat changes than predicted. Some biologists ar-
gued that each species occupies a separate guild and,
therefore, reacts differently to habitat changes (see
discussions of Mannin et al. 1984). Biologists should
not attempt to use guilding data to determine pre-
cise, site-specific changes in species abundance and
composition. Again, these data are best used to focus
attention on important habitat components, rather
than as precise predictive models.

Another related procedure is use of indicator
species. One use of this procedure involves selecting
a species termed an indicator species, which repre-
sents a group of species or guild. This species is then
sampled to determine how it is affected by land-use
practices. Data obtained for this species are then
applied to the entire group of species in the guild.
For example, alligator lizards (Gerrhonotus sp.) rep-
resent a guild of species that require downed leaf
litter within habitats they occupy. Therefore, in the-
ory, if we know how alligator lizard populations
will change with changes in leaf litter composition,
we know how all other members of that guild will
respond. Like guilding, this procedure assumes that
all species in a guild use habitat in an identical man-
ner and all have the same ability to use secondary
habitat. As mentioned earlier, Mannin et al. (1984)
found that species within guilds did not react simi-
larly to habitat changes. From a management per-
spective, the indicator species process saves time
and money since it allows the biologist to sample
only a few species and yet determine how the entire
wildlife community responds to habitat changes.
Biologically, however, this concept does not consis-
tently hold up. While I recommend using guilding
procedures to focus on important habitat compo-
nents, biologists should be cautious in using indica-
tor species to represent species guilds, especially
when accurate, site-specific predictions are needed.

The U.S. Forest Service, Region 6 (Albuquerque,
NM), uses indicator species in a different way. In-
stead of representing a group or guild of species, an

indicator species is used to indicate quality of a cer-
tain habitat component on a site. A group of indica-
tor species is selected so that all important habitat
components on a site are covered. Then, when a
certain site is altered by a land-use practice, the biol-
ogist can look at changes in indicator species to
determine effects on important habitat components.
Perhaps an easier way to determine the same effects
would be to measure changes in important habitat
components. The biologist identifies these important
components from existing studies and literature.

When it is possible to collect data on abundance
or density of a wildlife or taxonomic community
(e.g., raptors), species diversity and overlap indexes
can be calculated. Of the several types of diversity
indexes, the most commonly used is the Shannon-
Weaver index (Hair 1980). This index provides a
single value for a wildlife community based both on
the total number of species and the evenness of
abundance among the species. All the indexes allow
biologists to assess relative values of habitats based
on species diversity. Since the indexes are based
partially on relative abundance or density, they can
only be used within groups having abundance meas-
ured in similar ways. For example, a single diversity
index cannot be computed for big game and lizards,
because relative abundance or density is determined
in different ways.

Whereas similarity coefficients compare species
similarities between habitats, overlap indexes com-
pare proportions of species and their abundance
between habitats. Perhaps the most commonly used
overlap index is that described by Horn ( 1966).
Routledge (1980) and Ricklefs and Lau (1980) dis-
cuss limitations and bias associated with these types
of indexes. Biologists should consult these and other
sources before using diversity or overlap indexes.

Another type of data used to assess and compare
species composition and influence is biomass. In
many wildlife studies, where the biologist is trying to
determine how species contribute to trophic levels,
biomass is measured. For example, it may be impor-
tant to know how the loss or reduction of a certain
species will affect food available to other species
in the community. In trophic level and food niche
studies, biomass information is generally preferred
over abundance data; two species may provide simi-
lar energy to a system even though one is small and
abundant and the other is common and rare.

HABITAT DATA

In most cases, biologists collect habitat data to
make some type of prediction about animal popula-
tions. Many of the same kinds of measurements dis-
cussed for animals and animal communities can also
be applied to plant communities (habitat types) and

22

Data Types

individual plant species. In addition, habitat measure-
ments include abiotic physical features such as rock
and soils. I discuss measurements and data collection
in two general areas: ( 1 ) habitat area or type and
(2) habitat component.

Habitat Area or Type Data

These types of data apply to a habitat type or
area as a whole. Generally, a habitat type is an area,
delineated by a biologist, that has certain consistent
abiotic and biotic attributes such as dominant and
subdominant vegetation. The U.S. Bureau of Land
Management (BLM) terms these "habitat sites" (Kerr
and Brown 1977). Groups of habitat areas with simi-
lar attributes that support similar groupings of spe-
cies are termed "habitat types" or in BLM, "standard
habitat sites."

Similar to animal community measurements, the
simplest measurement of habitat type is presence/
absence. In most cases, this involves identifying and
mapping habitat sites within an area.

The following spatial measurements of habitat
types are commonly used to assess habitat availabil-
ity:

• size of habitat sites (area);

• location and position of habitat sites, including
distance between habitat sites and heterogeneity
within an area;

• edge influences created by habitat site interfaces
and ecotones;

• temporal availability of habitat sites (e.g., tempo-
rary lentic habitat sites).

All these data help biologists assess availability of
certain habitats and the potential for an area to sup-
port various types of wildlife.

Indexes have been used to quantify potential
influences of edges on wildlife habitat. Patton ( 1975)
provides a simple index of edge.

Habitat Component Data

Habitat component data refer to information
taken on abiotic and biotic attributes of habitat. Gen-
erally, data collection is limited to those components
known to affect wildlife distribution and population
fitness. Data can include spatial arrangement of these
components at a single point in time or over several
years, or physical characteristics of components such
as average height and width. Habitat component
data can represent an entire habitat type or a specific
area occupied by species. Table 2 provides a sum-
mary of these data. Examples of abiotic data
include —

• average surface rock cover,

• average size of surface rocks,

• tallus and rock outcrop abundance and size,

• soil depth,

• vertical and horizontal heterogeneity of soils
and rock.

Measurement and classification of terrestrial and
aquatic physical features are covered in Chapters
27 and 28 in this book. Examples of biotic data
include —

• average litter depth;

• average litter cover per unit area;

• snag density and size;

• average height of different vegetation life-forms
(grasses, shrubs, and trees);

• vertical and horizontal heterogeneity of litter,
snags, and live vegetation.

Other data often collected include attributes such as
elevation, latitude and longitude, slope, and aspect.

Many sampling methods can be used to obtain
presence/absence, abundance, density, size, spatial
arrangement, and habitat and habitat component
availability data. These methods fall into four general
categories:

Plant species richness, diversity, similarity, and
overlap can be computed for habitat sites or types in
ways similar to animal species composition in animal
communities. Plant species richness is the total num-
ber of species present within a habitat type. Similar-
ity coefficients can be computed by comparing
shared plant species richness between habitats. Plant
species diversity and overlap between habitat types
are computed from diversity indexes identical to
those previously discussed for animals. Often, these
data help assess forage quantity and quality between
habitat sites or types.

(1) Point (e.g., point intercept);

(2) Line (e.g., line intercept);
( 3 ) Area or plot;

(4) Plotless (e.g., point-center-quarter).

Generally, point data can be obtained quickly,
depending on the number of points needed. The
number of samples taken will depend on variability

Data Types

23

of the habitat component to be measured. For exam-
ple, a biologist can use several point samples to de-
termine leaf litter depth. These points can be
systematically (e.g., on a line) or randomly located.
One of the most widely used of these methods is the
toe-pace transect (Hays et al. 1981). Point methods
can also be used to determine frequency and relative
abundance of certain components. For example, a
toe-pace transect can be used to determine percent-
ages of certain substrate on habitat surfaces. This
type of procedure does not identify the density or
arrangement of habitat components on the surface
but rather frequency and relative abundance of these
components (expressed as a percentage). These
types of data are best used as a rough indicator of
component composition and frequency on a site.

Line transect methods also generate data on
habitat component composition, frequency, and
abundance. In addition, they provide data on hori-
zontal arrangement of habitat components (a func-
tion of distance between components). Similar to
point transects, these data are expressed as percent-
ages and can be used to characterize habitat compo-
nents on a site. The most common line transect
method is the line-intercept (see Hays et al. 1981 ).

Plot sampling techniques are frequently used to
describe habitat component arrangement, density,
and composition within an area. The area in which
habitat components are measured is usually prede-
fined, and area size can vary greatly. Methods used to
generate these data are usually time- and labor-inten-
sive, providing relatively accurate data but only for
a small area. They are most commonly used when
grass and forb composition and density data are
needed. Several plot or area methods have been de-
veloped. See Hays et al. (1981 ) for greater detail.
Data generated from plot methods are usually ex-
pressed as the number of plants per unit area. In
many forage studies, data are expressed in weight by
plant species per unit area.

Plotless sampling techniques produce density
estimates for habitat components, but as a function
of distance rather than density in a predetermined
area. One of the more commonly used plotless tech-
niques is the point-center-quarter method (Phillips
1959), which allows many samples to be collected
in less time than a plot method.

When set up as permanent samples and read
over several years, line, plot, and plotless methods
yield valuable information on changes in habitat
components on a site. Nonetheless, these habitat
measurement techniques have many biases and elim-
inations. Chapters 26 through 32 discuss habitat
component sampling methods in greater detail.

In all these methods, characteristics of individ-
ual habitat components can be measured. For exam-

ple, tree density determined from a point-center-
quarter method (plotless) is based on distances of
trees from the point. Each tree sampled for distance
can also be measured for physical characteristics
such as height and diameter. In some cases, habitat
components cannot be measured randomly or sys-
tematically. For example, a biologist might need
to determine physical characteristics of snags used
by nesting spotted owls (Strix occidentalis) so that
adequate habitat is provided in the area's snag man-
agement plan. In this case, the biologist would meas-
ure only those snags with known spotted owl nests.

Heterogeneity of both abiotic and biotic habitat
components can be computed from diversity in-
dexes. Foliage height diversity and patchiness in-
dexes are two of the more commonly used (see
Ralph and Scott 1981; Hays et al. 1981; Chapter 31
in this book ). Foliage height diversity' data give the
biologist an indication of how evenly vegetation
is distributed in vertical space. Patchiness diversity
indexes provide data on how evenly vegetation is
spread on horizontal space. Vertical spatial data are
extremely helpful in determining a site's ability to
support certain birds, while horizontal spatial data
help determine which small mammals and lizards
occupy a site.

In addition to other types of habitat data dis-
cussed, population structure (e.g., age ratios), pro-
ductivity, and condition (e.g., degree of hedging)
measurements can be taken on certain plant species.
These types of measurements are similar to those
taken on individual animal species, but are generally
easier to obtain because plants are stationary. These
measurements help biologists determine the condi-
tion and fitness of certain types of plants such as key
riparian tree species or important big game browse.

ANIMAL/HABITAT CORRELATIONS

Perhaps the most important consideration in
designing and conducting a habitat inventory or
monitoring project is ensuring that animal and habi-
tat data can be compared. By doing so, biologists can
then predict how habitat alterations will affect ani-
mal populations.

Two factors are extremely important if animal
and habitat data are to be compared. First, animal
and habitat data should be collected on the same
site. As samples accumulate, the biologist can deter-
mine how animal populations fluctuate with differ-
ences in physical and spatial arrangement of habitats.
If a biologist samples habitats and animals at different
sites, habitat and animal data cannot be compared.

Second, the level of sampling needed to com-
pare animal and habitat data should be carefully de-
termined. In most cases, animal and habitat sampling

24

Data Types

levels do not need to be similar. For example, most
inventories of wildlife eommunities in habitat types
require only presence and absence data for animals.
However, habitat measurements are generally more
specific than simple presence or absence of habitat
components; these tell little about habitat quality
and condition. More often, habitat measurements in-
clude spatial arrangement and abundance of habitat
components.

Conversely, sometimes biologists may need
more detailed data on animals than habitat. For ex-
ample, a biologist who is monitoring the effect of
new water developments on a pronghorn population
in an area would require detailed information on
movement, abundance or density, spatial arrange-
ment, and reproduction of pronghorns. Habitat data
collection would probably be limited to habitat map-
ping and presence or absence of water. In both
cases, the level or complexity of animal versus habi-
tat sampling depend on specific inventory and moni-
toring objectives. However, habitat data are usually
less costly and time-consuming to collect and more
accurate than animal data; plants are easier to locate
and are somewhat stationary. Therefore, most inven-
tories and studies generally have more detailed habi-
tat data. (Table 2) provides a summary of these data.

Many statistical procedures compare animal and
habitat data. Perhaps the simplest of these proce-
dures is to compile a list of species observed or col-
lected in a habitat component or habitat type. This
allows biologists to weigh relative wildlife values
of certain habitat components and habitat types,
which can then be used in evaluating proposed land
management.

Multivariate analyses, such as correlation, regres-
sion, and principal component analyses, can be used
to compare animal and habitat data. For example, a
biologist may want to determine which physical
attributes of habitat affect species richness. Jones et
al. (1985) used a step-wise, multiple regression to
determine how reptile species richness on habitat is-
lands was affected by habitat size and habitat compo-
nent structure. In this study, animal data consisted
of only presence and absence of species on sites,
while habitat data were more detailed and included
size, elevation, and detailed horizontal and vertical
habitat structures. Pimentel ( 1979) and Green
(1979) discuss these statistical procedures and pro-
vide examples. In addition, Jones and Glinski ( 1983),
discussing lizards; Rotenberry and Wiens ( 1980),
discussing birds; and Matthews and Hill (1981 ), dis-
cussing fish, provide examples of multivariate anal-
yses of animal/habitat data.

The U.S. Fish and Wildlife Service and U.S. For-
est Service have developed models that assess habitat
capability or suitability for certain species. The for-
mer has developed the Habitat Evaluation Proce-

dures (HEP) system, consisting of a series of habitat
suitability index ( HSI ) models developed for individ-
ual animals (Schamberger et al. 1982).

The HEP process and HSI models have some
drawbacks. First, habitat ratings are indexes, and are
normally derived from expert opinion and literature
review rather than from empirical data. Testing these
models is difficult, since the index cannot be verified
in the field.

Another type of habitat evaluation model was
developed by Williams et al. (1977). These are
known as Pattern Recognition ( PATREC ) models,
which, unlike HEP models, are based on statistical
inference. Data used in these models are often em-
pirical (e.g., population data). Models result in prob-
ability statements on an area's suitability to support
a certain species (e.g., an area in question has a
75% probability of supporting dinosaurs ). Because
PATREC models are developed from empirical data
(generally a population factor such as presence or
density), they are easier than HSI models to test.

The U.S. Forest Service has developed a series of
models (habitat capability relation models) that cor-
relate species with habitat components. The U.S.
Forest Service has also developed simulation models
that simulate changes in vegetation over time as a
result of plant succession and land-use practices on
an area. By combining these two types of models,
biologists can predict how wildlife habitat quality
and, thus, wildlife populations will change over time
given a certain land use.

This approach is limited in that habitat quality is
evaluated only by serai stages of cover types (habitat
types). The procedure does not account for relation-
ships between site-specific habitat components and
wildlife, and it is based on models developed for
individual species. Currently, only a few species have
been modeled. Therefore, I recommend caution
when using this system, especially when assessing
management alternatives on a site-specific basis. As
with guilding, I believe such models are most valua-
ble during major land-use planning, especially to
focus attention on major issues.

The area of animal/habitat correlations is cov-
ered in more detail in Chapter 38, Habitat Evaluation
Systems.

DISCUSSION

Generally, there are two approaches to data
collection and their use in recommending wildlife
and habitat alternatives. These two approaches have
been described by Bailey (1982) and are termed
"Linear-Comprehensive Management" and "Cyclic-
Incrementalism."

Data Types

25

Linear-comprehensive management is the accu-
mulation of data on all habitat components and wild-
life populations in all ecosystems (Figure 1). This
includes detailed, intensive data on population sizes
and productivities, seasonal distributions of animals,
forage productivity, cover availability, precipitation,
and all other ecological factors affecting biological
systems. This management philosophy also assumes
clear definitions of publicly agreed-upon manage-
ment goals; the biologist will have specific data that
support public opinions on issues. It also assumes
that all population and habitat treatment methods
are clearly understood and that the biologist has
a comprehensive knowledge of how each treatment
will affect the ecosystem and animals in question.

In general, linear-comprehensive management
generates large amounts of site-specific data without
regard to how the resulting data will be used in
management decisions. In some cases, models are
developed and used in planning without local test-
ing. Linear-comprehensive management does not
work in most wildlife management situations for sev-
eral reasons. First, wildlife is diverse in most ecosys-

INVENTORY

1

CENSUS

1

YIELD
DETERMINATION

1

DIAGNOSIS

*

CONTROL

HABITAT
MGMT

POPULATION
MGMT

terns. Simply stated, the biologist cannot collect data
on all population and habitat factors, nor can a biolo-
gist determine how each species will be affected by
every treatment. Some of the most comprehensive
studies have only censused a tiny fraction of the
animal population in an ecosystem. In addition, the
relationships between animals and ecosystems are
very complex. Most animals have several habitat
requirements, and these requirements are often in-
terrelated in complex ways. Therefore, predicting
how any particular animal will react to habitat
changes is difficult. As mentioned earlier, Mannan et
al. ( 1984) found that bird species within well-de-
fined guilds responded differently to certain manage-
ment prescriptions. Finally, wildlife budgets prohibit
collection of large amounts of wildlife data. Because
knowledge and budgets are limited, most wildlife
management programs are extensive, rather than
intensive.

Cyclic-incrementalism is defined as a slow, sub-
jective process of collecting data and knowledge
on how certain management prescriptions affect
wildlife (Figure 2). This process also allows for con-
tinuous adjustment of treatments and management
based on monitoring data. In this process, the biolo-
gist identifies a series of achievable goals (e.g., 30%
more browse in an area ). These goals are measured
in units that relate directly to management objec-
tives. The biologist then selects a series of methods
that are believed to achieve the management objec-
tive (e.g., prescribed burn). Since the biologist is not
100% confident that the prescription will have the
desired result, it is applied on a limited ( incremen-
tal) basis, and then the result is monitored. From the

PLANNED

TEST OF

PUBLIC RESPONSE

POPULATION,
HABITAT
STATUS

TREATMENT

GOALS
P

PI AMMFH

PLANNED

TEST OF

BIOTIC RESPONSE

Figure 1. The linear-comprehensive approach to

wildlife management (after Bailey 1982).

Figure 2. The cyclic-incremental approach to

wildlife management (after Bailey 1982).

26

Data Types

monitoring data, the biologist can determine if the
habitat management prescription or land use needs
to be modified.

In this process, the biologist determines the
concerns and needs of the public in formulating
management objectives. It is also extremely impor-
tant to inform the public about any findings made
during the monitoring of a management prescription.
This will help the biologist reinforce or modify man-
agement objectives. Because prescriptions are con-
ducted and monitored on a site-specific basis, the
biologist has a better understanding of the local situ-
ation. This type of procedure is more likely to gain
public support, as opposed to use of rigid proce-
dures developed for large regions.

In his description of these concepts, Bailey
(1982) points out that wildlife management is an art,
not a science. Wildlife managers apply the science
of biology, but the application of this science is an
art. The process of cyclic-incrementalism allows the
biologist to accumulate site-specific data relative to

publicly-accepted wildlife management objectives,
and slowly build a basis for management decisions.
Because this process does not require large-scale
data collection, it can be scaled to available budget.

The key to successful inventories and monitor-
ing studies is in establishing achievable and manage-
ment-related objectives, and in orienting the type
and level of data collected to these objectives. It
serves no purpose to determine the distribution of
all wildlife in several habitat types if the objective is
to determine how grazing practices affect certain key
wildlife species. Data should be collected to solve
specific management problems.

Throughout the chapter, I have highlighted dif-
ferent types and levels of data. This information
should be used in making decisions on the type of
data needed to answer specific management ques-
tions. I do not recommend using a standard set of
methods and data, but rather an assortment of tech-
niques and information, depending on individual
objectives, goals, priorities, and budget.

Data Types

27

LITERATURE CITED

BAILEY, J.A. 1982. Implications of "muddling through" for
wildlife management. Wildl. Soc. Bull. 10(4):363-369.

BROWN, D.E., C.H. LOWE, and C. PASE. 1979. A digitized
classification system for the biotic communities of
North America, with community (series) and associa-
tion examples for the Southwest. J. Arizona — Nevada,
Acad. Sci. 14, Supplement 1.

BURY, R.B. 1982. Structure and composition of Mojave
Desert reptile communities determined with a re-
moval method. Herpetological Communities, U.S. Dep.
Inter., Fish and Wildl. Serv., Wildl. Res. Rep. 1 3.

CAUGHLEY, G. 1977. Analysis of vertebrate populations.
John Wiley and Sons, London. 234pp.

EBERHARDT, L.L. 1978. Appraising variability in popula-
tion studies. J. Wildl. Manage. 42(2):207-238.

EMLEN, J.T. 1971. Population densities of birds derived
from transect counts. Auk 88.323-342.

GREEN, R.H. 1979. Sampling design and statistical meth-
ods for environmental biologists. John Wiley and
Sons, New York, NY.

GYSEL, L.W. and L.J. LYON. 1980. Habitat analysis and
evaluation. Pages 305-327 in Schemnitz, S.D., ed.
Wildlife Management Techniques Manual. The Wildl.
Soc, Washington, DC.

HAIR, J. D. 1980. Measurements of ecological diversity.
Pages 269-275 in Schemnitz, S.D., ed. Wildlife Man-
agement Techniques Manual. The Wildl. Soc, Wash-
ington, DC.

HAYS, R.L., C. SUMMERS, and W. SEITZ. 1981. Estimating
wildlife habitat variables. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-81/47. 111pp.

HENDRICKSON, DA., W.L MINCKLEY, R.R. MILLER, D.J.
SIEBERT, and PH. MINCKLEY. 1980. Fishes of the Rio
Yaqui Basin, Mexico and United States. J. Arizona,
Acad. Sci. 15(3>65-106.

HORN, H.S. 1966. Measurements of overlap in compara-
tive ecological studies. Am. Nat. 100:419-424.

JONES, KB. and C. GLINSKI. 1983. Microhabitats of lizards
and southwestern riparian community. Pages 342-
346 in Johnson, R.R., CD. Ziebell, DR. Patton, P.F.
Ffolliott, and R.H. Hamre, tech. coords. Riparian Eco-
systems and their Management: Reconciling Conflict-
ing Uses. U.S. Dep. Agric, Forest Serv., Gen. Tech.
Rep. RM-120.

, LP. KEPNER, and T.E. MARTIN. 1985. Species of

reptiles occupying habitat islands in western Arizona:
A deterministic assemblage. Oecologia. In press.

KERR, R.M. and K BROWN. 1977. Data requirements for
terrestrial wildlife habitat inventory in Classification,
Inventory, and Analysis of Fish and Wildlife Habitat.
Proc Natl. Symp. Phoenix, AZ, January 1977. U.S.
Govt. Printing Office, Washington, DC.

KUCHLER, AW. 1964. Map of potential natural vegetation
of the conterminous United States. Am. Geophysical
Soc, Spec. Publ. 36.

MANNAN,R.W„ M.L MORRISON, and EC. MESLOW. 1984.
Comment: The use of guilds in forest bird manage-
ment. Wildl. Soc. Bull. 12(4):426-430.

MATTHEWS, W.J. and L.G. HILL. 1980. Habitat partitioning
in the fish community of a southwestern USA river.
Southwestern Nat. 25(1): 51-66.

OTIS, D.L., KP. BURNHAM, GC. WHITE, and DR. ANDER-
SON. 1978. Statistical inference from capture data
on closed animal populations. Wildl. Monogr.
62:1-135.

PATTON, DR. 1975. A diversity index for quantifying
habitat "edge." Wild. Soc. Bull. 3(4):171-173.

PHILLIPS, E.A. 1959. Methods of vegetation study. Henry
Holt and Company, Inc. 107pp.

PIMENTEL, R.A. 1979. Morphometries: The multivariate
analysis of biological data. Kendall/Hunt Publ. Co.,
Dubuque, IA.

RALPH, C.J. and J.M. SCOTT, eds. 1981. Estimating num-
bers of terrestrial birds. Allen Press, Inc., Lawrence,
KS.

RICKLEFS, RE. and M. LAU. 1980. Bias and dispersion of
overlap indices: Results of some Monte Carlo simula-
tions. Ecology 6 1(5): 10 19- 1024.

ROTENBERRY, J.T. and J.A. WIENS. 1980. Habitat struc-
ture, patchiness, and avian communities in North
America steppe vegetation: A multivariate analysis.
Ecology 61:1228-1250.

ROUTLEDGE, R.D. 1980. Bias in estimating the diversity
of large, uncensused communities. Ecology
6l(2):276-281.

SCHAMBERGER, MA, AH. FARMER, and J.W. TERRELL.
1982. Habitat suitability index models: Introduction.
U.S. Dep. Inter., Fish and Wildl. Serv. FWS/OBS-82/10.

SEBER, G.A.F. 1973 Estimation of animal abundance and
related parameters. Griffin, London. 506pp.

SHORT, H.L 1983. Wildlife guilds in Arizona desert habi-
tats. U.S. Dep. Inter., Bur. Land Manage., Tech. Note
362. 258pp.

. 1984. Habitat suitability index models: The Ari-
zona guild and layers of habitat models. U.S. Dep.
Inter., Fish and Wildlife Serv. FWS/OBS-82/ 10.70.
37pp.

and KP. BURNHAM. 1982. Technique for structur-

ing wildlife guilds to evaluate impacts on wildlife
communities. U.S. Dep. Inter., Fish and Wildl. Serv.,
Special Sci. Rep. 244. Washington, DC.

SMITH, M.H., R.H. GARDNER, J.B. GENTRY, D.W. KAUF-
MAN, and M.J. OTARRELL 1975. Density estimation
of small animal populations. Pages 25-53 in Golley,
F.B., K Petrusewicz, and L. Ruszkowski, eds. Small
Mammals: Their Production and Population Dynamics.
Int. Biol. Prog. 5, Cambridge Univ. Press, London.

THOMAS, J.W. 1979. Wildlife habitats in managed forests:
The Blue Mountains of Oregon and Washington. U.S.
Dep. Agric, For. Serv., Agric. Handbook 553

TIPTON, A.R. 1980. Mathematical modeling in wildlife
management. Pages 211-220 in Schemnitz, S.D., ed.
Wildlife Management Techniques Manual. The Wildl.
Soc. Washington, DC.

WHITE, G.C., DR. ANDERSON, KP. BURNHAM, and D.L
OTIS. 1982. Capture-recapture and removal methods
for sampling closed populations. Los Alamos Natl.
Laboratory LA-8787-NERP.

WILLIAMS, G.L., KR. RUSSELL, and W.K SEITZ. 1977.

Pattern recognition as a tool in the ecological analysis
of habitat in Marmelstein, A., ed. Classification, Inven-
tory, and Analysis of Fish and Wildlife Habitat. U.S.
Dep. Inter., Fish and Wildl. Serv. FWS/OBS-78/76.
604pp.

28

Data Types

LITERATURE
REVIEW

Ora Wagoner1

Bureau of Land Management

Service Center

Denver, CO 80225-0047

Editor's Note. Too many biologists duplicate work
done by others. Libraries can help you determine
what's already been accomplished in your field

Consulting existing information is a mandatory
step in preparing for an inventory or monitoring
study. This involves more than a literature review.
Today's libraries search literature and other infor-
mation sources with electronic data bases, and they
pool their resources by operating in networks. They
can also provide you with unpublished data and
referrals to current researchers or other information
sources, such as museum records.

"Immediately after characterizing a problem, and im-
mediately before initiating any action, find out who
might already know the answer. Remember to ask
the librarian."

Alan Speigel

INTRODUCTION

Wildlife inventories and monitoring studies
should begin with searches of existing literature to
determine past research and current studies. The tre-
mendous growth in stores of knowledge has made
searching the literature a complicated procedure re-
quiring differing techniques and special tools.

A librarian's approach to literature searching is
systematic, to investigate all possible sources of in-
formation. This chapter provides biologists with
guidance on how scientific literature and other infor-
mation are accessed, and introduces BLM employees
to the services provided by the BLM Library (see
Appendix I).

Wildlife biologists need to use libraries and their
resources to ensure that they perform thorough liter-
ature searches and that they access other sources of
information essential to wildlife inventories and
monitoring studies. Library users must allot time for
literature searches to be performed, as materials
must often be obtained from distant sources.

The processes used to get information depend
on the nature and amount of data requested. In a
typical search, an information specialist first checks
printed sources available at the library; if not suc-
cessful, he or she then uses computer networks to
inquire about information from other locations. The
delivery of documents from other locations can take
4 to 6 weeks, so biologists should plan on this
amount of time to conduct their searches.

This chapter covers published materials includ-
ing serials, reports, books, and other documents, and
specialized systems of organized data such as com-
puter data bases and museums. It also provides sug-
gestions on accessing unpublished data.

1 Present Address: Dudley Knox Library, Naval Postgraduate School,
Monterey, CA 93943-5002.

Literature Review

29

PRINTED MATERIALS

Printed materials include published and unpub-
lished documents. Information about published mate-
rials is found in many different forms. Announce-
ments of new books or reviews provide the best
method for keeping abreast of newly published mate-
rial in a specific subject area. These can be found in
many journals. Information from book publishers, so-
cieties, and universities that publish in particular
areas of interest can also be invaluable for learning
of new publications. A review of Subject Guide to
Books in Print (published by R. R. Bowker Com-
pany, 205 E. 42nd St., New York, NY 10017) will re-
veal most new books in all subject areas as well as
older books that are still available from the publish-
ers. Appropriate subject headings to use in searching
Books in Print are wildlife management and animal
populations.

Because book material becomes quickly out-
dated, the most current printed information is found
in serials or periodicals. The process of finding ap-
propriate information begins with locating the appro-
priate periodicals.

There are various processes available for identi-
fying desired periodicals. Most libraries have access
to current subscriptions for the major serial publica-
tions in a given subject area. A look at the Table of
Contents in each of the appropriate periodicals will
identify new articles covering selected areas of inter-
est. For assistance in selecting periodical titles for
any subject area, as well as obtaining subscription in-
formation, Ulrich's International Periodicals Di-
rectory is a valuable source to use. Suggested sub-
ject headings to check in Ulrich's Directory are
Biology-Zoology, Environmental Studies, and
Conservation.

If you are not fortunate enough to have easy ac-
cess to a library that subscribes to the desired jour-
nals, personal subscriptions must be considered.
They can be expensive, however, and can require
time and space for maintenance.

You will probably want to consider some other
alternatives. One is to access Current Contents.
There are several subject-specific versions of Current
Contents, each of which is issued monthly. Each is-
sue reprints the Table of Contents from selected
journals within its specialty area.

Current Contents/Agriculture, Biology, and
Environmental Sciences, in particular, covers over
1 ,000 periodicals and 300 books of interest in wild-
life inventory and monitoring.

Another alternative is to access periodicals
through abstracting and indexing services. These are
subscription services, usually appearing monthly;

they provide an index to articles, reports, books, and
other documents related to specific subject areas.
Separate services exist for most subject areas, such
as Engineering Index for engineers, or Chemical Ab-
stracts for chemists.

Indexes reference citations under subject head-
ings and authors, giving full citations. For periodical
articles, a typical reference will list author, article ti-
tle, title of the periodical in which it was printed,
volume number, issue number, page numbers, and
date published.

Abstracting services provide brief summaries
about items in addition to information about how
they can be located. (Indexing services include only
title and bibliographic information. ) Some of the
more useful abstracting and indexing services to
consult for a literature search covering wildlife are

1. Agricultural Index

2. Bibliography of Agriculture

3. Bio Research Index (supplements Biological
Abstracts)

4. Biological Abstracts

5. Biological and Agricultural Index

6. Chemical Abstracts

7. Ecology Abstracts

8. Science Citation Index

9. Wildlife Abstracts (Cumulations of Wildlife
Review)

10. Wildlife Review

1 1 . Zoological Record.

Other indexing and abstracting services that can
be invaluable in a literature search are

1. Cumulative Subject Index to the Monthly Cata-
log of U.S. Government Publications 1900-1971

2. Supplementary Index to U.S. Publications

3. Masters Abstracts

4. Dissertation Abstracts.

These services cover government documents and
doctorate and masters theses. The indexed publica-
tions may be published or unpublished documents. A
biologist should not overlook the usefulness of these
types of publications and other unpublished mate-
rials that may be identified through other sources,
such as peer contacts.

Each of these services covers a different selec-
tion of publications that deal with the subject of
wildlife, including inventory and monitoring. The
BLM Library, any large public library, or a technical
research library can help you access these sources.
(See Appendix I for more detailed information on
these services.)

30

Literature Review

Bibliographies

Bibliographies covering specific subject areas
cannot be overlooked as sources of information. De-
pending upon the date of compilation, a bibliography
may be excellent for background information if the
subject in question must include older, but still ap-
plicable, materials. Some useful bibliographies for
wildlife subject areas are

1. U.S. — Canadian Range Management, 1935—1977:
A Selected Bibliography on Ranges, Pastures,
Wildlife, Livestock, and Ranching;

2. U.S. — Canadian Range Management, 1978-1980:
A Selected Bibliography on Ranges, Pastures,
Wildlife, Livestock, and Ranching; and

3. Wildlife Abstracts — the entire series, from 1935
through 1980.

Other, more subject-specific bibliographies (exam-
ple, a bibliography on the pronghorn antelope) are
valuable resources. They can be found in a variety of
places within library collections.

Another often-overlooked source of bibliogra-
phies is the references listed at the end of the article
or report the biologist already has on an individual
topic.

Once an article is identified, a biologist can re-
quest a copy of the article from the author, library,
or other appropriate organization. Preprinted reprint
request cards are useful for this purpose.

Some of the major journals published in wildlife
research are outlined in Appendix II. Not listed be-
cause of their number are journals that are specific
to a State; examples are California Fish and Game
and Wyoming Wildlife. These journals are highly use-
ful even for areas outside of the specified State. They
can be identified through the use of indexes at the
BLM Library or other large library facility.

The use of bibliographic computer data bases is
the most efficient method to develop current bibli-
ographies on any topic. Online data bases have ex-
isted since the mid 1960s, but they were primarily
private files accessible only to special groups. In the
early 1970s, online literature searching data bases
were made available to the general public, thus limit-
ing the time-consuming step of manually searching
tables of contents, indexes, and abstracts. With a few
exceptions, coverage of data bases is generally lim-
ited to literature from the 1970s to the present.
Therefore, any researching into literature published
before 1970 must be done manually.

To access the hundreds of data bases available
requires the use of a vendor with a software system
which provides the commands for searching. The
three most frequently used vendors for literature
searching are

DIALOG — operated by the Lockheed Corporation,

ORBIT — operated by SDC Search Service, and

BRS — operated by the Bibliographic Retrieval Ser-
vices Company.

Since knowledge of software is necessary to search
efficiently, most researchers access these systems
through trained librarians. For example, the com-
puter version of Biological Abstracts on DIALOG
goes back to 1969, whereas the printed version be-
gan publication in 1927. Experience has shown that
researchers sometimes become so impressed with
the computer that they tend to forget that not all the
published literature is contained within its files. If
you desire a comprehensive search, be sure to use
both manual and computer searches. Computer
searches may be adequate if only recent data are
desired.

COMPUTER SYSTEMS

After a researcher has reviewed the appropriate
manually-searchable sources, the next step is to con-
sult computer data bases. These data bases generally
fall into two major categories — bibliographic and
nonbibliographic.

Bibliographic data bases produce results similar
to searching the indexing and abstracting services
previously mentioned. These data bases contain cita-
tions to specific articles, books, and reports, and can
be searched by keywords, title words, authors, or
other appropriate identifiers. Nonbibliographic data
bases differ in that they contain actual data and usu-
ally do not refer the user to a printed source.

BLM librarian performing online computer search.

Literature Review

31

For interactive searching, librarians use termi-
nals and communicate directly with computers to
get the desired citations. Searches for wildlife infor-
mation will most frequently be performed in the
data bases AGRICOLA, BIOSIS, CAB, and CRIS, ex-
panding to other sources depending upon the topic.

Commercial sources represent only a portion of
the many computer sources which provide informa-
tion on wildlife inventory and monitoring. Other
sources of interest may or may not be bibliographic,
and there may be special requirements for access.
Noncommercial bibliographic files are basically
structured the same as the commercial sources dis-
cussed above. Data bases are often maintained and
updated by the originating organizations, and access
to the data bases is made available only through
their approval.

Several noncommercial data bases have been de-
veloped by Federal agencies to assist in meeting
their missions. Information on specific noncommer-
cial information systems can be found in DuBrock et
al. (1981). Three data bases of primary interest for
inventory and monitoring of wildlife habitat are the
Fish and Wildlife Reference Service, RUNWILD, and
the Procedures System. Other data bases pertaining
to wildlife-related subjects are summarized in
Appendix III.

Fish and Wildlife Reference Service (FWRS)

FWRS is a computer index of documents from
the Federal Aid in Fish and Wildlife Restoration Pro-
gram (Pittman-Robertson and Dingell-Johnson Acts),
the Anadromous Fish Conservation Program, the En-
dangered Species Grant Program, the Cooperative
Fishery and Wildlife Research Units, and State game
and fish agencies. The data base does not include
most routine publications of the U.S. Fish and Wild-
life Service.

The system allows access to research informa-
tion from the State level by providing an index to se-
lected State agency technical reports. Knowledge of
indexed publications can be gained from the FWRS
quarterly newsletter, which lists new publications, or
by obtaining a computer search that will also include
older publications. A computer search can be or-
dered directly from FWRS; there is a charge unless
the requester is employed by the U.S. Fish and Wild-
life Service or a cooperating agency. This data base is
available as a private file on the DIALOG System. The
BLM Library has received permission to access it
through DIALOG and can accomplish searches for
BLM employees.

RUNWILD

RUNWILD is a storage and retrieval system con-
taining wildlife information that can be accessed
from three levels: inventory, species-habitat associa-
tions, and management data. Developed in 1973 by
the USDA Forest Service, it is stored on a computer
in Fort Collins, Colorado. RUNWILD can be used for
875 vertebrate species and subspecies in the South-
west, but its format is applicable to any geographical
area. Casner et al. 1978, Lehmkuhl and Patton 1982,
and Patton 1978, 1979a,b provide details of the
system.

Access to RUNWILD is available to any Federal
employee and can be obtained by contacting the
USDA, Fort Collins Computer Center (FCCC) to es-
tablish an account. They will supply a form that re-
quires the requester to provide information neces-
sary for computer access codes and passwords.
Accounting and use information is periodically is-
sued to the user from the FCCC. Each account num-
ber is billed for use charges as necessary. The data
base manager, located in the Region 3 Forest Service
in Albuquerque, New Mexico, is the contact person
for questions about the system and for obtaining a
data base password. Because this is a nonbiblio-
graphic data base and the user is usually searching
for specific raw data, the BLM Library does not per-
form these searches for users. I recommend that one
account be opened for use by interested personnel
at a specific office.

New additions and updates are being made to
the file on a continuing basis. RUNWILD II was made
available in November 1978 and RUNWILD III in
1982. The data base contents have been made avail-
able on microfiche through the National Technical
Information Service (NTIS). Information on the use
of the microfiche is given in Patton ( 1979a).

Procedures

"A Procedure for Describing Fish and Wildlife"
(nicknamed Procedures) was developed by the U.S.
Fish and Wildlife Service for uniformly describing
characteristics of fish and wildlife populations
(Cushwa et al. 1980; Mason et al. 1979). The system
outlines procedures for collecting ecological facts
concerning species of vertebrates and selected inver-
tebrates and can be applied to information gathering
for species in any area. The Procedures system has
been used in data collection in several States, includ-
ing Pennsylvania, Colorado, Wyoming, Missouri, Illi-
nois, Virginia, Kentucky, and Tennessee. The data
bases are administered by the game and fish agency
for the State involved, and access is obtained by con-
tacting the data base manager at that organization to
set up an account for interactive searching or for re-
questing a batch search.

52

Literature Review

The Procedures system provides a useful and
valuable methodology for maintaining fish and wild-
life species data regardless of the geographic area
being considered. The system summarizes informa-
tion on wildlife habitat use and species distribution,
status, and life history. Workbooks that provide spe-
cific coding descriptions are used when preparing
data for input into the data bases. The standard clas-
sifications required for coding the information are
widely accepted, recognized standards that are easily
understood and applied. For categories in which
there were no accepted classification standards al-
ready in use, standards were developed and applied.

ACCESSING BIBLIOGRAPHIC COMPUTER
DATA BASES

Data bases may be accessed by anyone who has
a terminal and has made the necessary financial ar-
rangements. I recommend, however, that biologists
obtain bibliographic information through the ser-
vices of a librarian or information specialist. These
professionals are trained in techniques necessary for
efficiently and effectively using the variety of subject
data bases available. They also keep current on data
base changes by reading literature and attending up-
dated training sessions for experienced searchers. Bi-
ologists can best use their time reviewing the sub-
ject bibliographies that librarians provide for
publications of interest.

Many organizations perform literature searches
upon request. Most large public libraries and aca-
demic libraries offer the service. In most large cities,
there are commercial organizations that advertise
this kind of service. These sources usually charge ac-
tual costs for communications and offline prints and
add a percentage as a service charge. Some private
individuals also provide computer literature search-
ing for a fee. Consequently, anyone can have access
to bibliographic data bases and thus to current re-
search data. The key is knowing where to find the
organizations or individuals that will perform the
searching.

Within BLM, access to computer literature
searching services is available through the BLM Li-
brary. The Library maintains contracts with DIALOG,
SDC, and BRS, enabling access to more than 250 data
bases. The requester simply telephones the BLM Li-
brary with details of the request or sends a com-
pleted Library Search Request Form to the Library
(Figure 1).

To conduct an effective search, a librarian must
obtain specific information from the requester con-
cerning the data needed. In addition to a general
statement of the overall objective of the search re-
quest, the requester must provide key words. Key
words are any meaningful words that are important

to the topic being researched. They are terms or
phrases that must be present in the citations for the
resultant bibliography to be pertinent to the search
request. Names of authors and organizations may
also serve as specific keys upon which a search can
be based.

Librarians develop search strategies from the re-
lationships that exist among key words, grouping like
concepts. Search strategies are executed in various
data bases that cover subjects of interest, and bibli-
ographies are returned to requesters. Employees
then have the opportunity to review the listings,
mark those items of interest, and return the lists to
the Library. The BLM Library then obtains copies or
loans of the desired items. (Appendix I describes
some of the data bases available through the BLM Li-
brary that contain wildlife information. )

OTHER INFORMATION SOURCES

Information from other than published and elec-
tronic information sources can be used to great ad-
vantage in research for wildlife inventory and moni-
toring studies. Two major areas to be considered
will be discussed in the remainder of this chapter.

Museums

Museum data can be a useful source of informa-
tion for the wildlife biologist (See Figure 2). Several
museums have important systematic biological re-
search resources that can be used for natural history
information on specimens. Of special interest in the
consideration of this data source is information on
species location. Information such as place of collec-
tion, date of collection, specific locality, and county
are required in the records for a species.

Museum collections that would be useful for
wildlife inventory research are found in many differ-
ent institutions (see Appendix IV). These organiza-
tions produce many publications that contain infor-
mation about their collections. Information regarding
what they publish can be obtained from a variety of
sources, mentioned by Moore (1980). Assistance in
obtaining any of the references mentioned can be
obtained from a large public or a technical library.
BLM personnel should seek help from the BLM Li-
brary at the Service Center.

Once a museum source is identified, it is possi-
ble to locate catalogs of that museum's holdings or
specimen listings. One can then access the museum's
records covering the species or geographical areas of
interest. The searcher should be aware of a major
problem that sometimes exists with museum data. If
only general localities are recorded for an observa-
tion, the usefulness of the record for studies of distri-
bution and habitat may be lost. Accurate association

Literature Review

33

J

4

M/

Form 1279-2
(April 1979)

UNITED STATES
TMENT OF THE W
BUREAU OF LAND MANAGEM

DEPARTMENT OF THE INTERIOR ^ J-- V\ i.. \

lEQUESr^ ky nA

ftp ^

w

1. Requested by (name) £^J^ g ' /^ZJ^,

Date

9^^/ 7, /far

Office* and Division

relepnone (include nieti <

£ t\

Telepffone (include aika code)

Oo'l)2V?- 757ft

a. Needed by ^,^^,/^r

' A , z. Subject of search (describe)

I 3. Give5 specific terms and/or phrases that are considered relevant. Include synonyms, closely related phrases, and

generic names when appropriate. Use scientific and technical as well as common vocabulary. (Attach additional
sheet, if necessary.) ■

rgC\llDE 4£ HhHy^fc^t^, dsiUU-rLfa , rruw^c^, tYUun^TOfy ^fattas,

pftf C l PZ. I fa >State limitations desired (e.g. geographical, language, dates, etc.)

5. Give two or three references already located which you consider relevant to the subject requested.

iUO'tdt A.L. Icl7(j,. £lMs bt-hciis-Lrt- i^ Su.£att.e-*s 1~o

^ uUtcruo T^oul^ ^curiae* £r^ &bi*rTh,Cj,<*£Ci.eU.' UJiftCrrv-nty-

n&A/t 7JU.IAA, r<Uvr<UiU<u <%*i, UQV^U jrf^

P-

-me ^ta^cAor, trvrf ftu^ tuxt not rttusurJii .

6. Desired scope (check appropriate item)

[T^fBroad search designed to retrieve as many as possible of the relevant citations, but which might also retrieve
many irrelevant citations.
J Narrow search designed to retrieve some of the relevant citations, but with few accompanying irrelevant
citations.

Remarks JL ,/W^ L

:-y\ L*^-

GPO 650-423

Figure 1. A completed BLM Library Search Request Form.

34

Literature Review

MUSEUM OF VERTEBRAU ZGJLGGY DATA FUR fclRDS CGLLEUEU IN CULURAUO

Ob/v-7/bb

PAGt 1

NC. OF ITLMS IN CUtRY RESPONSE: 7
NO. OF ITEMS IN DATA BANK: 2743b
PERCENTAGE OF kESPGhSE/TGTAL DATA BANK:

0.03*

x*v

PARUL1NAE

DENDR01CA CORONATA

MVZ CAT. NO.: 37399
SEX: h age : juv.

COUNTRY: COLORACO

MEMGRABlLlS
SKIN: STUDY SKEL.: NONE

COLLECTOR: DILLE FRED M.

STA*t"/CO./PRLV.

CULL .n j. ;
LARlMtR CU.

COLL. NO.
LARIMER CO.

ELEV: ; LOCALITY: ESTES PARK

DATE: 4 AUGUST 19J? M.: RlFRC.:

FAT: MOLT:YES S TOM.COM . :NG CLLORS:nO

REMARKS: ED:MGLT-StE SPECIMEN.
DENDR01CA CGRUNATA MEMLRAblLlS

MVZ CAT. NO.: 44533 SKIN: STUDY SKEL.: NONE

SEX: K AGE : U AD. COLLECTOR: DILLE FRED Ml

COUNTRY: COLORADL STAJE/CC ./PROV ,

llev: --- --: LOCALITY: ESTES park

date: 6 july 1904 mt.: reprc.

fat: molt:no stum.ccnt . :n0 colcrs:no

remarks: ed:mst.
oehorgica coronata mem0rab1l1s

mvz cat. no.: 78903 skin: stuoy skel.: n[
sex: f age : u ad. collector :f£ktr f.h. coll. no.:

country: colorado ,, state/co. /prov . : larimtk co.

ELEV: — — ; LOCALITY^fBERTHCUD t

DATE: 18 AUGUST 1897 ^W I ■ I m" REPRC.

FAT: MOLT:YES ST0M.COM. :N0 COL0RS:NO

REMARKS: EO:MOLT-SEE SPECIMEN.
N1LS0MA PUSILLA P1LE0LATA

MVZ CAT. NO.: 76915) SKIN: STUOY SKEL.: NONE

SEX: L> AGE : U COLLECTOR: FOWLER F.H.

COUNTRY:

COLORADO

ELEV:

— --;

OATE:

9 JULY

FAT:

COLL. NO. !
LARIMER CO.

STATE/CC./PROV.:

LOCALITY: LONG'S PEAK; ROCKY MOUNTAIN NAT. PARK

1899 NT.: kEPRO.:

MOLT:NO STQM.COM.:NO COLORS:no
remarks: basin above timberline
n1ls0ma pusilla pileolata

mvz cat. no.: 7e916 skin: study skel.: nene

SEX: M AGE : U COLLECTOR: FO*LER F.H. COLL.nu.:

COUNTRY: COLORADO ST ATE/CO ./PROV . : LARIMER CO.

ELEV: --; LOCALITY: LONG'S PEAK; ROCKY MOUNTAIN NAT. PARK

DATE: 10 JULY 1899 HT . : REFRU.:

FAT: MOLT i Hi — fcimj.CCM.:NO COlCRS:nO

REMARK S/*6AS IN AEOVE TIMBERLINE^
TURD IN AE
MYADESTES TCWNSENDI TCmRSENCI

MVZ CAT.NC.: 107381 SKIN: STUDY sKi.: NCNE

SEX: M AGE : U COLL EC TOR r^^- LULL.Nl|

COUNTRY: COLORACO ^,-"-»^i^ ST ATL/CO./PKOV . : LAklMtk CU . ..fl^

ELEV: ; locality: lovelanc

DATE: -- MAY 1904 NT.: kEPRL.:

FAT: --- MQL1»NU _ST0M.COM . :N0 C0LCRS:nO

REM/

MQLl»NO S

U.:

LAKlMcR CU. _.A\

MUiLl

rt^bM ZGjLGGY DATA FOR clRDS CCLLiCTtU IK COLORADO

Ot>/C7/6b

PAGt _

TURD1NAE

S1ALIAX LUkRUCulb-S

MVZ CAT.NC.: 396t>i
JEX: F AGE : U

jUNTRY: COLORADO

■ i LUCnia

fit : 1 APRIL 1900

-f AT i -■■ '

FEMARKS:

SKIN: STUOY SKtL.: NCNL

CCLLEC10F: DILLE F.M,

FT. COLL INS

M.: KEFKL.:

NO stom.ccm.:no culcrs:nl

STATE/Cu./PkLV.:

CULL .N^i.
LARIMER CU.

144

20t)

it>2

318

319

Figure 2. Major scientific museums have computer bases on their collections that contain valuable

information on species distribution; an example from the Museum of Vertebrate Zoology at the University
of California, Berkeley is shown here.

Literature Review

55

of the specimen with a particular habitat type may
be impossible. The site named may be some distance
from the actual collection or observation point, or,
in some instances, more than one point in the gen-
eral area might have the same name. Care should
also be exercised in checking the records to be sure
they correctly represent the specimen being
reported.

Personal Contacts

The value of knowing who is prominent in a
specific area of research cannot be overlooked as a
potential source for valuable information. Generally
these specialists are willing and anxious to tell any-
one about their research or send copies of their pub-
lications upon request. A way to identify specialists
is to make note of recurrent names in subject bibli-
ographies. Libraries can conduct author searches us-
ing the data bases discussed earlier; thus, a bibliog-
raphy of writings done by a specific person can be
generated. In addition, the CRIS Data Base and the
Federal Research in Progress Data Base, which are
available through DIALOG, are data bases that cover
current research information. To learn more about
the research project listed there, the specialist can
contact the researcher whose name is given.

The key step in using this method is locating
the specialist. Some useful non-computer tools in
your library that can correctly identify and locate an
individual are --

• directories of organizations

• information in proceedings of meetings, sympo-
sia, and conferences

• subject matter directories such as the Conserva-
tion Directory

• who's who-type publications

• introductory remarks, short biographical infor-
mation, and title page information in some pub-
lications

• inquiries to people who have already been iden-
tified.

In many instances, conversation with a colleague can
lead to information sources that would be impossible
to otherwise identify.

The telephone becomes an important research tool.

SUMMARY

The importance of thorough literature searches
cannot be overemphasized. When biologists have
conducted studies that are similar to yours, you can
realize considerable savings in time and resources. In
some instances there may be little or no information
available in response to an inquiry, but that in itself
is useful to know. If no previous studies exist, the re-
sults of your study may be invaluable.

One of the major focuses of libraries today is in-
terlibrary cooperation and shared resources. Even if
one has no idea where to begin to look for some
person, publication, or resource, a check with your
library or information center is a good way to begin.

Most large libraries and many technical libraries
have computer services that enable them to identify
book and periodical holdings of other libraries
throughout the Nation. These services will save time
and increase the thoroughness of literature reviews.
At the BLM Library, the needs of field employees are
given highest priority. For library and information
needs, it is the place to begin.

36

Literature Review

LITERATURE CITED

CASNER, W.B., B. KULONGOWSKI, DR. PATTON, and S.J.
PINKERTON. 1978. RUNWILD— For the UNIVAC
1100 Series: implementation and maintenance. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. RM-51A. 31pp.

CUSHWA, C.T., C.W. DUBROCK, N.D. GLADWIN, G.R.
GRAVATT, R.C. PLANTICO, R.N. ROWSE, and LJ.
SLASKI. 1980. A procedure for describing fish and
wildlife for Pennsylvania: summary evaluation report.
U.S. Dep. Inter., Fish and Wildl. Serv. FWS/OBS-79/19A.
15pp.

DUBROCK, C.W., D.N. GLADWIN, W.T. MASON, Jr., and
C.T. CUSHWA. 1981. State-of-the-art of fish and wild-
life species information systems in the United States.
Trans. North Am. Wildl. Nat. Resour. Conf. 46:156-
170.

LEHMKUHL, J.F. and DR. PATTON. 1982. RUNWILD
Wildlife/habitat relationships: user's manual for the
RUNWILD III data storage and retrieval system. U.S.

Dep. Agric, For. Serv. Southwestern Region. Wildl.

Unit Tech. Rep. 68pp.
MASON, W.T. Jr., C.T CUSHWA, LJ. SLASHI, and D.N.

GLADWIN. 1979. A procedure for describing fish and

wildlife: coding instructions for Pennsylvania. U.S.

Dep. Inter., Fish and Wildl. Serv. FWS/OBS-79/19.

21pp.
MOORE, J. L. 1980. Wildlife management literature. Pages

7.38 in S.D. SCHEMNITZ, ed. Wildlife management

techniques manual. 4th ed. Wildl. Soc. Washington,

DC. 686pp.
. 1978. RUNWILD: a storage and retrieval system for

wildlife habitat information. U.S. Dep. Agric, For. Serv.

Gen. Tech. Rep. RM-51. 8pp
PATTON, DR. 1978. RUNWILD: a storage and retrieval

system for wildlife habitat information. U.S. Dep.

Agric, For. Serv. Gen. Tech. Rep. RM-51. 8pp.
. 1979a. How to use RUNWILD data files stored on

microfiche. U.S. Dep. Agric, For. Serv. Res. Note RM-

377. 2pp.
. 1979b. RUNWILD II: a storage and retrieval system

for wildlife data. Trans. North Am. Wildl. Nat. Resour.

Conf. 44:425-430.

Literature Review

37

APPENDIX I. BLM Library

History

The BLM Library was established as part of the
BLM Service Center in 1970. The Library's services
continue to grow with advances in library technol-
ogy and information management. Today the BLM Li-
brary offers a full range of library services, including
reference, circulation, aid in acquisition/purchasing,
cataloging, interlibrary loan, plus automated and
manual literature searching.

Library Collection

The BLM Library contains more than 30,000 vol-
umes and subscribes to over 200 periodicals. This
collection covers a wide range of subjects covering
all aspects of land management, research and devel-
opment, computer science, and administration. The
library's collection is arranged according to the Li-
brary of Congress Classification System and is cata-
loged through the online shared cataloging system of
Online Computer Library Center (OCLC) Inc., a so-
phisticated online computer network that links over
2000 participating libraries.

Loans

The BLM Library can deliver any published ma-
terial to field offices. Field personnel are encouraged
to use their local resources first; if a publication is
not easily identified or obtained, they should contact
the BLM Library for assistance.

Most materials in the Library's collection can be
loaned for 4 weeks. If the item is small, a photocopy
may be made and sent to the requester for retention.

If the publication you need is not owned by the
Library, it will be located at another library and bor-
rowed on interlibrary loan. If the publication is avail-
able free of charge, you may be provided with the
information necessary to obtain your own copy.

Literature Searching

The BLM Library has literature searching capa-
bilities that enable the staff to access over 250 data
bases and quickly prepare bibliographies on almost
any subject. This service is available to all BLM per-
sonnel and may be requested by filling out a "Library
Search Request" Form (BLM Form No. 1279-2), or
by calling the Library.

The requester provides the librarian with a gen-
eral statement of the information needed and with
key words that can be used in building a search
strategy for entering data bases. The key words pro-
vided are important in that they determine how suc-

cessful the search will be. Terms must be pertinent
to the subject area to locate citations on the desired
request.

Reference

A full range of reference services is provided by
the Library staff. Requests involve short, specific in-
formation questions that may be answered by phone
or general reference questions that require more ex-
tensive searching. Comprehensive information re-
quests require identification and location of mate-
rials; response time will vary from several hours to
many days.

Referral

Often the most current or relevant information
you seek may not yet be published and may be avail-
able only from a specialist. The Library's staff can lo-
cate researchers or organizations that specialize in
your subject area. The staff has a good knowledge of
potential referrals from the personal contacts they
make, and from the computerized networks they
access.

Current Awareness

The Library's current awareness services are an
extension of literature searching. Users can be kept
current in selected subject areas through computer
based services called Selective Dissemination of In-
formation (SDI). Search strategies are stored and
saved as SDIs in selected data bases. As updated ma-
terials are added to these data bases, the SDI pro-
grams are executed and the Library is provided with
offline printouts that are forwarded to the requester.

Special Collections

Presidential Executive Orders (Numbered) June
1985-Present

Federal Register-All volumes available —
1936-Present

Code of Federal Regulations (All titles)

Memoranda — Instruction and Information —
Washington Office: 1964-Present

BLM Service Center: 1965-Present

38

Literature Review

Wildlife information data bases available through the BLM Library.

Data Base

Description

Coverage
Dates

Producer

Update
Frequency

Comments

AGRICOLA

The cataloging and
Indexing data base of
the National
Agricultural Library;
contains publications
of the USDA Forest
Service. Covers
worldwide journal and
monographic literature
in agriculture and
related subjects,
including animal
science, forest and
plant-related areas,
entomology, and
agricultural
engineering

1970-Present

National

Agriculture

Library

Monthly

Records are citations
only.

BIOSIS
Previews

Covers all aspects of
the life sciences,
drawing on all forms of
original published
literature for citations.
Corresponds to the
printed publications
Biological Abstracts
and Biological
Abstracts/RRM
(formerly entitled
Bioresearch Index),

1969-Present

Bioscience
Information
Service

Semimonthly

Records primarily
citation only, but
abstracts given for the
Biological Abstracts
records from July 1976
to the present.

CAB
Abstracts

A comprehensive file of
agricultural and
biological information
containing all records
in the 26 main abstract
journals published by
the Commonwealth
Agricultural Bureau.
Other publications,
including books and
journals, are scanned.
In some instances, less
accessible literature is
abstracted by
scientists working in
other countries.
Included in this data
base are Forestry
Abstracts.

1972-Present

Commonwealth

Agricultural

Bureau

Monthly

Significant papers
abstracted; less
important works
reported as citations
only.

Literature Review

39

Wildlife information data bases available through the BLM Library (continued).

Data Base

Description

Coverage
Dates

Producer

Update
Frequency

Comments

Comprehensive
Dissertation
Index (CDI)

Provides access to
records of more than
1,000 scientific and
technical papers
presented at over
1,000 major regional,
national, and
international meetings
each year. Primary
subject areas covered
include life sciences
and chemistry.

1973-Present

Cambridge

Scientific

Abstracts

Monthly

Records are citations
only, but availability
and ordering
information also given.

Current

Research

Information

Service

(CRIS)

A current awareness
data base for
agriculturally-related
research sponsored or
conducted by U.S.
Department of
Agriculture research
agencies, State forestry
schools, and other
cooperating State
institutions. Currently
active and recently
completed projects
(within the last 2 years)
are in the file. Included
among the subjects
reported are biological
sciences, natural
resource conservation
and management, and
environmental
protection.

Last two years

U.S. Department
of Agriculture,
Washington, DC

Quarterly

Abstracts included with
the records.

ENVIROLINE

Covers more than
5,000 international
primary and secondary
source publications
reporting on all
aspects of the
environment. Has a
section on wildlife
publications, as well as
many other subject
overages.

1971 -Present

Environment
Information
Center, Inc.

Monthly

Records contain
abstracts.

40

Literature Review

Wildlife information data bases available through the BLM Library (continued).

Data Base

Description

Coverage
Dates

Producer

Update
Frequency

Comments

Federal
Research
in Progress

Provides access to
information about
ongoing Federally-
funded research
projects in physical
sciences, engineering,
and life sciences.
Project descriptions
include: project title,
keywords, start date,
estimated completion
date, principal
investigator, performing
and sponsoring
organizations,
summary, and
progress report.

Current
reserch

National
Technical
Information
Service

Semiannual
(reload)

Abstracts included with
the records.

Fish and
Wildlife
Reference
Service

Includes indexed
documents from the
Federal Aid in Fish and
Wildlife Restoration
Program (Pittman-
Robertson and Dingell-
Johnson Acts), the
Anadromous Fish
Conservation Program,
the Endangered
Species Grants
program, Cooperative
Fishery and Wildlife
Research Units, and
State fish and wildlife
agencies. Documents
include reports,
published papers,
technical publications,
theses, and special
materials, such as
endangered species
recovery plans.

1950s-Present

U.S. Fish and
Wildl. Serv.,
Division of
Federal Aid

Quarterly

GPO Monthly
Catalog

The machine-readable
equivalent of the
printed Monthly
Catalog of United
States Government
Publications. Contains
records of reports,
studies, fact sheets,
maps, handbooks,
conference
proceedings, etc.,
issued by all U.S.
Government agencies,
including the U.S.
Congress.

1976-Present

U.S.

Government
Printing Office

Monthly

Records contain
abstracts.

Literature Review

41

Wildlife information data bases available through the BLM Library (concluded).

Data Base

Description

Coverage
Dates

Producer

Update
Frequency

Comments

Life Sciences
Collection

Gives worldwide
coverage of journal
articles, books,
conference
proceedings, and
report literature.
Contains information in
the fields of animal
behavior, biochemistry,
ecology, entomology,
genetics, immunology,
microbiology,
toxicology, and
virology, among others.

1978-Present

Cambridge

Scientific

Abstracts

Monthly

Records contain
abstracts.

SCISEARCH

A multidisciplinary
index to the literature
of science and
technology that
contains all the records
published in Science
Citation Index and
additional records from
the Current Contents
series of publications
that are not included in
the printed version of
Science Citation Index.

1974-Present

Institute for

Scientific

Information

Biweekly

Records are citations
only.

Water

Resources

Abstracts

Prepared from
materials collected by
over 50 water research
centers and institutes
in the United States.
Covers a wide range of
water resource topics
including water
resource economics,
ground and surface
water hydrology,
metropolitan water
resources planning
and management, and
water-related aspects
of nuclear radiation
and safety.

1968-Present

U.S. Department
of the Interior

Monthly

Citations contain
abstracts.

Zoological
Records

Provides extensive
coverage of the world's
zoological literature
with particular
emphasis on
systematic/taxonomic
information.

1978-Present

Biosciences
Information

Bimonthly

Records are citations
and contain systematic
classification of up to
six levels for the
organism cited.

42

Literature Review

,

APPENDIX II. Publications useful for wildlife studies.

Title

Frequency

Source

Publications

in which Indexed

Comments

American Birds

Bimonthly

National Audubon Society
950 Third Avenue
New York, NY 10022

Biological Abstracts

Major areas of interest are
the changing distribution,
population, migration, and
rare occurrences of birds
throughout the Americas.
Formerly issued as
Audubon Magazine,
Section 2; Audubon Field
Notes.

American Fisheries
Society Transactions

Bimonthly

American Fisheries

Society
5410 Grosvenor Lane
Suite 110
Bethesda, MD 20814

Biological Abstracts
Chemical Abstracts
Current Contents
Excerpta Medica
Oceanic Abstracts
Pollution Abstracts

Objectives of the Society
are conservation,
development, and wise
utilization of recreational
and commercial fisheries,
promotion of all branches
of fisheries science and
practice, and exchange
and dissemination of
knowledge about fish,
fisheries, and related
subjects.

Auk

Quarterly

American Ornithologists'

Union
%University of Illinois at

the Medical Center
Department of Anatomy
Box 6998
Chicago, IL 60680

Biological Abstracts
Chemical Abstracts
Current Contents

Scholarly articles on
various topics concerning
birds basic to the study of
animal or bird behavior.
Covers studies throughout
the world, but articles are
all in English.

Condor

Quarterly

Cooper Ornithological

Society, Inc.
Meriden Rd.
Lebanon, NH 03766

Biological Abstracts
Chemical Abstracts
Current Contents
Zoological Record

A scholarly journal
devoted to the biology
and behavior of birds in
the wild.

Copeia

Quarterly

American Society of
Ichthyologists and
Herpetologists

Department of Ichthyology

American Museum of
Natural History

New York, NY 10024

Biological Abstracts
Chemical Abstracts
Current Contents
Oceanic Abstracts

Concerned with fishes,
amphibians, and reptiles.

Fisheries

Bimonthly

American Fisheries

Society
5410 Grosvenor Lane
Bethesda, MD 20814

Current Contents

Articles are aimed toward
scientific research and
enlightened management
of aquatic resources.

Herpetologics

Quarterly

The Herpetologists'

League, Inc.
Department of Biology
University of Southwestern

Louisiana
Lafayette, LA 70504

Biological Abstracts
Chemical Abstracts
Current Contents
Zoological Record

The pumal is dedicated
to furthering knowledge of
the biology of amphibians
and reptiles.

Journal of
Herpetology

Quarterly

Society for the Study of
Amphibians and
Reptiles

Ohio University

Dept. of Zoology

Athens, OH 45701

Biological Abstracts
Current Contents
Zoological Record

Contains original research
articles and

comprehensive reviews on
amphibians and reptiles.

Literature Review

43

APPENDIX II. Publications useful for wildlife studies (continued).

Title

Frequency

Source

Publications

in which Indexed

Comments

Journal of
Mammology

Quarterly

American Society of

Mammologists
Vertebrate Museum
Shippensburg State

College
Shippensburg, PA 17257

Biological Abstracts
Biological and

Agricultural Index
Current Contents
Index Medicus
Oceanic Abstracts
Pollution Abstracts
Selected Water

Resources Abstracts

Covers the entire field of
mammalian biology, in-
cluding marine mammals.
Papers on each issue
may involve taxonomy,
ecology, genetics, behav-
ior, physiology, or distri-
bution of species.

Journal of Wildlife
Management

Quarterly

The Wildlife Society
5410 Grosvenor Lane
Bethesda, MD 20814

Biological Abstracts
Biological and

Agricultural Index
Chemical Abstracts
Science Citation Index

Topics ranging from habi-
tat management to the
effect of toxic agents on
animals and plants and
other wildlife management
issues.

North American
Journal of Fisheries
Management

Quarterly

American Fisheries

Society
5410 Grosvenor Lane
Suite 110
Bethesda, MD 20814

The society promotes the
development of all
branches of fishery sci-
ence and practice, and
the conservation, devel-
opment, and wise utiliza-
tion of fisheries, both
recreational and commer-
cial. Journal reports and
papers reflect these ob-
jectives.

North American
Wildlife and Natural
Resources
Conference
Transactions

Annually

Wildlife Management

Institute
709 Wire Bldg.
Washington, DC 20005

The institute promotes the
better management and
wise use of all renewable
natural resources in the
public interest.

Raptor Research

Quarterly

Raptor Research
Foundation, Inc.

Dept. of Zoology

161 WIDB

Brigham Young
University

Provo, UT 84602

Focuses on predatory
birds. Covers all aspects
of general ecology, natu-
ral history, management,
and conservation of diur-
nal and nocturnal preda-
tory birds.

44

Literature Review

APPENDIX II. Publications useful for wildlife studies (concluded).

Title

Frequency

Source

Publications

in which Indexed

Comments

Wildlife Monographs

Irregular

The Wildlife Society
5410 Grosvenor Lane
Bethesda, MD 20814

Current Contents

Each Monograph covers
an individual subject
relating to resource
conservation and wildlife
management.

Wildlife Review

Quarterly

Editorial Office

U.S. Fish and Wildlife

Service
Aylesworth Hall
Colorado State University
Fort Collins, CO 80523

Alerts wildlife biologists of
current worldwide
literature of wildlife
management and
conservation. The
individual issues of the
Wildlife Review are
accumulated into Wildlife
Abstracts.

Wildlife Society
Bulletin

Quarterly

The Wildlife Society
5410 Grosvenor Lane
Bethesda, MD 20814

Biological Abstracts

Covers all aspects of
management, law
enforcement, education,
economics, administration,
philosophy, contemporary
problems, and other
topics related to wildlife.

Wilson Bulletin

Quarterly

Wilson Ornithological

Society
Department of Ornithology
Royal Ontario Museum
1000 Queen's Park
Toronto, Ont. M552C6

Biological Abstracts
Current Contents
Science Citation Index
Zoological Record

Articles based on
research or field
observations throughout
the world.

Literature Review

45

APPENDIX III. Other data bases useful for wildlife research.

Title

Source

Emphases

BIO-STORET

Environmental Protection Agency

Zooplankton, macroinvertebrates,
some vertebrates.

Data base of animal and plant
species of concern to Office of
Endangered Species

Brookhaven National Laboratory

Review and candidate species for
endangered list and those listed as
proposed threatened or endangered

Information system for regional
energy-related assessment and
planning (GEOECOLOGY)

Environmental Sciences Division Oak
Ridge National Laboratory

Wildlife data — includes breeding bird
information, endangered and mammal
range maps and distribution data.

Integrated Habitat Inventory and
Classification System (IHICS)

Bureau of Land Management

Site-specific wildlife species
information — amphibians, birds,
reptiles, mammals.

National Data Base (RPA)

Forest Service

Birds, mammals, reptiles, amphibians,
fish, invertebrates

Raptor Management Information
System (RMIS)

Bureau of Land Management

Management and impact information
on all nocturnal and diurnal raptors
worldwide. Provides for retrieval of
annotated bibliographies of published
and unpublished information
concerning raptors.

Sensitive Wildlife Information System
(SWIS)

U.S. Army Corps of Engineers

Biology and distribution of
approximately 100 selected mammal,
bird, reptile, amphibian, fish, and
invertebrate species. Emphasizes
Federal "endangered" or "threatened"
species.

Species Data Bases

Migratory Bird and Habitat Research
Laboratory (MBHRL)

Avian Information

Species Data Bases

National Coastal Ecosystems Team

Aquatic species — fish, molluscs,
aquatic crustaceans

Species Data Bases

Office of Migratory Bird Management
(MBMO)

Avian Information

Species Systems

Northern California and Pacific
Northwest Coastal Characterization
species systems

Vertebrate species and selected
invertebrates

Terrestrial Species Database

Western Energy Land Use Team
(WELUT)

Amphibians, reptiles, birds, mammals

Western Sierra Wildlife/Habitat
Relationships Program (WHR)

Forest Service Region 5

Birds, mammals, reptiles, amphibians

WILD RAM

Forest. Service Region 4

Birds, mammals, reptiles, amphibians,
fish

Wildlife-Habitat Relationships Data
Base (WILDHAB)

Forest Service Region 6

Birds, mammals, reptiles, amphibians,
fish

Wildlife Management Information
System (WMIS)

Forest Service Region 9

Birds, mammals, reptiles, amphibians,
fish

WILD01

Forest Service Region 1

Birds, mammals, reptiles, amphibians,
fish

46

Literature Review

APPENDIX IV. Museums useful for wildlife inventory research.

American Museum of Natural History
79th Street and Central Park West
New York, NY 10024
212-873-1300

California Academy of Science

Golden Gate Park

San Francisco, CA 941 18

415-221-5100

Carnegie Museum of Natural History
4400 Forbes Avenue
Pittsburgh, PA 15213
412-622-3131

Field Museum of Natural History
Roosevelt Road at Lake Shore Drive
Chicago, IL 60605
312-922-9410

Los Angeles County Museum of Natural History
900 Exposition Boulevard
Los Angeles, CA 90007
213-744-3414

Museum of Vertebrate Zoology
University of California
2593 Life Sciences Building
Berkeley, CA 94720
415-642-3567

Museum of Zoology
University of Michigan
1109 Washtenau
Ann Arbor, MI 48109
313-764-0476

National Museum of Natural Sciences
Victoria Memorial Building
McLeod and Metcalfe Streets
Ottawa, Ontario, K1A OMB
CANADA

U.S. National Museum of Natural History

Smithsonian Institute

10th Street and Constitution Ave. NW

Washington, DC 20560

202-357-2664

Literature Review

47

HABITAT
MAPPING

Richard M. Kerr

Editor's Note: Mapping and classifying land areas
into distinct habitat types is a necessary part of
most wildlife activities. This is particularly true of
wildlife inventories, but it is also generally neces-
sary for management, monitoring and research.
Although the principles of habitat mapping and
classification have not changed since Leopold
(1933) wrote about them, many new tools are now
available. This chapter covers the principles and
provides a guide to the new techniques available.

U.S. Bureau of Land Management
Las Cruces District Office
Las Cruces, NM 88004

"Facts concerning game distribution, behavior, his-
tory, and management can often be accumulated on
maps or tables to better advantage than in notes."

— Aldo Leopold, Game Management

INTRODUCTION

Aldo Leopold noted that facts necessary for
wildlife management can better be recorded as maps
and tables rather than notes because notes tend to
be unorganized and difficult to analyze. He went
further to describe type mapping and the sorts of
wildlife information that could be placed on base
maps (GLO plats, U.S. Geological Survey quads, or
aerial photography; Leopold 1933).

Mapping is done for several purposes or uses in
wildlife habitat management:

• to show geographic locations of wildlife habitat
types;

• to show relationships of types to other types;

• to show community ( types of habitat ) intersper-
sion;

• to quantify types of wildlife habitat;

• to overlay wildlife habitat types with other re-
source inventories; and

• to provide geographic locators in which to re-
cord site-specific animal occurrence and use.

Mapping various types of habitat by drawing
boundaries around these types not only provides
data for inventory and analysis, but also provides the
habitat manager with information for monitoring
management direction or problems. With habitat
maps, the biologist can identify and locate the prob-
lem, quantify the size of the problem and provide,
in some instances, certain types of baseline data.

Most wildlife data bases or inventory systems
used before 1970 described the life history of a spe-
cies and, as part of that discussion, related the kinds
of habitat that the species used. This is an impracti-
cal organization of data for a land manager because
as many as 400 vertebrate species may occur in a
U.S. Bureau of Land Management (BLM) District.
Mapping the habitat individually for each one of

Habitat Mapping

49

these species would be an impossible job. Ranges
would have to be delineated by observation or cap-
ture, and this would be difficult for 1 species and
impossible for 400.

A system that delineates, measures, and locates
on a map different types of habitat and describes the
occurrence of animal species in each type is more
useful to the land manager. Such a system can allow
sampling of habitats and extrapolation of data to
similar habitats.

In addition, land managers must have a system
that allows them to record site-specific information
(e.g., 200 elk [Cervus elaphus] winter in this area).
Also, special features in the landscape like cliffs,
caves, and springs need to be recorded in the
system.

CLASSIFICATION SYSTEMS FOR MAPPING

Mapping from the Top Down: Continental
or Regional Classifications

In the late 1890s, Clinton Hart Merriam pub-
lished "Life Zones of the San Francisco Mountains."
This elevational zone classification was one of the
first ecosystem classifications to be used by ecolo-
gists in North America (see Figure 1) and is still
useful in ecological analyses (Brown 1982). The
early ideas of Clements (1928) and, later, Kuchler
(1964), tended to locate regional ecosystems de-
scribed by dominant natural potential vegetation.

The recent efforts of Brown et al. ( 1977), al-
though at a larger scale (1:1,000,000), are still re-
gional, geographic vegetation types based on natural
potential vegetation (Figure 2).

N

METERS
3965

Alpine Tundra
Timberline

ARTIC- ALPINE

HUDSONIAN

Interior
Chaparral

3355

2745

2135

-1525

-915

Figure 1. Merriam's Life Zones, although helpful as a general characterization, cannot be used as an inven-
tory because of individual site variations within elevational zones (Merriam 1890).

50

Habitat Mapping

/- \

BIOTIC
COMMUNITIES

OF THE
SOUTHWEST

TUNDRA FORMATION
H ai i-ini TUNOnAS
FOREST FORMATION
i^B PtTRAN '.ntiAi ciNI CONIIIH f Olll ST
■H SIEHHAN SUDAll'INt CONIFER FOREST
9BK PETHAN MONTANf CONIFER FORES1
HI SlLHRAN MONTANf CONIFER FORES1
■H S1NALOAN DECIDUOUS FOHEST
WOODLAND FORMATION
ESS GREAT BASIN CONNER WOODLAND
rSBS\ MAOntAN EVERGREEN WOODLAND
■U CALIFORNIAN hVLHGRLEN WOODLAND

SCRUB FORMATION

■■ GREAT UASIN MONTANF SCRIM

131 t CALIFORNIAN CHAPARRAL

IM:: CALIFORNIAN COAS1ALSCRUB

1JS53 INTERIOR CHAPARRAL

JSS2 SINALOAN TMOHNSCRUD

GRASSLAND FORMATION

■■ SUBALPINF GRASSLAND

■ PL A
PL A

^■BR SFMIOFSIHI (.HASSLAND

DESERTSCRUB FORMATION
ISM- GREAT BASIN UESEHTSCRU8

111.1- MOHAVF DlblMTSCHUB

■iU.tJ CHtHUAHUAN DLStRISCHUU
SONONAN VI SIR I SCHUU

I5«-H Uv«<>t Colorado fii*«i Subdivision

[TSfcJj, A../on.i Upland SubdlviMon

KHSS PUtn*. of Sonora Subdivision

nSiJl Ccot.jl Goll Coo-.! Subdivision

KH3H Vi/catno Subdivision

(fSTfifi Maodaleno S«bd*vi%Mv>

Figure 2. A regional classification map called Biotic Communities of the Southwest (Brown et al. 1977),
although at a larger scale than most regional maps, is still too small for field management information.
The State of Arizona is shown as an example.

Robert Bailey's "Ecoregions of the United States"
(1976) does not predominantly use vegetation data
as most do but uses a variety of biotic and abiotic
factors to develop a single geographic classification
at a scale of 1:7,500,000. This again could be consid-
ered a regional ecosystem map or classification
(Figure 3).

The BLM has taken AW. Kuchler's (1975) map
of Potential Natural Vegetation (Figure 4) and re-
gionalized it by using physiographic region bounda-
ries based on soils, vegetation, and regional
topography. These regionalized potential vegetation
types are used as ecosystem boundaries within
which to aggregate, analyze, and extrapolate informa-
tion or data.

The trouble with most of these delineations and
classifications that start mapping from the top down
(from regional or continental ecosystems) is that the
scale is extremely small for use by field biologists,
and mapping is based on potential vegetation and
not on the existing vegetation. Consequently, accu-
racy in the field is very limited. The major benefit
then of these preclassified geographic systems is to
use their boundaries for aggregating data for limiting
extrapolation of information.

Mapping from the Ground Up: Integrated
and Component Classifications

The classifications of Merriam (1890), Kuchler
(1964), Brown et al. (1977), and Bailey (1976)

Habitat Mapping

51

'*\7~~-7s' 'tif /v \'\

^^^xvV

V\ ^of"

/d\\ "

44 dp=

*A

^9k

r\ WARM

1/ /r^

"VA CONTINENTAL

*0» ^

3

rO c

T'c

vM Z^—

^C:

«> ^

\s

/v^£]

>j

/ \^sX — "~-~-^*

> ' \ -

-y^&t>v

IA

|/7_jAc/iC

^ -o
33

s>7

o 1/

>
1 ^

><ff\

^^ O V- ■ 1 /I r"

1 m |

l^"\ ./ A

-A ^<C

// $\ • P-^

y

yk ^ rn S

s^y

<

j\ y^

Highlands

A/i\

S

^*"^,

^r^

\7_&M- mountains

.'. jP - plutcau

O^/^A/, RAINFOREST

i — — | A - altiplano

0

400

800 MILES "^T^vJ

L

1

Figure 3.

terns.

Ecoregions of the United States are a classification that represents a hierarchy of regional ecosys-

might be considered to be integrated because they
are geographically specific, and their boundaries
encompass all habitat components within the
mapped ecosystem. Field-level classifications that use
all habitat components for delineating mapping units
are also integrated systems.

Another type of classification system has been
constructed where each component (vegetation,
landform, soil, climate, etc.) is mapped individually
and independently from local to continental levels,
but is not "integrated" into one mapping unit to
form an ecosystem. This is called a hierarchical com-

52

Habitat Mapping

I

Figure 4. AW. Kuchler's Potential Natural Vegetation is further regionalized by using physiographic region
boundaries.

Habitat Mapping

53

ponent classification system. Hence there are inte-
grated and component classifications.

Since integrated classifications are maps or de-
pictions of geographically located ecosystems, they
are best for wildlife habitat mapping. This is because
animals respond to all components within an area,
not just one. In addition, they react to the size and
location of these small ecosystems.

Within the last decade, integrated concepts have
been so mixed with hierarchical component con-
cepts that it is now difficult to determine a partially
integrated system from the hierarchical component
system. A good example of this is "A Component
Land Classification for the United States" (Driscoll et
al. 1983). This system, authored by five government
agencies, is meant to be a hierarchical component
classification system; yet two components, soil and
vegetation, are integrated through habitat types and
range sites. Other component hierarchies are not yet
developed. What is called a component system now
appears to be an integrated system.

In the field, less time should be spent on re-
gional classification systems and more time on local
inventories, as far as wildlife habitat data are con-
cerned. A classification can be built from a good
inventory, but an inventory cannot be built from a
classification.

MAPPING TOOLS

To inventory wildlife habitat, a biologist must
construct a map to quantify and geographically lo-
cate habitat. Describing animal communities found
within plant communities by composition is merely
a characterization of different kinds of habitat. An
inventory of wildlife habitat has not been completed
until a reasonably accurate map of the habitat has
been constructed. Similarly, monitoring studies initi-
ated must have a map to be of continuing value;
anything less is incomplete or unusable.

The map may be an overlay of a base map, a
computer graphic printout, delineators of habitat on
a U.S. Geological Survey quadrangle map, a typed
aerial photo (see Glossary for definitions), a Landsat
image, a typed planimetric map, a plane table map,
or typed orthophoto maps.

A map scale for inventorying and monitoring
should be at the range somewhere between 1/60,000
to 1/2,000, depending on the intensity and purposes
of the study. Maps at a smaller scale than 1 :60,000
may provide some assistance in planning, but are not
large enough to solve actual habitat management
problems.

Some maps have been constructed by using a
scale of 1/2 inch to equal a mile or 1/100,000. These
were commonly used in BLM's Unit Resource Analy-
sis (URA) system. At this scale, these maps can only
give spatial relationships of gross kinds of habitat.
They are not suitable for recording and analyzing
wildlife habitat data.

Aerial Photos

The most commonly used mapping tools are
aerial photos from 1/2,000 to 1/63,831, and 7.5- and
1 5-minute U.S. Geological Survey quadrangles. Ob-
viously, the larger the survey area in acres, the
smaller the scale will be to meet economic criteria.
The easiest and best photo-map combination for
broad inventories over large acreages of land 8,000
to 400,000 ha (20,000 to 1 million or more a.) is
the U.S. Geological Survey orthophoto quadrangle
(1/24,000) (Figure 5) and the 7. 5-minute topo-
graphic quadrangle (Figure 6). The orthophoto quad
is a corrected copy of an aerial photo at 1 :24,000
scale, which allows the habitat typing to be placed
directly on a corrected aerial photo which is the
same area, scale, and control as the base map. Habi-
tat type lines can be traced or plotted from the or-
thophoto quadrangle to the topographic quadrangle
of the same area.

Orthophoto maps (Figure 7) are corrected aer-
ial photos with topographic map information super-
imposed. They thus combine all features of both 7.5-
minute topographic quadrangles and orthophoto
quadrangles. These are ideal for habitat mapping but
are only available for a few areas.

If orthophoto quadrangles are not available at
1/24,000, a 7. 5-minute U.S. Geological Survey quad-
rangle map is used; type lines are transferred from
aerial photos with a plotter or zoom transfer scope.

A complete description of the principles behind
aerial photography can be found in Burr (1976).

The kind of aerial photos used may vary as to
the situation and region. The first habitat inventories
were completed with black and white aerial photos
(Figure 8), and these are still very useful and accept-
able. A recent innovation is the use of color-infrared
film which is particularly useful if one wants to em-
phasize riparian vegetation or seasonally dominant
vegetation. This film produces a brighter shade of
red on the print where more growth is taking place
(Figure 9).

In the Southwest, where vegetation is sparse and
soil highly visible, the use of true color film is desira-
ble. Color often gives better definition between habi-
tat types than black and white photos give (Figure
10).

54

Habitat Mapping

Figure 5. Orthophoto quadrangles like these are excellent corrected photos for habitat mapping.

Habitat Mapping

55

Figure 6. U.S. Geological Survey 7.5-minute topographic quadrangles are the mainstay for base maps for
wildlife habitat inventory.

56

Habitat Mapping

Figure 7. The U.S. Geological Survey 7.5-minute orthophoto map provides a corrected aerial photo with
topographic and other map features.

Habitat Mapping

57

Figure 8. Black and white aerial photo from Piceance Basin, Colorado.

58 Habitat Mapping

Figure 9. Color infrared aerial photo of same area as Figure 8.

Habitat Mapping

59

Figure 10. True color aerial photo of same area as Figure 8.

60

Habitat Mapping

The photo scale recommended for overall in-
ventories on public lands has historically been close
to 1/24,000 because this ties in well with the scale
of U.S. Geological Survey quads, which make excel-
lent base maps for typing vegetation. This medium-
scale photography is normally available from various
government agencies (Appendix I).

Larger-scale photos about 1/2,000 to 1/4,000 are
beneficially used where typing is necessary on highly

valuable habitats such as riparian areas in semiarid
regions. Color-infrared aerial photography for these
areas may be valuable (Figure 11). This large scale is
also extremely useful for habitat monitoring because
individual trees and shrubs can easily be seen as
well as stream channel changes. This large-scale pho-
tography is normally not available from government
agencies and must be obtained by contract. In some
instances, larger scale photography has been used for
monitoring vegetation damage.

Figure 11. Large-scale (1/2,000) color infrared aerial photo of riparian areas along Black Canyon Creek,
Arizona.

Habitat Mapping

61

A complete description of low-level aerial pho-
tography for riparian inventory and monitoring was
given by Cuplin (1981).

In addition to mapping, aerial, oblique, and
ground-based photographs are indispensable as rec-
ords for baseline data and for composition, density,
and structure of habitat. A properly identified photo-
graph may be the best record for a baseline study
of a variety of monitoring studies. The use of a cover
or profile board can show the decreasing density of
perennial grass on a duck-nesting habitat study,
whereas a series of yearly photos from a photo hub
may show the invasion of grass into a shrub type
used for deer winter feeding. "A picture is worth a
thousand words."

Landsat Images

Landsat images are sometimes used to delineate
gross vegetation cover types. Reflected light is re-
corded in four bands and can be used to generate
Landsat images that can be manually interpreted or
the digital reflectance data can be computer-classi-
fied. Pixels of about 0.04 ha (1.1 a.) picture reflected
light and are density-sliced and artificially classified
into false colors for various classifications. Because of
the characteristics of the Landsat reflections, these
scenes do not provide all the benefits of a photo-
graph at medium scales. They show no texture,
shadow, or highlights, but just the color of a particu-
lar classification (Figure 12). Also, the discerning
abilities of the Landsat scanners are not capable of
separating vegetation species or areas that give the
same reflectance. Detailed classifications are not pos-
sible with Landsat images.

A common standard used to describe levels of
classification detail has been a "Land Use and Land
Cover Classification System for Use with Remote
Sensor Data" (Anderson et al. 1976). (Note: The au-
thor gives description titles to only Levels I and II,
leaving the development of Levels III and IV to local
workers [Table 1].)

Information given in the latest Manual of Re-
mote Sensing (Colwell 1983) states "Efforts to clas-
sify Level III categories using Landsat generally result
in overall accuracy figures of less than 70% , a figure
often considered unacceptable." In actuality, vegeta-
tion cover classifications for land management use
require more definition than Level III (Anderson et
al. 1976). Level IV approaches the level necessary
for management use, and it is even less accurate than
Level III, perhaps 50 percent or less. If Level II was
described as shrubland, then the Level III classifica-
tion descriptor might be rabbit brush (Chtysotham-
nus sp.) while the Level IV descriptor might be
rabbit brush-snakeweed {Gutierrezia sp.).

Unlike more easily interpreted photographic
scenes, Landsat scenes are highly dependent on
ground truthing. The more ground truthing, the
more accurate is the final classification scene. The
problem of signature extension has not been solved
for all areas and vegetation types, hence a poorly
ground-truthed scene of vegetation may be highly
inaccurate.

Vegetation signatures (reflection shades) many
times do not remain accurate even over reasonable
distances (3 to 5 mi.); therefore, for accurate classifi-
cations, signatures must be classified for small areas
by using a great deal of ground truthing and super-
vised classifications (Short 1982). Because of this, it
may be more economical to use aerial photos to
lend more accuracy to final products. Also, because
definitions on photos exceed those of pixel classifica-
tions, ground control (section corners, etc.) is better
accomplished with photos than Landsat images.

Although Landsat images may be an efficient
advance in providing a mapping base where there
are no existing base maps (i.e., Alaska or Africa),
most areas in the western U.S. are already accurately
mapped with 7.5-minute U.S. Geological Survey topo-
graphic quadrangles. These maps, constructed
through photo interpretation, in many instances have
some vegetation mapped at a level comparable to
the Level III mapping used by contemporary Landsat
projects. This mapping is much more accurate and
at a larger scale than the Landsat images available
either by standard order or by computer construc-
tion and enhancement. Computer-enhanced Landsat
images therefore must be used with discretion be-
cause they do not provide mapping at the accuracy
and level (or scale) for most BLM management
needs. However, they may supplement or provide an
additional tool for enhancing management data such
as information on snow cover, cloud cover, soils, and
repetitive synoptic views. If aerial photos were to
be totally replaced by Landsat scenes as a basic habi-
tat inventory tool, the land management profession
would have taken a giant step backward.

Sources of Supply

Aerial photos and Landsat images may be ob-
tained from affiliates of the National Cartographic In-
formation Center (Appendix I). Assistance for BLM
employees may be obtained from the BLM Service
Center. For U.S. Geological Survey maps of all kinds,
including orthophoto quads for areas west of the
Mississippi, write to the U.S. Geological Survey, Box
25286, Denver Federal Center, Denver, CO 80225.

62

Habitat Mapping

HIGH DENSITY SAGEBRUSH

LOW DENSITY SAGEBRUSH

SALTBUSH / GREASEWOOD / SAGE

SALTBUSH

ASPEN / OPEN ASPEN

CONIFER

RIPARIAN / AGRICULTURE

BARREN

WATER

Figure 12. An unsupervised Landsat classification.

Habitat Mapping

63

Table 1. Land use and land cover classification system for use with remote sensor data (modified from
Anderson et al. 1976).

Level I1

Level II2

Urban or built-up land

11. Residential3

12. Commercial and services

13. Industrial

14. Transportation, communications, and utilities

15. Industrial and commercial complexes

16. Mixed urban or built-up land

17. Other urban or built-up land

Agricultural land

21. Cropland and pasture

22. Orchards, groves, vineyards, nurseries, and orna-
mental horticultural areas

23. Confined feeding operations

24. Other agricultural land

Rangeland

31. Herbaceous rangeland

32. Shrub and brush rangeland

33. Mixed rangeland

Forest land

41 . Deciduous forest land

42. Evergreen forest land

43. Mixed forest land

Water

51. Streams and canals

52. Lakes

53. Reservoirs

54. Bays and estuaries

Wetland

61 . Forested wetland

62. Nonforested wetland

Barren land

71. Dry salt flats

72. Beaches

73. Sandy areas other than beaches

74. Bare exposed rock

75. Strip mines, quarries, and gravel pits

76. Transitional areas

77. Mixed barren land

Tundra

81. Shrub and brush tundra

82. Herbaceous tundra

83. Bare ground tundra

84. Wet tundra

85. Mixed tundra

Perennial snow or ice

91. Perennial snowfields

92. Glaciers

'Level 1 Typical data characteristics at this level would be found on Landsat types of data.

2Level 2 Typical data characteristics at this level would be found on high-altitude data at 40,000 ft. (12,400 m) or above (less than
1 80,000 scale)
3Numbers and titles refer to standard land use codes and definitions from U.S. Urban Renewal Administration et al. 1965.

THE INTEGRATED MAPPING PROCESS

The traditional integrated mapping approach fits
well with the concept of habitat inventory, which
has five basic steps:

(1) mapping;

(2) stratification and classification;

(3) inventory of habitat-species relationships;

(4) site-specific inventory; and

(5) data storage, retrieval, and analysis.

Mapping

To provide for the aggregation of wildlife habi-
tat data at local levels and up into higher levels of
field classification, a mapping unit is used to collect

64

Habitat Mapping

;

and quantify data. This mapping unit is delineated on
aerial photographs as a homogeneous polygon —
homogeneous as to existing vegetation, landform,
and soil. Although size of the mapping unit has no
required range, it varies between 0.8 ha and perhaps
4,000 ha (2a. and 10,000a.). There may be some
misunderstanding of what the mapping unit is called.
In BLM's Integrated Habitat Inventory and Classifica-
tion System (IHICS), it is called a habitat site (Hamil-
ton et al. 1983). In other systems, it is called a
mapping unit, whereas in the BLM's vegetation in-
ventory, it is called a site writeup area.

Mapping units are first pretyped on aerial photos
(1:24,000) or orthophoto quads. This is done by
examining the photos and drawing boundary lines
around areas homogeneous in terms of existing vege-
tation, landform, and soils (if soil inventory overlays
or ecological site delineations are available). To re-
duce displacement caused by relief, the effective
area or central portion of a photo pair should be
used for interpretation.

Once all areas of the photos are pretyped, the
boundaries or delineations that were put on in the
office are checked in the field by visual observation.
It is also desirable to use some management bounda-
ries to break the mapping unit by forming one of
its boundaries (e.g., a grazing allotment boundary
should not be crossed when delineating mapping
units because many times data need to be accumu-
lated and analyzed according to grazing allotment).
Similarly, Resource Area boundaries should not be
crossed when delineating mapping units. During
a field check, section corners are located and pin-
pricked through the photograph. Identification is
placed on the back of the photo for control. If ortho-
photo quads are used, this is not necessary; each
mapping unit is assigned a unique designation within
the inventory area.

Certain habitat data are recorded for each map-
ping unit while checking boundaries in the field.
Landform, site elevation, dominant and subdominant
vegetation (by visual observation or vegetation data),
aspect (by cardinal direction), and slope data are
recorded and entered on the appropriate form.
Acreage should be computed and recorded as soon
as the unit boundary is mapped and becomes final.

Once mapping units have been checked and are
considered final, they are transferred to a 1:24,000
U.S. Geological Survey quadrangle map if they are
not already on an orthophoto map. This is done by a
zoom transfer scope or other type of plotter and by
using the located section corners for control.

Where ecological sites (potential vegetation)
have been delineated on the ground and units of
existing vegetation have been delineated on maps,
the same units and numbers can be used for habitat

inventory data base. If this option is available, the
work described in the preceding paragraphs is not
necessary except when collecting the described
habitat data.

Stratification and Classification

After all mapping units have been given an iden-
tification number and initial habitat information has
been gathered and recorded, the mapping units are
placed in strata which will ultimately be a field level
classification. To determine the number and types
of strata, the biologist in charge will need to make a
field review, noting which types of habitat should
be designated as habitat strata. The partially com-
pleted habitat information forms can be used for
stratifying by reviewing the habitat site information
on them, particularly the dominant and subdominant
vegetation, landform, or other factors based on the
needs of the inventory situation. Once criteria for
grouping strata are determined, classifications are
made by assigning each mapping unit to the stratum
that applies.

The BLM IHICS can take input forms showing
habitat data for mapping units and automatically sort
strata by selected criteria. For example, if all map-
ping units located on a mesa where ponderosa pine
{Pinus ponderosa) is dominant and Columbia need-
legrass (Stipa columbiana) is subdominant are se-
lected as a stratum, the computer will easily sort and
form the stratum. Each classification (stratum) is
assigned a number and name. Each mapping unit
then has its classification number recorded on the
appropriate form. New strata may be devised at any
time by using different criteria and reassembling
the mapping units.

Although existing vegetation and landform are
important, criteria for forming strata may vary. If
a particularly important animal is present, a stratum
of habitat may be made for it by using an important
characteristic of its habitat. For instance, if Gila
monsters (Heloderma suspectum ) are found in a
creosote-bursage habitat, but only in areas that con-
tain abundant ground litter, and if abundant litter
is used as a criterion within this cover type, a stra-
tum is constructed that will represent potential Gila
monster habitat.

Sufficient habitat data (described below) are
collected for each map unit. Strata can be con-
structed by using habitat data in manual or auto-
matic data processing (ADP) operations (i.e., sort all
map units that are ponderosa pine and pine needle-
grass [Stipa pinetorum] at 7,000 to 9,000 ft [2,121
to 2,727 m] elevation). Consequently, biologists have
the ability to construct strata for sampling based on
a variety of criteria once map units are delineated
and initial habitat information is recorded.

Habitat Mapping

65

The mapping unit is the lowest level of ecosys-
tem classification systems. The stratum level is the
aggregation of mapping units having similar vegeta-
tion, landform, soil, etc. The two lower levels (map
unit, stratum) provide a means of categorizing field
inventory data to facilitate data analysis for land-
use planning and decisionmaking. They provide a
long-term accumulation and aggregation of new data
from the mapping unit level to upper levels of conti-
nental or regional systems.

Inventory of Habitat-Species Relationships

Once delineation and classification of mapping
units have been made, the next step is to collect
information that will reveal species-habitat relation-
ships. For each stratum, representative mapping units
are selected for sampling vegetation and wildlife.

Vegetation composition and structural data
should be sampled in the same habitat sites where
the animal sampling occurs. If these data are avail-
able from other vegetation inventories, then sam-
pling is not needed.

A hypothetical list prepared through literature
review by using upper-level classifications (Kuchler
1964; Brown et al. 1977) can be used as a basis for
sampling animal occurrence. A sampling plan is de-
veloped that verifies the occurrence of animal spe-
cies listed hypothetically. If a complete wildlife
inventory is conducted, each animal class is invento-
ried for one or more seasons, depending on its life
history. Generally, inventories are conducted for
mammals, birds, reptiles, and amphibians. When time
and money are limited, only those classes deter-
mined important in the management situation are
inventoried. Species thought to occur in the selected
habitat sites are listed in the data base as hypotheti-
cal. Unique or important animals expected to occur
in the inventory area should be verified by using
appropriate inventory methods.

Sampling of animals in a stratum will relate oc-
currence of animals to certain canopies or other
habitat structure, or other special features within the
habitat. One way to do this is to use inferential sta-
tistics and compare occurrence of certain animals
with occurrence of a particular layer of habitat or
special feature. This is accomplished best if the biol-
ogist completes the same vegetation (habitat) inven-
tory at each animal sampling location. After sufficient
numbers of samples have been taken, a statistical
comparison can be made. Vegetation or habitat sam-
pling methods may be developed by region or area
of inventory, considering those things in the habitat
that appear to be important in that particular region.

A simple list of animals occurring in an area is
helpful, but the land manager needs additional infor-

mation. This includes season of use; function of the
habitat for a particular species; determination of
whether the habitat is crucial for the species; and
the relative abundance of the species, compared
with other habitats.

If stratification and sampling have revealed that
two or more strata have the same animal community
and uses, then these may be combined into a single
habitat stratum.

If the purpose of mapping is to monitor a wild-
life habitat problem, then the same procedures of
mapping apply except that vegetation sampling is
done by a system that monitors those aspects of the
habitat that are being managed or affected.

Because the BLM IHICS uses inventories of ex-
isting vegetation, the potential vegetation situation is
normally portrayed by using habitat types (plant
associations) for forests and range or ecological sites
for rangelands. Comparing the existing vegetation
with this potential vegetation will give possibilities
for vegetation management, which would direct
animal use or animal community progression toward
a desired level. This potential determination may
be available for broad areas or can be determined
individually for small management areas (such as
deer feeding areas).

Much money and time can be spent acquiring
potential information on the total area of rangelands
within a BLM administrative unit jurisdictional
boundary, when in practice, only a small amount of
this information will be used in habitat management
decisions. Biologists, therefore, cannot afford to
spend their time delineating potential for all areas.
Biologists should restrict their efforts toward deter-
mining vegetation potential only in areas where in-
formation is not available from the U.S. Soil
Conservation Service (SCS) range site system or the
forest habitat type system, and the information is
necessary for specific immediate wildlife habitat
management analysis.

It is important to understand that vegetation
progression under commercial forest management
changes structure rapidly, whereas composition and
structural changes under range management situa-
tions take long periods and the effects are much
more subtle and generally less predictable.

Site-Specific Inventories

Mapping units provide the mechanism for re-
cording site-specific data because each has a unique
number which locates it geographically. For those
species important to the management of public lands
and for which information cannot be safely extrapo-
lated, site-specific data must be collected. These

66

Habitat Mapping

species are generally highly mobile species (such as
raptors and wild ungulates) whose presence is not
well correlated with just a habitat or vegetation type
(e.g., because all ponderosa pine stands do not con-
tain deer, we must find ones that do and record by a
mapping unit number).

The form for recording information for mapping
units and site-specific data should also allow for con-
tinuing entry of animal occurrence data as they are
gathered, either in complete inventories or oppor-
tune sightings.

Sometimes features of the habitat that are too
small to map (springs, walls, power poles, cliffs, etc.)
have a major effect on whether certain species of
wildlife occur in an area. The effect may be either
negative or positive. These are called special habitat
features, and a complete habitat inventory should
record them. Time and money will not allow a total
inventory of all special habitat features. The biolo-
gists must choose those features that are important
to wildlife in their inventory areas and inventory
them. (For instance, all special habitat features that
are water-related should be inventoried.) Species
that are benefitted or discouraged by the special hab-
itat feature are listed. This gives the biologists the
second form of site-specific data for their data bases.

Data Analysis, Storage, and Retrieval

In the next stage, field forms for animal use,
occurrence, relative abundance, and crucial habitat
are assembled and, depending on the system, sum-
marized on additional forms. At this stage, the biolo-
gist may need to modify the original strata
designations based on the additional information
collected.

At this stage, data is normally entered into some
sort of computer system. In the BLM IHICS, the sum-
mary forms are entered into a computer data base.
Regardless of the system used, forms should be
checked carefully prior to key-entry.

In the BLM's IHICS, summary data are edited
automatically and then used to create data bases that
are available in the field on either a time-sharing
computer system or on a field office microcomputer.
Separate data bases are maintained for habitat sites
(mapping units), standard habitat sites (mapping
strata), and special habitat features for each inven-
tory area.

The use of ADP has generally proceeded in two
reasonable directions as far as wildlife habitat inven-
tory and monitoring are concerned. Some agencies
have emphasized digitized graphic displays (maps) of
types of habitat. This was done by using systems
such as the Map Overlay and Statistical System
(MOSS) to construct polygons (habitat mapping
units from digitized data derived from aerial photos).
For large study areas, 400,000 ha ( 1 million a. ) or
more at 1/24,000 accuracy, digitizing needed for
constructing computer-generated maps is expensive
and time-consuming, but its use is increasing at the
field level.

Other agencies emphasized the measurement of
biological components within strata and mapping
units. These data are normally input as tables and
other narratives whereas the maps that have been
used with them have been manually constructed.
The BLM IHICS is a good example of a system that
sorts several dozen habitat characteristics.

Both types of systems are valuable where large
amounts of data are handled, and the automation
entices field offices to use the data repeatedly for a
variety of uses, thereby obtaining greater benefit.

The present move toward balanced emphasis on
computer mapping and definitive biological data
will make the data even more usable and valuable.
This type of balanced emphasis could provide a very
usable Geographic Information System (GIS). Pro-
cesses such as the digital elevation model (DEM) can
be included or used in conjunction with GIS thereby
enhancing the utility of both data bases.

DISCUSSION

The overall purpose of mapping is to organize
wildlife habitat data to manage wildlife habitat and to
mitigate or eliminate impacts of other uses. The hab-
itat map is the basic tool for the land-management
biologist. Geographic locations, quantification of
habitat, and specific habitat characterization cannot
be accomplished without a map. In addition, maps
are necessary to compare wildlife habitat with other
resources for multiple-use decisions. Even though
contemporary map-making methods may change
from a manually-constructed paper map to a com-
puter-generated graphic, a map is only as accurate as
its human constructor.

Habitat Mapping

67

LITERATURE CITED

ANDERSON, J.A., E.E. HARDY, J.T. ROACH, and RE. WIT-
MER. 1976. A land use and land cover classification
system for use with remote sensor data. U.S. Dep.
Inter., Geological Survey, Professional Paper 964.
28pp.

BAILEY, RG. 1976. Ecoregions of the United States. U.S.
Dep. Agric, Forest Service. Ogden, UT 84401. 77pp.

BROWN, D.E., ed. 1982. Desert plants, biotic communities
of the American Southwest, United States and Mexico.
Vol. 4, Numbers 1-4, Special issue. University of Ari-
zona, Tucson. 342pp.

, C.H. LOWE, and C.P. PASE. 1977. Biotic communi-
ties of the Southwest. U.S. Dep. Agric, Gen. Tech.
Rep. RM-41. 342pp.

BURR, R.D. 1976. The use of aerial photographs. Technical
Note 287. U.S. Dep. Inter., Bur. Land Manage. BLM
Service Center, Denver, CO 80225. 34pp.

CLEMENTS, F.E. 1928. Plant succession and indicators.
The H.W. Wilson Co. New York, NY. 453pp.

COLWELL, R.N., ed. 1983- Manual of remote sensing. 2nd
Ed., Vol. II. American Society of Photogrammetry.
Falls Church, VA. 2440pp.

CUPLLN, P. 1 98 1 . The use of large-scale color infrared
photography for stream habitat and riparian vegeta-
tion inventory. Technical Note 325, U.S. Dep. Inter.,
Bur. Land Manage. BLM Service Center, Denver, CO
80225. 7pp.

DRISCOLL, R.S., D.L. MERKLE, J.S. HAGIHARA, and D.L.
RADLOFF. 1983 A component land classification for
the United States: Status report. Technical Note 360.

U.S. Dep. Inter., Bur. Land Manage. BLM Service Cen-
ter, Denver, CO 80225. 30pp.

HAMILTON, C.K., R.M. KERR, and LA. PETERSON. 1983.
IHICS, the Bureau of Land Management's habitat in-
ventory system. Presented at National Workshop
on Computer Uses in Fish and Wildlife Programs — A
State of the Art Review. Virginia Polytechnic Institute
and State University, Blacksburg, VA. 1 3pp.

KUCHLER, AW. 1964. Potential natural vegetation of the
conterminous United States, map and manual. Special
Publication 36, Am. Geographic Soc. New York, NY.
122pp.

. 1975. Potential natural vegetation of the contermi-
nous United States, revised map. Am. Geographic
Soc. New York, NY.

LEOPOLD, A. 1947. Game management. Charles Scribner's
Sons. New York, NY. 481pp.

MERRIAM, C.H. 1890. Results of a biological survey of the
San Francisco Mountain region and desert of the
Little Colorado in Arizona. North Am. Fauna No. 3-
U.S. Dep. Agric, Washington, DC (Reprinted in se-
lected works of Clinton Hart Merriam. Natural Sci-
ences in America, K.B. Sterling, ed.) Arno Press. New
York, NY 1974.

SHORT, N.M. 1982. The Landsat tutorial workbook, basics
of satellite remote sensing. NASA Reference Publica-
tion 1078. National Aeronautics and Space Adminis-
tration. Washington, DC.

U.S. URBAN RENEWAL ADMINISTRATION, HOUSING
AND HOME FINANCE AGENCY, and BUREAU OF
PUBLIC ROADS. 1965. Standard land use coding man-
ual, a standard system for identifying and coding
land use activities. Washington, DC. 1 1 1pp.

68

Habitat Mapping

APPENDIX I. Sources of Supply

National Affiliate Offices

State Affiliate Offices (cont.)

National Cartographic Information Center
U.S. Geological Survey
507 National Center
Reston, VA 22092.

3. Montana

Montana Bureau of Mines and Geology

Montana Tech

Main Hall, Room 200

Butte, MT 59701

Eastern Mapping Center

National Cartographic Information Center

U.S. Geological Survey

536 National Center

Reston, VA 22092.

Nebraska

Director and State Geologist
Conservation and Survey Division
University of Nebraska
Lincoln, NE 68508

Mid-Continent Mapping Center

National Cartographic Information Center

U.S. Geological Survey

1400 Independence Road

RoUa, MO 65401

5. New Mexico

University of New Mexico
Technology Applications Center
2500 Central Avenue, S.E.
Albuquerque, NM 87131

Rocky Mountain Mapping Center
National Cartographic Information Center
U.S. Geological Survey
Stop 504, Box 25046, Federal Center
Denver, CO 80225.

6. Nevada

Nevada Bureau of Mines and Geology
University of Nevada, Reno
Reno, NV 89557-0088

Western Mapping Center

National Cartographic Information Center

U.S. Geological Survey

345 Middlefield Road

Menlo Park, CA 94025.

Oregon

Oregon State Library
Public Services
Salem, OR 97310

Alaska Office National Cartographic Information Center
U.S. Geological Survey
Skyline Building
218 E Street
Anchorage, AK 99501.

Utah

Utah Geological and Mineral Survey

606 Black Hawk Way

Research Park

Salt Lake City, UT 84108-1280

7. EROS Data Center
U.S. Geological Survey
Sioux Falls, SD 57198

State Affiliate Offices

Washington

(User Services)

Washington State Library
Information Services Division
Olympia, WA 98504

1. Arizona

Arizona State Land Department
Information Resources Division
1624 West Adams, Room 302
Phoenix, AZ 85007

(Data Acquisitions)

Division of Management Services
Photos, Maps, and Reports Section
QW-21, Public Lands Building
Olympia, WA 98504

2 Idaho

Idaho State Historical Library
325 West State
Boise, ID 83702

1 0. Wyoming
State Engineer
Barrett Building
Cheyenne, WY 82002

Habitat Mapping

69

II MAJOR HABITATS

5 Forests

6 Rangelands

7 Deserts

8 Tundra

9 Riparian Habitats

10 Marshes

1 1 Streams

12 Lakes

FORESTS

Jack Ward Thomas

Pacific Northwest Forest and Range Experiment Station
U.S. Forest Service
LaGrande, OR 97850

Jared Verner

Editor's Note: This is the first in a series of chapters
dealing with habitats — other upland habitats in-
clude deserts, rangelands, and tundra These are not
clearly demarked ecosystems; they are important
basic habitat types that overlap geographically and
descriptively. As such, they require different em-
phases when planning inventory or monitoring
studies of wildlife resources.

While more emphasis has been placed on inventory
and monitoring of forested habitats than on others,
there are no generally accepted or standardized
techniques. Biologists still need to design studies
based on local needs, while taking advantage of the
experiences of others.

Pacific Southwest Forest and Range Experiment Station
U.S. Forest Service
Fresno, CA 93710

"Forest ecosystems are far too complicated for us to
ever fully understand and precisely predict future
conditions. Inventories of current conditions, and
technologies for analyzing resource potentials and
conducting management actions, are often incom-
plete or less than desired. . . . None of these are ex-
cuses for lack of action. They are simply the
circumstances under which resource managers
operate."

— Hal Salwasser

INTRODUCTION

The question to be answered in this chapter is,
what are the important attributes of forest habitat to
be measured if inventorying or monitoring of wild-
life habitat is the goal? The answer will vary depend-
ing on the questioner's interests. For example, if
the aim is to obtain a general idea of the habitat and
which species groups are (or could be) present, it
may be sufficient to determine the extent of each
forest type and its attendant successional stages or
structural conditions and the presence or absence of
animal groups of interest. This could be accom-
plished either through observation or from informa-
tion relating vertebrate species to forest plant
communities and attendant seres or structural condi-
tions (see Thomas 1979b or Verner and Boss 1980
as examples).

If, however, one desires sufficient information to
predict the response of a single species to habitat
alterations, more specific habitat attributes must be
considered. For example, if Rocky Mountain elk
(Cervus elaphus nelsoni) in the heavily forested
areas of the intermountain West were of particular
interest, information should be collected on stand
size, stand height, canopy closure, density of roads
open to vehicular traffic, cover/forage ratios, distance
to edges between cover and forage, and, in some
instances, quantity and quality of forage. In short, the
more specific the question, the more detailed the
habitat information required.

Jenkins ( 1977:43) quoted Frank Egler as saying,
"Ecosystems are not only more complex than we
think, they are more complex than we can think."
He then said,

". . . to deal with this literally incomprehensible
complexity, for whatever purpose, none of us
have any choice but to take a strenuously re-
ductionist approach which means making

hard choices about which aspects to emphasize

Forests

73

and which to disregard. No single set of
choices can ever satisfy everyone or every
purpose, so there will never be a single classifi-
cation system ... it is both necessary and legiti-
mate to tailor our efforts to our own
purposes . . ."

First, we need a clear idea of what to inventory
or monitor and why. There are no pat answers to
the questions — what should be monitored? What
approach should be used? How should the inventory
and monitoring be accomplished? The answers de-
pend on the legal requirements, management needs,
objectives of the wildlife management, resources,
and skills of persons available to do the work. There
is not and should not be a list of species, guilds, or
life forms that should be monitored on each forested
area. There is not and should not be specification of
a vegetative classification system to be used. There is
not and should not be a prescribed list of the right
techniques to be used to inventory and monitor
either habitats or associated wildlife species.

It sounds trite, but the only realistic answer
it all depends. This entire area of activity is so new
and so dynamic, in terms of management need and
technology available, that the best any reviewer can
do is to array the existing information and let the
user choose, considering the particular objectives,
needs, and resources available.

One needs to settle on a forest habitat classifica-
tion system to structure the effort. This is necessary
to provide —

• a basis for stratifying forest habitats for
sampling;

• a framework for communication among various
disciplines and the public;

• a mechanism for aggregating information from
local areas into district, regional, and even
larger assessments.

These first steps of deciding what and why to
inventory and of selecting an appropriate habitat
classification system to guide that effort are critical
and must be carefully considered. Decisions should
be predicated upon these questions:

( 1 ) How much will it cost and can it be afforded?

(2) What steps can be taken to ensure the useful-
ness of a data base that can be developed at
a realistic cost?

Intensive and frequent sampling of forest habitat
is expensive. Sampling that includes statistically valid
estimates of animal numbers can be prohibitively

expensive (Verner 1983). In many situations, then,
only relatively simple, nonintensive surveys are
affordable.

Further, it is likely that wildlife biologists will
continue, at least in the foreseeable future, to oc-
cupy a distinctly minority position in the major land
management agencies. If they want to be effective,
therefore, they will recognize the advantages of mak-
ing their data bases compatible with those developed
by the disciplines that dominate the agency — e.g.,
foresters and planners. Among the advantages are —

• increased probability that forest wildlife and
wildlife habitat can be more readily considered
in land-use planning and management decisions;

• better communication among disciplines and
with the public;

• possible reduction in the cost of gathering
needed information about wildlife habitat.

CLASSIFICATION SYSTEMS

The habitat classification system used in inven-
tory and monitoring of forest wildlife and their habi-
tats is usually selected on the basis of the classi-
fication system being used by forest inventory or
timber managers, usually both. Forest habitat classifi-
cation is complex and little agreement exists on a
standard approach (Bailey 1977). The confused sta-
tus of the situation is apparent from a perusal of a
1977 symposium on "Classification, Inventory, and
Analysis of Fish and Wildlife Habitat" (Marmelstein
1977).

Several criteria have been used to develop sys-
tems of ecosystem classification, including biological
(Brown et al. 1980; Garrison et al. 1977) and physi-
cal (Godfrey 1977). The most commonly used stand-
ard classification system for western forests based
primarily on vegetation was developed by Dauben-
mire (1968). With some modification, Daubenmire's
system has been extended to half or more of the
forested land of the western U.S. The primary classifi-
cation systems in use in the 1 1 western States and
Alaska are shown in Table 1.

Bailey (1982:16) noted that—

"Past wildlife studies and inventories have pro-
ceeded without the benefit of an integrated
system. Biologists often had to depend on any
available, sometimes inadequate, information or
devise their own classification, usually a map
featuring forest cover. Many investigators gath-
ered disconnected bits of descriptive informa-
tion on habitat without a classification frame-
work to give them meaning. Without such a

74

Forests

framework, it was very difficult and sometimes
impossible to integrate wildlife information
with other information for evaluating trade-offs
or interactions within the fish and wildlife re-
source and between it and other resources. As
of 1970, there was no national approach to
integrating wildlife information ..."

What was true in 1970 is still true today. But
improvements have been made and efforts are un-
derway to develop an acceptable National approach
(Hoekstra et al. 1983).

What can be done when it makes sense to use a
finer or coarser breakdown of habitats than fits the

Table 1. Major classification systems for forests of the western States.

System

Description

Source

Habitat types

Daubenmire 1968;
Pfister 1977

Land systems

Involves delineation,
description, and evaluation of
relatively homogeneous units
of land at local or regional
scale. Hierarchical
components exert the most
control at the top.

Wertz and Arnold 1972; Bailey
1976, 1978; Crowley 1967

ECOCLASS

A method of integrating
several systems to identify
homogeneous land units.
Developed by the U.S. Forest
Service.

Cortiss 1974; Buttery 1978

ECOSYM

Links classification to land
management needs.

Henderson et al. 1978

National Site (Land)
Classification System

Like ECOCLASS, a component
system that involves
vegetation, soil, landform, and
water. Developed jointly by the
U.S. Forest Service, Bureau of
Land Management, Fish and
Wildlife Service, Geological
Survey, and Soil Conservation
Service.

Driscoll et al. 1978

Integrated Habitat Inventory
Classification System

A six-level system for
organizing species occurrence
data from plant communities
to physiographic regions.
Developed by Bureau of Land
Management.

U.S. Department of the Interior,
Bureau of Land Management
1978

Land resource regions and
major land resource areas

Groups organizationally related
and geographic factors related
to land use, topography,
climate, water, and soil.
Developed by the Soil
Conservation Service.

Austin 1965; Kuchler 1964

Land use and cover mapping
program

Broad-based information, i.e.,
not intended for wildlife.
Developed by the U.S.
Geological Survey.

Anderson et al. 1976

Forest cover types

Classification based on
dominant tree species.

Society of American Foresters
1954

Forests 75

classification system in use by dominant forces in the
agency? A mechanism exists to allow data collection
at the most meaningful level for specific wildlife
purposes. One such mechanism, developed for use in
the Blue Mountains of the Pacific Northwest, has
proved useful there and elsewhere (Table 2). In
such a system, the classifications are "crosswalked"

so that information gathered under one system can
be appropriately applied to others.

Given such information as that shown in Table
2, data on wildlife habitat can be developed at the
most appropriate level for the wildlife resource and
then, if appropriate, can be applied to other land

Table 2. Relationships among four plant community classifications in the Blue Mountains of Oregon and
Washington (adapted from Thomas et al. 1979a).

Forest and range
ecosystems as
described by Garrison et
al. (1977)

Plant community
designations from
Thomas et al. (1979a)

Plant community types
of the Blue Mountains as
described by Hall (1973)

Potential natural
vegetation as described
by Kuchler (1964)

Pinyon-juniper'

Western juniper
Juniper/stiff sage scab-
land
Juniper/low sagebrush
Juniper/big sagebrush

Juniper/bunchgrass

K24 Juniper steppe
woodland

Ponderosa pine

Ponderosa pine

Ponderosa pine/fescue

Ponderosa pine/bitter-
brush/Ross sedge

Ponderosa pine/blue
wildrye

Ponderosa pine/wheat-
grass

K10 Ponderosa shrub
forest

Douglas-fir

Mixed conifer

Ponderosa pine/Doug-
las-fir/elk sedge

Ponderosa pine/Doug-
las-fir/snowberry-
oceanspray

Ponderosa pine/Doug-
las-fir/ninebark

K12 Douglas-fir forest
(interior)

Larch

Mixed conifer/pinegrass-

ash soil
Mixed conifer/pinegrass-

residual soil

K14 Grand fir/Douglas-fir
forest

Spruce-fir

White fir (grand fir)

White fir/twinflower
White fir/big huckleberry
White fir/grouse huckle-
berry

Subalpine fir/big huckle-
berry

Subalpine fir/grouse

huckleberry
Subalpine fir/whitebark

pine/sedge

K15 Western spruce-fir
forest

Lodgepole

Lodgepole pine

Lodgepole pine/pine-
grass-grouse
huckleberry

Lodgepole pine/big
huckleberry

Lodgepole/grouse
huckleberry

No provision in Kuchler
(1964)

1See specific reference citations for a listing of scientific names of plants noted in this table

76

Forests

classification systems. The rule-of-thumb, then, for
selecting the appropriate habitat classification system
to be used as the basis for inventorying and monitor-
ing has two components:

( 1 ) Select the system in use by the dominant profes-
sional group in the agency in order to enhance
communication.

( 2 ) If another finer or coarser grained classification
is used as the framework for the wildlife-habitat
data base (for whatever reason), make sure it
can be "crosswalked" with other classification
systems (Verner 1984).

The ultimate objective is to present wildlife
habitat in such a manner that it will be effectively
considered by land-use planners and managers. The
selection of a habitat classification system that may
be technically superior for wildlife-habitat evaluation
purposes is a strategic error if it cannot be easily
translated into the system used by the dominant
resource.

This requirement for crosswalking is addressed
by Bailey (1980:17) in his discussion of the ecosys-
tem concept of classification:

"This concept regards the earth as a series of
interrelated systems in which all components
are linked, so that a change in any one compo-
nent may bring about a change in other com-
ponents and in the operation of the whole. An
ecosystem approach to land evaluation stresses
the interrelationship among components
. . . Since ecosystems are spatial systems, they
will be consistently inserted, or nested, into
each other. Each level subsumes the environ-
ment of the system at the level below it. At
each level, new processes emerge that were
not present or not evident at the next lower
level . . ."

In this example (Table 2 ), the dominant land
management agency (U.S. Forest Service) active in
the area used Hall's ( 1973) description of plant com-
munities as the basic classification system to guide
forestry and range management decisions. On the
whole, the divisions were too fine to be germane to
the classification of information about wildlife-habitat
relationships. Therefore, for relating wildlife to habi-
tats, resource professionals developing information
on wildlife-habitat relationships (Thomas et al.
1979a) modified the classification system derived by
Garrison et al. (1977). They thought that these
broader community descriptors were more suitable
to the existing state-of-knowledge concerning wild-
life-habitat relationships. The biologists involved
recognized that the vegetative classification schemes
developed by Kuchler ( 1964 and 1970) were receiv-
ing favorable attention by persons dealing with na-

tional assessments, but they thought that such a
system was too broad to be meaningful in the con-
text of wildlife-habitat relationships.

The decision about what classification system to
use may have already been decided by the existence
of a wildlife-habitat data base for the area in ques-
tion. The development of such data bases was stimu-
lated by requirements of the National Environmental
Policy Act of 1969 (which required environmental
impact statements for federally funded projects), the
Forest and Rangelands Renewable Resource Planning
Act of 1974, the National Forest Management Act of
1976, and the Federal Land Policy and Management
Act of 1976. The first such data bases for the west-
ern States were developed by Patton ( 1978) for the
Southwest and Thomas ( 1979b) and associates for
the Pacific Northwest. These initial publications were
quickly followed by similar efforts for the western
Sierra Nevada in California (Verner and Boss 1980),
the Great Basin of southeastern Oregon (Maser and
Thomas 1983), and western Oregon and Washington
(Brown, in press). Parallel efforts are underway in
several other areas in the western U.S. and Canada.
The Fish and Wildlife Habitat Relationships Program
of the U.S. Forest Service was established to encour-
age the development of similar programs throughout
the U.S. and to enhance development of additional
habitat analysis techniques. Cushwa and DuBrock
(1982) detailed various wildlife data bases in exist-
ence or being developed by various State and Fed-
eral agencies.

Various permutations of this basic approach to
wildlife-habitat analysis seem to be most widely used
in dealing with analysis and management of forested
habitats of the western U.S. The basis of the original
system was to consider wildlife habitats at three
levels:

( 1 ) The relationship of groups and individual spe-
cies to major plant communities and their
successional stages or structural conditions for
activities, including feeding and reproduction;

(2) A description of special habitats that were not
adequately considered through plant communi-
ties and successional stages;

(3) The habitat requirements of featured species,
i.e., the species that receive special management
attention.

MAJOR SPECIES GROUPS

The purpose of this section is to review and
evaluate the existing approaches to forming groups
of species for inventory and monitoring in forests so
as to reflect habitat conditions.

Forests

77

When dealing with the vertebrate component of
the forest ecosystem, it is necessary to determine
which species or group of species need to be inven-
toried. Obviously, there are insufficient knowledge,
skill, and resources to deal with inventory and moni-
toring of all the species.

When wildlife-habitat data bases are developed,
each individual species must be considered. But if
the relationships of all vertebrate species within an
area to habitat features are considered, the detail and
sheer volume of data can be overwhelming. When
possible, it would be more convenient and cheaper
to use groups of species rather than individuals.

Featured Species

The desirability of the featured species approach
(examining the relationship of groups of species to
habitat variables) is in keeping with the necessity for
the "strenuous reductionist approach" mentioned
earlier in relation to habitat classification. A group
called the Committee of Scientists, who recom-
mended regulations as required by the National For-
est Management Act of 1976, called for use of
indicator species in setting up and monitoring forest
management plans. These indicators were to serve
as surrogates for the welfare of a group of species.

Assume that the spotted owl (Strix occiden-
talis) was chosen as the indicator species for species
that thrive in old-growth Douglas-fir (Pseudotsuga

menziesii) forests in western Washington. If the
spotted owl's habitat requirements were met, ac-
cording to the indicator species concept, other spe-
cies that require old-growth Douglas-fir would be
provided for simultaneously.

Verner (1984) reviewed the use of the guild
concept as applied to management and essentially re-
jected the indicator species concept on practical
and technical grounds. A guild is made of species
within an order that jointly exploit the same re-
source in an ecosystem. He did believe that the use
of whole guilds (i.e., each bird of any species within
a guild is tallied in sampling) to evaluate habitat
conditions might be useful but recommended that
(p. 1 ) "much testing must be done before it is ap-
plied." This means that the research community has
not yet tested the hypotheses inherent in the indica-
tor species approach. Field biologists who are using
the approach should admit that to themselves and to
the users of that work.

Guilds. The guild concept assumes that animal
species within an order can be grouped by how
similarly they use environmental resources (Root
1967). Severinghaus (1981 ) applied the guild
concept to species of vertebrates according to their
response to habitat and habitat change, delineating
30 mammal and 31 bird guilds. He said (p. 189),
"The usefulness of guilds . . . relies on the principle
that . . . actions impacting one member of the guild
should impact . . . other members in a similar way."

Short and Burnham ( 1982) and Short (1983)
presented a technique for structuring guilds for use
in evaluating impacts of habitat alterations on wild-
life communities. Described are mathematical and
computer aids to form guilds, as opposed to the
more intuitive classifications proposed by Thomas et
al. ( 1979a), Severinghaus ( 1981), and others.

The Short-Burnham and other mathematically
derived techniques, although holding promise, are
probably more complex than needed for application
by most field biologists. Mathematical sophistication
may exceed biological sensibility in this approach
(Verner 1984). It might be best to contract for these
services if this approach is broadly applied in
management.

Thomas et al. ( 1979a), in an effort to respond to
the recommendations of the Committee of Scientists,
modified a suggestion of ornithologist Antti Haapa-
nen ( 1965, 1966) that species of birds could be
grouped according to their habitat use into "life
forms." They extended the concept to include all
vertebrates and grouped the species (379) that oc-
curred in the Blue Mountains of Oregon and Wash-
ington into 16 life forms. Life forms include
groupings of vertebrates — regardless of order — that
require similar habitat attributes.

78

Forests

Oversophisticated mathematics may shed great darkness
on the subject.

What those initial life forms were is not as im-
portant as the concept; for habitat management, ver-
tebrate species could be grouped by the general
type of habitat they used for feeding and reproduc-
tion, irrespective of taxonomic considerations. The
concept assumes only that each species in a life form
will be present in suitable habitat. It does not, how-
ever, assume that species within the life form use the
habitat similarly or that numbers of the various spe-
cies will rise and fall simultaneously.

Life forms are constructed to fill a management
need, and the number of life forms is limited only by
the amount of knowledge available on the relation-
ship of the individual species to habitat. Life forms
can be created or eliminated, or species can be
moved between life forms to fit the need or to re-
spond to new information. It is also important to
understand that in using the life form approach, it is
the habitat condition that is inventoried and moni-
tored, not the species in the life form.

The three approaches discussed for grouping
species in order to relate them to habitats (indicator
species, guilds, and life forms) share the idea that
animal species can be grouped according to re-
sponse to habitat. If so, that grouping can be a real
aid in the simplification of habitat analysis across the
spectrum of the vertebrate species present in the
area to be considered.

Other approaches have included those of Root
(1967), MacMahon et al. (1981), andjaksic (1981).
Techniques used to formulate guilds include cluster
analysis (Crome 1978), principal components analy-
sis (Holmes et al. 1979; Landres and MacMahon
1980; Short and Burnham 1982), canonical correla-
tion (Folse 1981), discriminant function analysis
(DeGraaf and Chadwick 1984; Dueser and Shugart

1979; Lindeman et al. 1980; Nie et al. 1974; Rao
1966; Williams 1980), and polar-ordination (Ander-
son 1971; Austin and Orloci 1966; Bannister 1968;
Bray and Curtis 1957; DeGraaf and Chadwick 1984;
Gauch 1982). Note that all these techniques produce
investigator-defined guilds, "a fact that introduces
circularity into their use for examining ecological
questions" (Verner 1984:1 ). This simply means that
there is a danger, if the investigator defines the guild,
of altering the expected to fit the observed species-
habitat conditions and vice-versa.

It seems inappropriate to present lists of life
forms or guilds here. They should be developed by
the user from the available wildlife data base on the
occurrence of animal species and their relationship
to various habitat conditions. The only advantage
of dealing with life forms or guilds is the enormous
savings that can result from monitoring whole guilds
instead of indicator species. Conversely, if the deci-
sion is made to monitor indicator species, the spe-
cies groupings delineate those species that the
indicator species represents. We reject the indicator
species approach to monitoring as neither reasonable
nor practical for the reasons presented by Verner
(1984). In practice, most management analysts have
fallen back on consideration of individual species —
either en masse or of a select few of particular
interest.

We recommend the following:

( 1 ) Do not use the indicator species approach un-
less regulations require it. If the indicator spe-
cies approach is used, recognize that it is merely
a hypothesis and that the users must provide
the test.

(2) Use the guild approach, as it seems to be the
most economical and practical approach to
monitoring species as reflective of habitat
condition.

(3) Use the life form approach when monitoring is,
for financial or other reasons, restricted to vege-
tative conditions; it is essential to understand
that the approach involves a hypothesis that
must be tested by the user.

CRITICAL HABITAT FEATURES

Soil

Soil is a primary variable contributing to the
development of plant communities. With the excep-
tion of fossorial vertebrates, the soil type per se does
not appear to be a controlling factor in animal habi-
tat. The relationships between soils and wildlife are
thoroughly discussed by Robinson and Bolen (1984).

Forests

79

Thomas et al. (1979a:26) suggested that —

"Plant communities and their successional
stages have unique environmental conditions
that are ecologically important as niches for
wildlife species . . . The niches are a product of
the plant community, its successional stages,
and other environmental factors — including soil
type, moisture regime, microclimate, slope,
aspect, elevation, and temperature. The com-
plex interactions of site and plant community
structure could be dissected and the more
precise of each on the animal community de-
termined. If such information existed, it would
probably be too complex to use readily. The
plant community type, however, can be consid-
ered an integrator of the many factors interact-
ing on the site."

The factors that plant communities integrate
include soils. We recommend that soils not be a part
of monitoring unless there are particular reasons for
including them, i.e., the soil texture has particular
significance to a wildlife species of sufficient interest
that monitoring is deemed appropriate. Even then,
it is important to identify what soil attribute is signif-
icant to the occurrence of the species in question.
This could include such things as texture, size of soil
particles, cohesiveness, moisture retention proper-
ties, depth, and drainage.

Physical Features

Maser et al. (1979a, b) noted that geomorphic
and edaphic habitat features occurred within the
general vegetative mosaic and have special values to
wildlife. These "unique" or "geomorphic" habitats
included cliffs, caves, talus, lava flows, sand dunes,
and playas. Each of these habitats supports one or
more species not found in the general forest
environment.

Talus is the accumulation of rocks at the base of steep
slopes.

Verner and Boss (1980:2-3) called such compo-
nents "special habitat requirements," adding ground
burrows, friable soils, moist soil, earthen bank, rock
outcrops, limestone outcrops, and crevices. The spe-
cies to which such features are important were iden-
tified. For example, friable soils are important to
burrowing animals; dead-and-down woody material is
critical to reptiles, amphibians, and small mammals;
caves harbor bats; and cliffs provide nest sites and
roosts for raptors.

Plant Communities and Successional Stages

The forest habitat types (from whatever classifi-
cation scheme is chosen for use) are the first vegeta-
tive component considered. Forest condition within
each habitat type is described by successional stage.
Thomas et al. (1979a) identified six such stages:

( 1 ) grass forb

(2) shrub seedling

(3) pole-sapling

(4) young

(5) mature

(6) old-growth

Verner and Boss ( 1980) suggested four such stages
for forest habitats:

(1) grass/forb

(2) shrub/seedling/sapling

(3) pole/medium tree

(4) large tree

Special Habitat Features

According to Thomas et al. (1979a:21), special
habitats are "biological in nature, can be manipulated
by the forest manager, and play a critical role in the
lives of at least some species. These habitats include
riparian zones, edges, snags, logs, and other dead
woody material on the forest floor." Verner and Boss
(1980) added forest openings, elevated perches,
nest cavities, hollows, and litter to the list of what
they called "vegetation elements." Which of these
special habitats should be included in inventories or
monitoring operations depends on the wildlife spe-
cies or groups of species considered important by
the responsible land manager. For example, snags
would be important if cavity-nesting birds received
management emphasis, and logs and other dead
woody material would be of interest if reptiles and
amphibians were of interest. Of these, we believe
that only riparian zones, edges, snags, logs, and forest
openings will generally be considered critical inven-
tory elements.

Riparian Zones. Riparian zones are those areas
identified by the presence of vegetation that requires
free or unbound water. When they occur within

80

Forests

forested areas, they generally create well-defined
mesic areas within much drier surrounding areas.
They make up a minor portion of the area, are more
productive of biomass than surrounding forests, and
are a dramatic source of diversity. Of the 378
vertebrate species in the Blue Mountains of Oregon,
285 were identified as dependent on riparian zones
or as using them more frequently than other habitats
(Thomas et al. 1979b).

Edges. Edges are places where different plant
communities or successional stages meet and are
identified as particularly rich in wildlife. Leopold
(1933:131) stated that "game wildlife is a
phenomenon of edges . . . We do not understand the
reason for all of these edge effects, but in those cases
where we can guess the reason, it usually harks back
either to the desirability of simultaneous access to
more than one environmental type, or the great
richness of border vegetation or both." Patton
(1975) suggested that the amount of edge per unit
of area might serve as an index to diversity.

Thomas et al. (1979d) divided edges into inher-
ent edges (those resulting from the meeting of two
plant communities) and induced edges (those be-
tween two successional stages within a plant com-
munity). They suggested further that the degree
of contrast in height of vegetation along an edge
would be correlated with the diversity of animals
encountered along the edge and in adjacent stands.
Initial testing supports this hypothesis.

Snags. Snags are dead or partly dead trees at least
10.2 cm (4 in.) in diameter at breast height (dbh);
snags at least 1.8 m (6 ft) tall are critical to many
species of forest wildlife. For example, in the Blue
Mountains, 39 bird and 23 mammal species (16% of
the 379 species occurring) use snags for nesting or
shelter. Snags are now recognized by managers of
public forest land as a critical habitat component.

Recommendations on size, species, and numbers
of snags required to support various population lev-
els of wildlife dependent on snags are beginning to
emerge (Thomas et al. 1979b). The following infor-
mation on snags should be gathered:

• species

• height

• dbh inside bark

• degree of decay ( hard or soft )

• top condition (whole or broken)

• surrounding vegetative structure

• density

• use by wildlife

Logs. Logs (dead-and-down woody material) are
dead trees or portions of trees lying on the forest

Logs partly submerged in water form an important link
between terrestrial and aquatic habitats for some species.

Snags are critical habitat components for many species.

Forests

81

floor. Logs are important wildlife habitat components
(Maser et al. 1979c). Elton (1966:279) said—

"When one walks through the rather dull and
tidy woodlands . . . that result from modern
forestry practices, it is difficult to believe that
dead wood provides one of the two or three
greatest resources for animal species in a natu-
ral forest, and that if fallen timber and slightly
decayed trees are removed, the whole system
is gravely impoverished of perhaps more than a
fifth of its fauna."

In the Blue Mountains, for example, 5 amphibi-
ans, 9 reptiles, 116 birds, and 49 mammals (47%
of the 379 species present) make some use of dead-
and-down woody material (Maser et al. 1979c) (Fig-
ure 1 ). This is evidently common because Elton
(1966:279), speaking of England, said "... indexes of
the Geological Survey contain 456 species of animals
(including invertebrates) at Wytham living in wood
or under bark where decay had begun or already
gone far. Another 518 species are known to occur in
this habitat elsewhere in Britain." Maser et al.
(1979c) presented a state-of-knowledge review
about dead-and-down woody material as wildlife
habitat. They also presented information about the
role of such material in forest ecology.

Forest Openings. Forest opening areas within a
forest that do not support trees (for whatever
reason) are an important source of diversity within
the forest. They frequently support species that do

Openings provide diversity in a forest.

not occupy forested habitats per se, and they are
important to species associated with their edges. For
example, more elk and mule deer (Odocoileus
hemionus) occur within about 183 m (600 ft) of
edges between forest cover and forage areas
(Reynolds 1966; Harper 1969; Leckenby 1984). This
was a key factor in the development of elk and deer
habitat models that are receiving wide application
(see Thomas et al. 1979c for an example). The
inventory of forest openings can be accounted for
under the inventory of plant communities,
successional stages, and edges discussed earlier.

Structural Conditions of the Forest

Even greater detail may be required for struc-
tural conditions of the forest beyond what can be
inferred from knowledge of the successional stage of
a given plant community. Included are the following:

• dbh of trees, treating conifers and broadleafs
separately

• tree height and height of the base of the
crown

• canopy volume — the space occupied by the
crowns of trees

• canopy closure

• understory volume — the space occupied by
the leaves and limbs treating shrubs separately
from grasses and forbs

• sight distance

• stand size

Dbh. Dbh can be important in dealing with
management of snags and dead-and-down woody
material. Living trees are the snags and logs of the
future. Division into conifers and broadleaved trees
can be important, as many birds and some arboreal
mammals are influenced by this variable (Thomas et
al. 1977).

Tree Height and Height of Base of Crown. Tree
height and the height of the base of the crown must
be known if the relationship between vegetative
volume and wildlife species is to be considered.

Canopy Volume. The importance of considering
canopy volumes, particularly for purposes of
explaining bird occurrence, has been emphasized by
a number of investigators (reviewed in Larson
1981). MacArthur (1964) believed that variability in
bird occurrence due to foliage height diversity could
be accounted for by considering three layers of
vegetation: 0-60, 60-750, and 750 cm (0-2, 2-25, and
25 ft). Thomas et al. (1977), however, found that
considering vegetative volumes in 1.5-m (5-ft) layers

82

Forests

The trunk provides
a food source for
woodpeckers, particularly
pileated woodpeckers.

The root wad is used by
flycatchers for perching,
by grouse for dusting, and
by juncos for nesting

Elevated areas are
used as lookouts
and feeding sites.

Limbs are used as
perches, and if hollow,
as nest cavities

The spaces between loose bark
and wood are used as hiding
and thermal cover by invertebrates
and small vertebrates, such as the
Pacific treefrog.

Protected areas under the
log are used as nesting
cover by grouse and as
hiding and thermal cover
by snowshoe hares.

Figure 1. Downed logs furnish many structural features important to wildlife.

77k- myriad of openings and vegetation types in this urea form edges important to wildlife.

Forests

83

markedly increased the variability in bird occurrence
explained by vegetative volumes. Mawson et al.
(1976) described a technique called HTVOL (height-
volume) whereby trees were measured within a
plot in a way that vegetative volumes in various
"slices" could be determined. Each tree to be
sampled was assigned to a geometric form and
appropriate measurements taken. The degree of
detail needed to determine vegetative volumes
depends on the investigator's need.

Canopy Closure. Canopy closure provides
information that is useful in describing the stand
quality in terms of thermal cover, i.e., cover that is
used by ungulates to thermoregulate during
extremes of ambient temperature (Thomas et al.
1979e) (Figure 2). Such information can also be
useful in predicting the productivity of understory
vegetation within such stands.

/Thermal cover conditions

/ $ / /o

/ ••£■ / / o

/ <P> / £ / £

' £ / § / $

„ IUU

c

CD

o

g 80

1/5

f /

1/1

CD

cB 60

-

' J*

>

<F

o
% 40

A<

CD

>

O

o „„

_ 20

—

CO

E

CD

£ n

■ i i

i i

• i

0

20 40 60 80 100

Canopy closure (percent)

Figure 2. Effectiveness of thermal cover for deer
and elk is closely related to canopy cover.

Understory Volume. The volume of understory
vegetation (particularly shrubs) is useful in
predicting bird occurrence (MacArthur 1964; also
see review by Larson 1981) and as a measure of
forage available for herbivores. Forage availability
can be expressed as kg/ha (lb/a.) of air-dried forage.

Sight Distance. Sight distance is the distance at
which 90% or more of a standing deer or elk is
hidden from view. Sight distance can be useful in
evaluating the effectiveness of hiding cover for deer
or elk. Hiding cover has been denned as existing

when an elk or deer is 90% or more hidden at a
distance of 61 m (200 ft) or less from an observer
(Black et al. 1976).

Stand Size. Stand size is an important habitat
variable for several reasons. For example, optimum
stand sizes are recommended to ensure use by elk
and deer. These sizes usually range from 10 to 16 ha
(25 to 40 a.). Some consider stand width to be a
more important measure of stand size than area, as
this characteristic, because of edge effect, has the
most influence (positive and negative) on animal
occurrence and density (Leckenby 1984). For
example, recommendations for optimum habitat for
deer and elk call for stand width to be held at 183-
366 m (600-1,200 ft; Thomas et al. 1979c). In
addition, stand size is related to the number of
species that will be found in the stand and the
diversity of species within the stand. The number of
species that are found within a stand is positively
correlated with the size of the stand up to some size.
This relationship is called the species/area curve;
see Cain and Castro (1959), Greig-Smith (1964),
Galli et al. (1976), and Whitcombe et al. (1981) for
reviews of this concept.

Species Composition

It is sometimes desirable to consider the species
of plants involved in a habitat. This information is
gathered in addition to categorization of the plant
community (identified by characteristic dominant
vegetation in the climax stage) and the successional
stage. The following examples show why data about
species composition may be valuable in wildlife habi-
tat assessment:

• variability in palatability and food value in for-
age plants between consumer species;

• differences in values of various species of trees
to cavity nesters;

• indication of range condition.

Remember, however, that collection and analysis of
inventory data including species composition in-
creases costs many times. In some instances, such as
the classification of plant communities, it may be
sufficient to record plant species as present or absent
or to deal with estimates of relative abundance.

INVENTORY

Broadscale Systems

The broadscale inventorying and monitoring of
wildlife populations and wildlife habitats are man-
dated by law on National Forest lands but little has

84

Forests

been done to date (Salwasser et al. 1983)- There are
no well developed, well tested systems to accom-
plish those tasks, although some efforts are underway
to produce the tools to accomplish them (Hoekstra
et al. 1979, 1983; Verner 1983). The systems that
are available have been called "working hypotheses,"
and their developers advise caution in their
application.

Special Habitat Features

This first level of analysis may not be sufficient
to deal with special habitat features or with the de-
tailed requirements of individual species of special
concern to management. Special and unique habitat
features are then considered. These were listed and
discussed earlier.

The basic system that has received the broadest
application on forested land in western North Amer-
ica was initially described by Thomas ( 1979a) and
associates and has become known as the Wildlife and
Fish Habitat Relationship (WFHR) approach. The
system has been and is being modified in the permu-
tations being developed for other ecosystems in the
U.S. and Canada. Examples include the application
for the western Sierra Nevada (Verner and Boss
1980), western Oregon and Washington (Brown, in
press), and the Oregon portions of the Great Basin
(Maser and Thomas 1983). This general approach
has been used in many other areas in the West, but
the adapted models and data bases have not been
formally published. Scientists in British Columbia are
working on the system to help describe the ecosys-
tems of British Columbia.

All applications of this general approach attempt
to identify habitat needs for feeding and reproduc-
tion (some adaptations include other activities such
as resting) by each vertebrate species for each plant
community, according to successional stage or struc-
tural condition. Most approaches use the formation
of life forms or guilds to facilitate analysis and moni-
toring. Expected changes in habitat can be superfi-
cially evaluated by comparing the wildlife in the
existing habitat to those in the anticipated habitat.
Predicted changes in species occurrence are, indeed,
the evaluation. Some species will be helped, some
will be hurt, and some will not be affected by any
major change in a given habitat. The impact on each
species, or group of species, can be evaluated at a
crude level.

The required habitat inventory for this level of
analysis is relatively quick and easy to accomplish by
simple delineation of plant communities and their
successional stages or structural conditions. Aerial
photo-interpretation is easily applied, and imagery
from multispectral scans obtained from satellites
(LANDSAT) shows promise in this regard and is
being effectively applied in the Blue Mountains of
Oregon and Washington by the Oregon Department
of Fish and Wildlife and the U.S. Forest Service in
monitoring big game habitat conditions. It has been
particularly useful in deriving forage-cover ratios,
sizing and spacing of cover and forage areas, and
quantification of cover quality.

Here we use snags as an example. Standard for-
est management regimes are unlikely to provide
snags in adequate numbers to support viable popula-
tions of all members of the life forms that excavate
cavities in dead trees for nesting or for those that
occupy existing cavities for nesting. Therefore, a
snag inventory should be routinely done in managed
stands if the welfare of the wildlife species that use
snags is of concern. This can be accomplished by
regular inventorying by forest crews whose primary
job is to provide data to support the timber manage-
ment program (and who, importantly, are well fi-
nanced to do that job) or, with less accuracy, by use
of aerial photography.

Admittedly, snags are an obvious and relatively
easy special habitat feature to inventory and monitor.
They are also a feature that has received considera-
ble management attention. Other features, such as
dead-and-down woody material, edges, talus, and
others have not been routinely inventoried. Whether
these features should be inventoried depends on
management objectives. If there are wildlife species
that depend on such features that are of significant
interest to management, the associated habitat fea-
ture should be monitored. The relationship of wild-
life species to several special habitat features is
detailed in such publications as Thomas (1979b).
The techniques for inventory, if they exist, have not
become standardized. Also, there is no standard list
of special habitat features. Whether to inventory and
monitor special habitat features, and which ones to
address, are determined by the habitat management
objectives of the area in question. Prime candidates
for inventory and monitoring are those that are criti-
cal to featured species, indicator species, or manage-
ment guilds. Significance and priority for attention
are determined by emphasis offered by the manage-
ment plan.

When the needs of individual species must be
considered, additional habitat features germane to
management of that species will become important.
This is illustrated by the discussion of featured spe-
cies that follows.

Featured Species

Featured species, or species of special interest
to managers, usually require special consideration.

Forests

85

These usually include game species, threatened or
endangered species, management indicator species,
and ecological indicator species. The National Forest
Management Act and the regulations issued pursuant
to that Act require the maintenance of viable popula-
tions of all native and desirable non-native species
in National Forest management. The regulations
specify that progress toward achievement of that
goal be evaluated by monitoring the occurrence and
density of management indicator species. Astute
selection of management indicator species and assur-
ance of their subsequent welfare were assumed to
be an appropriate mechanism to meet legal require-
ments for maintenance of viable populations of all
wildlife. Included within this group are some rather
vaguely denned "ecological indicators" which, by
implication, represent a larger group of species. We
believe this approach to be impractical and unrealis-
tic for the reasons specified by Verner ( 1 984 ). It is,
however, the law and the U.S. Forest Service must
attempt to comply. A rational approach to carrying
out such an effort has been spelled out by Verner
( 1983). Agencies not bound by the National Forest
Management Act are advised to use other approaches
such as management guilds (discussed later), or stay
with individual species monitoring without inference
to other species.

No general comprehensive list of habitat attri-
butes for featured species can be listed here. These
attributes are unique to the species being consid-
ered. The description of the system by Thomas et al.
( 1979c) used elk and mule deer as examples of fea-
tured species. The habitat attributes to be measured,
and mentioned throughout this chapter include —

• percentage of the area consisting of thermal
cover, hiding cover, and forage areas,

• stand size,

• density of roads open to vehicular traffic,

• cover/forage area ratios.

If, however, the pileated woodpecker (Dryoco-
pus pileatus) were chosen as a featured species,
other habitat attributes would become important,
such as —

• snags greater than 51 -cm (20-in.) dbh,

• species of snags,

• height of canopy,

• dead-and-down woody material,

• the presence of carpenter ants ( Camponotus
sp.).

Each featured species will introduce a different set of
habitat attributes to be considered.

The U.S. Department of Interior, Fish and Wild-
life Service (1980) has developed "Habitat Evalua-
tion Procedures (HEP)" focusing on habitat
requirements of fish and wildlife species applicable
to project planning and impact evaluation (Scham-
berger and Farmer 1978). This was the impetus for
development of Habitat Suitability Index (HSI)
models (Schamberger et al. 1982) for select wildlife
species. These models provide a good source of in-
formation on habitat variables that one should meas-
ure when using one of the particular species as an
indicator or featured species. Examples of species for
which such habitat models exist include beaver
{Castor canadensis) (Allen 1982a), marten (Martes
americana) (Allen 1982b), and black-capped chicka-
dee (Parus atricapillus) (Schroeder 1983).

The relationship between these habitat evalua-
tion procedures and the habitat evaluation proce-
dures of Thomas (1979b), Verner and Boss ( 1980),
and others is explored in detail by Thomas (1982).

MONITORING

A wag once said that monitoring was an inven-
tory done over and over. So it is. The National Forest
Management Act of 1976, and regulations issued
pursuant thereto, contain instructions to monitor
indicator species to determine ( 1 ) if management
objectives are being met, and (2) if the management
indicator species are changing in status. The costs
of such monitoring of individual wildlife species are
likely to be very high in personnel and dollars. Ver-
ner (1983) estimated that for the western Sierra
Nevada it would take some 300,000 10-minute point
counts to detect a 10% change in numbers of such
species of special interest such as the pileated wood-
pecker or willow flycatcher (Empidonax traillii). If
20 counts per observer per day were made, 1 5,000
observer days would be required. Because the suita-
ble counting period within any one season is about
60 days, some 250 temporary employees capable
of locating the sample points and capable of recog-
nizing birds by sight and sound would be needed.
Salary costs alone are estimated at $825,000 per
year. Verner (1983:357) summed up by saying:

"Not only are the costs and personnel needs
for such an effort out of reason, but also it
is unreasonable to find such a large number of
counting points because each must be indepen-
dent of the others to satisfy assumptions of the
statistical models. This fantasy can be extended
to include all of the various species we may
be concerned about and in all of the various
types of habitats where they may occur. Ob-
viously other approaches are required."

86

Forests

Note also that even after a change in the popula-
tion is detected, this does not assure information
about cause and effect. Is the change due to weather,
food supply, habitat change, or some other factor or
combination of factors? Are such changes indicative
of something that requires management action or
merely normal perturbations in population numbers?

Toward More Practical Monitoring

Verner (1983) suggested various ways to make
monitoring more feasible in terms of costs and man-
agement need. These are summarized below:

( 1 ) Separate species into categories of high and low
risk to be reduced below acceptable levels in
terms of range or density by forest management
activities. We believe that the vast majority of
species will fall in the low risk category and that
habitat monitoring will suffice to monitor trends
in their status. High risk species will probably
require monitoring of the species itself.

(2) Design monitoring activities to detect only de-
clining trends. Only in rare circumstances, such
as with ungulates (which are hunted ) or go-
phers (which cause damage to young trees),
would increases be a cause for concern. Besides,
data on animal damage are very apt to be avail-
able from sources such as forest surveys for

the forestry program. The monitoring should be
designed to protect against statistical type II
errors (concluding that a population is stable
when it has really declined).

Costs vary with the power of the test and the
magnitude of the decline to be detected. We
recommend a power of 0.8 (failure to detect a
specified decline 20% of the time) as a good
compromise between tolerable costs and man-
agement needs. In all situations, the sampling
design should be standardized to produce as
nearly identical conditions as possible every
time the sampling is done (probably annually).
The following should be considered:

• sampling intensity

• time of day

• season

• sample sites

• weather

• order of sampling sites

• variability due to observers

(3) Select "high probability" sites for sampling when
indicator species are monitored. Management
indicator species, with the exception of game

species, are apt to be uncommon; in fact, they
may have been selected because they were rare,
threatened, or endangered.

Therefore, average counts of such species are
generally so low in a randomly selected set of
sampling points that costs preclude sampling
enough points to detect declining trends with
statistical confidence (Verner 1984). A pre-mon-
itoring inventory should be used to identify
sites where target species occur and monitoring
stations should be randomly selected from
among those sites.

(4) Monitor "management guilds" rather than indi-
vidual species. Management guilds are species
that respond in a similar way to changes in their
environment. Verner (1983:360) said, "The
primary use of such management guilds should
be to monitor trends in the suitability of various
zones of a habitat to support wildlife species.

As such, guilds will probably make their greatest
contribution to a monitoring system as indica-
tors of the quality and quantity of certain habi-
tats that are most likely to be changed by
management activities."

One obvious advantage of monitoring manage-
ment guilds is that counts of a group of species
are higher than counts of any of the component
species. Therefore, fewer samples are required.
Further, the biologist obtains a more complete
listing of the species assemblage than would
result if only indicator species were counted. At
least for birds, the cost of obtaining the more
complete list is no greater than tallying just one
or a few species (Verner 1983). This assumes, of
course, that personnel are available who can
readily identify the species by both sight and
sound. This gives managers at least some hint of
the changing status of individual species.

(5) Monitor only during a single season. Ideally,
managers should have knowledge of the year-
round status of habitats and occupying species
that would require at least four monitoring ef-
forts per year. Likely, only one effort per year
can be afforded. For birds, that effort should
correspond with the breeding season (Verner
1983). For game species, particularly big game,
monitoring is best done as close to hunting
season as feasible to allow time for adjustment
in hunting regulations.

For birds, Verner (1983:361) hypothesized that
"If transients and winter residents use the same
zones of a habitat for feeding and cover in basi-
cally the same ways as the breeders do, mainte-
nance of the populations of all breeding species
should ensure maintenance of suitable habitat
for transients and winter residents."

Forests

87

(6) Monitor trends in habitat. The assumption in
such monitoring is that if enough is known
about species-habitat relationships, trends in
populations (or potential populations) could be
inferred from monitoring habitat trends. These
systems are probably adequate for low risk spe-
cies, at least for the time being. Nonetheless
they are still in a rudimentary stage. They re-
quire testing, verification, and continuous
correction.

Ultimately, it may be possible to effectively
monitor trends in most wildlife populations almost
exclusively through habitats, even for management
indicator species. This will be the most cost-effective
method because trends in habitats must be moni-
tored for information on timber resources. Thus, the
cost of monitoring wildlife trends can be mainly
absorbed by monitoring timber resources.

An Example of a Monitoring System

The Sierra National Forest has developed a wild-
life monitoring plan that seems well conceived, ade-
quate, realistic, and promising (Verner 1983). The
plan calls for 5 years of inventory to precede institu-
tion of monitoring efforts. This inventory effort will
provide the sampling framework and baseline infor-
mation on which monitoring plans are based. The
plan is in three phases:

( 1 ) Monitoring of management indicator species.
This is required by law, and the minimum legal
requirement will be met by monitoring seven
species, five of which are rare, threatened, or
endangered, and two of which are game species.

Monitoring nesting habitats of rare, threatened, or endan-
gered species should be done yearly.

DISCUSSION

(2) Monitoring, at 2 -year intervals, management
guilds of birds in three habitats considered espe-
cially vulnerable to alteration by management.
These habitats include late-successional mixed-
conifer forests, riparian habitats, and meadow
edges. Each would include 200 sampling points.
It is hypothesized that stability in bird assem-
blages indicates similar stability in assemblages
of other vertebrate groups.

(3) Monitoring of most wildlife species in most
habitats by monitoring trends in habitat defined
by plant community, successional stage, and
special and unique habitat features such as snags
and dead-and-down woody material. Such an
approach requires the development of data
bases of site-specific wildlife-habitat relationships
such as those described by Patton ( 1978),
Thomas (1979b), or Verner and Boss ( 1980).

The inventorying and monitoring of forest wild-
life habitats and forest wildlife (except for game
species) are in their infancies. The literature reveals
more speculation than action at this point. In fact,
most of what has been discussed in this chapter is
based on hypothesized wildlife-habitat relationships
formulated from the synthesis of ecological theory
and some basic data on how different species of
wildlife react to their habitats. The result has been
called "a working hypothesis, a place to start, and a
way to derive tentative answers to questions for
which there are no certain answers" (Thomas
1979b:7). But we do have a place to start. We are
confident that, as speculation gives way to action, ex-
perience will allow improvements in the currently
primitive systems, and the art of monitoring will
evolve to a science. An overriding consideration in
all of this must be its cost, because if economical
ways cannot be found to do the job, it will not get
done.

88

Forests

LITERATURE CITED

ALLEN, AW. 1982a. Habitat suitability index models:

beaver. FWS/OBS-82/ 10.30. U.S. Dep. Inter., Fish and
Wildl. Serv. Washington, DC. 20pp.

. 1982b. Habitat suitability index models: marten.

FWS/OBS-82/ 10. 11. U.S. Dep. Inter., Fish and Wildl.
Serv. Washington, DC. 9pp.

ANDERSON, A.J.B. 1971. Ordination methods in ecology.
J. Ecol. 59:713-726.

ANDERSON, JR., E.E. HARDY, J.T. ROACH, and RE. WIT-
MER. 1976. A land use and land cover classification
system for use with remote sensor data. U.S. Dep.
Inter., Geol. Surv. Prof. Pap. 964. Reston, VA. 28pp.

AUSTIN, M.E. 1965. Land resource regions and major land
resource areas of the United States (exclusive of
Alaska and Hawaii). USDA Soil Conserv. Serv. Agric.
Handb. 296. Washington, DC. 82pp.

AUSTIN, MP. and L. ORLOCI. 1966. Geometric models in
ecology. II. An evaluation of some ordination tech-
niques. J. Ecol. 54:217-227.

BAILEY, R.G. 1976. Ecoregions of the United States (map).
U.S. Dep. Agric. For. Serv., Intermountain Region.
Ogden, UT.

. 1 977. A new map of the ecosystem regions of the

United States. Pages 1 21- 1 28 in Classification, Inven-
tory, and Analysis of Fish and Wildlife Habitat — the
Proceedings of a National Symposium. FWS/OBS-
78/76. U.S. Dep. Inter., Fish and Wildl. Serv. Washing-
ton, DC.

. 1978. Description of the ecoregions of the United

States. U.S. Dep. Agric, Forest Service, Intermountain
Region. Ogden, UT. 77pp.

. 1980. Integrated approaches to classifying land as

ecosystems, in Proc. IUFRO/ISS workshop on land
evaluation for forestry. Wageningen, The Netherlands.
1982. Classification systems for habitat and ecosys-

tems. Pages 16-26 in Mason, W.T. ed. Research on
Fish and Wildlife Habitat. Office of Research and
Development, U.S. Environmental Protection Agency.
Washington, DC.

BANNISTER, P. 1968. An evaluation of some procedures
used in simple ordinations. J. Ecol. 56:27-34.

BLACK, H. Jr., R SCHERZINGER, and J.W. THOMAS. 1976.
Relationships of Rocky Mountain elk and Rocky
Mountain mule deer habitat to timber harvest in the
Blue Mountains of Oregon and Washington. Pages
105-118 in Proceedings of the Elk-Logging-Roads
Symposium. Univ. Idaho, Moscow.

BRAY, JR. and J.T. CURTIS. 1957. An ordination of the
upland forest communities of southern Wisconsin.
Ecol. Monogr. 27:325-349.

BROWN, D.E., C.H. LOWE, and C.P. PASE. 1980. A digi-
tized systematic classification for ecosystems with an
illustrated summary of the natural vegetation of North
America. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-73. 93pp.

BROWN, H.R., ed. In Press. Wildlife habitats in managed
forests — western Oregon and Washington. U.S. Dep.
Agric, For. Serv., Pacific Northwest Region. Portland,
OR.

BUTTERY, R.F. 1978. Modified ECOCLASS— A Forest Ser-
vice method for classifying ecosystems. Pages 157-
168 in Integrated Inventories of Renewable Natural
Resources. Tucson, AZ. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-55.

CAIN, S.A. and G.M. DE OLIVERIA CASTRO. 1959. Manual
of vegetation analysis. Harper and Brothers. New
York, NY. 325pp.

CORTISS, J.C. 1974. ECOCLASS — a method for classifying
ecosystems. Pages 264-271 in Foresters in Land-Use
Planning. Proc. of the 1973 Natl Conv., Soc Am. For.
Washington, DC.

CROME, F.H.J. 1978. Foraging ecology of an assemblage of
birds in lowland rainforest in northern Queensland.
Aust. J. Ecol. 3:195-212.

CROWLEY, J.M. 1967. Biogeography. Can. Geogr. 11:312-
326.

CUSHWA, C.T. and C.W. DUBROCK. 1982. Design of com-
puterized fish and wildlife species data bases by state
and federal agencies. Pages 37-46 in Mason, W.T.
ed. Research on Wildlife Habitat. EPA-600/8-82-022.
U.S. Environmental Protection Agency. Washington,
DC.

DAUBENMIRE, R. 1968. Plant communities: a textbook of
plant synecology. Harper and Row. New York, NY.
300pp.

DEGRAAF, R.M. and N.L. CHADWICK. 1984. Habitat clas-
sification: a comparison using avian species and
guilds. Environ. Manage. 8(6):51 1-518.

DRISCOLL, R.S., J.W. RUSSELL, and M.C. MEIER. 1978.
Recommended national land classification system for
renewable resource assessments. U.S. Dep. Agric, For.
Serv. Rocky Mtn. For. and Range Exp. Sta. Fort Col-
lins, CO (unpublished manuscript). 44pp.

DUESER, R.D. and H.H. SHUGART, Jr. 1979. Niche pattern
in a forest -floor small-mammal fauna. Ecology 60:108-
118.

ELTON, C.S. 1966. Dying and dead wood. Pages 279-305
in The Pattern of Animal Communities. John Wiley
and Sons, Inc. New York, NY.

FOLSE, L.J. Jr. 1981. Ecological relationships of grassland
birds to habitat and food supply in east Africa. Pages
160-166 in Capen, D.E. ed. The Use of Multivariate
Statistics in Studies of Wildlife Habitat. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. RM-87.

GALLI, A.E., C.F. LECK, and R.T.T. FORMAN. 1976. Avian
distribution patterns in forest islands of different sizes
in central New Jersey. Auk 93(2 ):356- 364.

GARRISON, G.A., J.J. BJUGSTAD, DA. DUNCAN, M.E.
LEWIS, and DR. SMITH. 1977. Vegetation and envi-
ronmental features of forest and range ecosystems.
U.S. Dep. Agric, For. Serv. Agric Handb. 475. U.S.
Govt. Print. Off. Washington, DC. 68pp.

GAUCH, H.G. Jr. 1982. Multivariate analysis in community
ecology. Cambridge Univ. Press, Cambridge. 298pp.

GODFREY, A.E. 1977. A physiographic approach to land
use planning. Environ. Geol. 2:43-50.

GREIG-SMITH, P. 1964. Quantitative plant ecology. 2nd
ed. Butterworths, London. 256pp.

HAAPANEN, A. 1965. Bird fauna of the Finnish forests in
relation to forest succession. I. Ann. Zool. Fenn.
2(3):153-196.

. 1 966. Bird fauna of the Finnish forests in relation

to forest succession. II. Ann. Zool. Fenn. 3(3): 176-
200.

HALL, F.C. 1973- Plant communities of the Blue Mountains
in eastern Oregon and southeastern Washington. U.S.
Dep. Agric, For. Serv. Reg. 6 Area Guide 3-1. 62pp.

HARPER, JAMES A. 1969. Relationship of elk to reforesta-
tion in the Pacific Northwest. Pages 67-71 in Black,
H.C. ed. Wildlife and Reforestation in the Pacific
Northwest. School of Forestry, Oreg. State Univ.
Corvallis.

Forests

89

HENDERSON, J.A., L.S. DAVIS, and EM. RYBERG. 1978.
ECOSYM: a classification and information system for
wildland resource management. Utah State Univ.
Logan. 30pp.

HOEKSTRA, T.W., D.E. CHALK, C.L. HAWKES, and S.A.
MILLER. 1983- Monitoring regional wildlife and fish
habitats and populations for national assessments and
appraisals. Pages 308-314 in Trans. 48th North Am.
Wildl. and Nat. Resour. Conf.

, D.L SCHWEITZER, C.T. CUSHWA, S.H. ANDER-
SON, and R.B. BARNES. 1979. Preliminary evaluation
of a national wildlife and fish data base. Pages 308-
314 in Trans. 44th North Am. Wildl. and Nat. Resour.
Conf.

HOLMES, R.T., RE. BONNEY, Jr., and S.W. PACALA. 1979.
Guild structure of the Hubbard Brook bird commu-
nity: a multivariate approach. Ecology 60:512-520.

JAKSIC, F.M. 1981. Abuse and misuse of the term "guild"
in ecological studies. Oikos 37:397-400.

JENKINS, R.E. 1977. Classification and inventory for the
perpetuation of ecological diversity. Pages 41-51
in Classification, Inventory, and Analysis of Fish and
Wildlife — the Proceedings of a National Symposium,
U.S. Dep. Inter., Fish and Wildl. Serv. Washington, DC.

KUCHLER, AW. 1964. Potential natural vegetation of the
conterminous United States (map and manual). Am.
Geogr. Spec. Publ. 36. 11 6pp.

. 1970. Potential natural vegetation. (Map, 1966.)

Pages 90-91 in The National Atlas of the United States
of America. U.S. Dep. Inter., Geological Survey. Wash-
ington, DC.

LANDRES, P.B. and J.A. MACMAHON. 1980. Guilds and
community organization: analysis of an oak woodland
avifauna in Sonora, Mexico. Auk 97:351-365.

LARSON, T.A. 1981. Ecological correlates of avian commu-
nity structure in mixed-conifer habitat: an experimen-
tal approach. Ph.D. Dissertation. Illinois State Univ.
Normal.

LECKENBY, D.A. 1984. Elk use and availability of cover
and forage habitat components in the Blue Mountains,
Northeast Oregon, 1976-1982. Wildl. Res. Rep. 14.
Oregon Dep. of Fish and Wildl. Portland. 40pp.

LEOPOLD, A. 1933. Game management. Charles Scribner
Sons. New York, NY. 481pp.

LINDEMAN, R.H, P.F. MERENDA, and R.F. GOLD. 1980.
Introduction to bivariate and multivariate analysis.
Scott, Foresman and Co. Glenview, IL 444pp.

MACARTHUR, R.H. 1964. Environmental factors affecting
bird species diversity. Am. Nat. 98:387-397.

MACMAHON, J.A, D.J. SCHIMPF, DC. ANDERSON, KG.
SMITH, and R.L. BAYNJr. 1981. An organism-centered
approach to some community and ecosystem con-
cepts. J. Theor. Biol. 88:287-307.

MARMELSTEIN, A. 1977. Classification, inventory, and
analysis of fish and wildlife habitat — proceedings of a
national symposium. U.S. Dep. Inter., Fish and Wildl.
Serv. Washington, DC. 604pp.

MASER, C, J.M. GEIST, DM. CONCANNON, R. ANDER
SON, and B. LOVELL 1979a. Wildlife habitats in man-
aged rangelands — the Great Basin of southeastern
Oregon: geomorphic and edaphic habitats. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. PNW-99. 84pp.

, J.E. RODIEK and J.W. THOMAS. 1979b. Cliffs, talus,

and caves. Pages 96-103 in Wildlife Habitats in Man-
aged Forests — the Blue Mountains of Oregon and
Washington. U.S. Dep. Agric, Agric. Handb. 553- U.S.
Govt. Print. Off. Washington, DC.

— , R.G. ANDERSON, K CROMACK Jr., J.T. WILLIAMS,
and RE. MARTIN. 1979c. Dead and down woody
material. Pages 78-95 in Wildlife Habitats in Managed
Forests — the Blue Mountains of Oregon and Washing-
ton. U.S. Dep. Agric, Agric Handb. 553- U.S. Govt.
Print. Off. Washington, DC.

— and J.W. THOMAS. 1983. Wildlife habitats in man-

aged rangelands — the Great Basin of southeastern
Oregon: Introduction. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. PNW-160. 15pp.

MAWSON, J.C., J.W. THOMAS, and R.M. DEGRAAF. 1976.
Program HTVOL — the determination of crown vol-
ume by layers. U.S. Dep. Agric, For. Serv. Res. Pap.
NE-354. 9pp.

NIE, N.H., C.H. HULL, J.G. JENKINS, K STEINBRENNER,
and D.H. BENT. 1974. SPSS statistical package for the
social sciences. 2nd ed. McGraw-Hill. New York,
NY. 675pp.

PATTON, DR. 1975. A diversity index for quantifying
habitat "edge." Wildl. Soc Bull. 3(4): 171- 173-

. 1978. RUNWILD: a storage and retrieval system for

wildlife habitat information. U.S. Dep. Agric, For.
Serv. Gen. Tech. Rep. RM-51. 8pp.

PFISTER, R.D. 1977. Ecological classification of forest land
in Idaho and Montana. Pages 329-358 in Ecological
Classification of Forest Land in Canada and North-
western USA. Univ. British Columbia. Vancouver,
Canada.

RAO, C.R. 1966. Influence on discriminant function coeffi-
cients. Pages 587-602 in R.C. Rose, I.M. Chakrovarti,
P.C. Mahalanobis, C.R. Rao, and KJ.C Smith eds. Es-
says in Probability and Statistics. Univ. North Carolina
Press, Chapel Hill.

REYNOLDS, H.G. 1966. Use of openings in spruce-fir for-
ests of Arizona by deer, elk and cattle. U.S. Dep.
Agric, For. Serv. Res. Note 78, 4pp.

ROBINSON, W.L and E.B. BOLEN. 1984. Wildlife ecology
and management. MacMillan Publ. Co. New York,
NY. 478 pp.

ROOT, R.B. 1967. The niche exploitation pattern of the
blue-gray gnatcatcher. Ecol. Monogr. 37:317-350.

SALWASSER, HAL, C.K HAMILTON, W.B. KROHN, J. LIP-
SCOMB, and C.H. THOMAS. 1983. Monitoring wildlife
and fish: mandates and their implications. Pages 297-
306 in Trans. 48th North Am. Wildl. and Nat. Resour.
Conf.

SCHAMBERGER, M. and A. FARMER. 1978. The habitat
evaluation procedures: their application in project
planning and impact evaluation. Pages 274-283 in
Trans. 43rd North Am. Wildl. and Nat. Resour. Conf.

, , and J.W. TERRELL. 1982. Habitat suitability

index models: introduction. FWS/OBS-82/10. U.S.
Dep. Inter., Fish and Wildl. Serv. Washington, DC.
2pp.

SCHROEDER, R.L 1983. Habitat suitability index models:
black-capped chickadee. FWS/OBS-82/10. 37. U.S. Dep.
Inter., Fish and Wildl. Serv. Washington, DC. 1 2pp.

SEVERINGHAUS, WD. 1981. Guild theory development
as a mechanism for assessing environmental impact.
Environ. Manage. 5:187-190.

SHORT, H.L 1983. Wildlife guilds in Arizona desert habi-
tats. U.S. Dep. Inter., Bur. Land Manage. Serv. Cen.
Tech. Note 362. 258pp.

and KP. BURNHAM. 1982. Technique for structur-
ing wildlife guilds to evaluate impacts on wildlife
communities. U.S. Dep. Inter., Fish and Wildl. Serv.
Spec. Sci. Rep. 244. Washington, DC. 34pp.

90

Forests

SOCIETY OF AMERICAN FORESTERS. 1954. Forest
cover types of North America. Society of American
Foresters. Washington, DC. 67pp.

THOMAS, J. W. 1979a. Introduction. Pages 10-21 in Wild-
life Habitats in Managed Forests — the Blue Mountains
of Oregon and Washington. U.S. Dep. Agric, Agric.
Handb. 553- U.S. Govt. Print. Off. Washington, DC.

ed. 1979b. Wildlife habitats in managed forests —

the Blue Mountains of Oregon and Washington. U.S.
Dep. Agric, Agric. Handb. 553- U.S. Govt. Print. Off.
Washington, DC. 510pp.

. 1982. Species/habitat relationships — a key to con-
sidering wildlife in planning and land management
decisions. Pages 27-36 in Mason, W.T. ed. Research
on Wildlife Habitat. EPA-600/8-82-022. U.S. Environ-
mental Protection Agency. Washington, DC.

, R.M DEGRAAF, and J.C MAWSON. 1977. Determi-
nation of habitat requirements for birds in suburban
areas. U.S. Dep. Agric, For. Serv. Res. Pap. NE-357.
15pp.

, R.J. MILLER, C. MASER, R.G. ANDERSON, and B.E.

CARTER. 1979a. Plant communities and successional
stages. Pages 22-39 in Wildlife Habitats in Managed
Forests — the Blue Mountains of Oregon and Washing-
ton. U.S. Dep. Agric, Agric. Handb. 553. U.S. Govt.
Print. Off. Washington, DC.

, R.G. ANDERSON, C MASER, and EL. BULL. 1979b.

Snags. Pages 60-77 in Wildlife Habitats in Managed
Forests — the Blue Mountains of Oregon and Washing-
ton. U.S. Dep. Agric, Agric. Handb. 553- U.S. Govt.
Print. Off. Washington, DC.

, H. BLACK Jr., R.J. SCHERZINGER, and R.J. PEDER

SEN. 1979c Deer and elk. Pages 104-127 in Wildlife
Habitats in Managed Forests — the Blue Mountains
of Oregon and Washington. U.S. Dep. Agric, Agric.
Handb. 553. U.S. Govt. Print. Off. Washington, DC.

, C. MASER, and J. RODIEK. 1979d. Edges. Pages 48-

59 in Wildlife Habitats in Managed Forests — the Blue
Mountains of Oregon and Washington. U.S. Dep.

Agric, Agric Handb. 553. U.S. Govt. Print. Off. Wash-
ington, DC.

, , and . 1979c Riparian zones. Pages

40-47 in Wildlife Habitats in Managed Forests — the
Blue Mountains of Oregon and Washington. U.S. Dep.
Agric, Agric Handb. 553. U.S. Govt. Print. Off. Wash-
ington, DC.

U.S. DEPARTMENT OF INTERIOR, BUREAU OF LAND
MANAGEMENT. 1978. Integrated habitat inventory
and classification system, BLM Manual Section 6602.
Washington, DC. 37pp.

U.S. DEPARTMENT OF INTERIOR, FISH AND WILD-
LIFE SERVICE. 1980. Habitat evaluation procedures
(HEP). ESM 121. U.S. Dep. Inter., Fish and Wildl. Serv.
Washington, DC.

VERNER, J. 1983- An integrated system for monitoring
wildlife on the Sierra National Forest. Pages 355-366.
in Trans. 48th North Am. Wildlife, and Nat. Resour.
Conf.

. 1984. The guild concept applied to management

of bird populations. Environ. Manage. 8( 1):1-14.

and A.S. BOSS. 1980. California wildlife and their

habitats: western Sierra Nevada. U.S. Dep. Agric, For.
Serv. Gen. Tech. Rep. PSW-37. 439pp.

WERTZ, W.A. and J.F. ARNOLD. 1972. Land systems inven-
tory. U.S. Dep. Agric, For. Serv., Intermountain Re-
gion. Ogden, UT. 12pp.

WHITCOMB, R.F., C.S. ROBBINS, J.F LYNCH, B.L. WHIT-
COMB, M.K. KLIMBIEWICZ, and D. BYSTRAK 1981.
Effects of forest fragmentation on avifauna of the
eastern deciduous forest. Pages 125-205 in Burgess,
R.L. and DM. Sharpe eds. Forest Island Dynamics
in Man-Dominated Landscapes. Springer-Verlag. New
York, NY.

WILLIAMS, B.K. 1980. Discriminant analysis in wildlife
research: theory and applications. Pages 59-71 in
Capen, D.E. ed. The Use of Multivariate Statistics in
Studies of Wildlife Habitat. U.S. Dep. Agric, For. Serv.
Tech. Rep. RM-87.

Forests

91

RANGELANDS

Henry L. Short

U.S. Fish and Wildlife Service
Western Energy and land Use Team
Ft. Collins, CO 80526

Editor's Note: Most range studies have focused on
problems of livestock management. Most wildlife-
oriented efforts have concentrated on alleviating
impacts from grazing and have focused on prob-
lems such as forage allocation between livestock
and wild ungulates.

This chapter provides techniques to inventory and
monitor habitat structure. Because structure (trees,
shrubs, etc.) on range lands is very limited, it is
extremely important to wildlife. Areas with struc-
ture are where biologists should concentrate their
efforts. Systems based on structural diversity will be
the most useful on rangelands.

"As demands have grown for the products of range-
lands, it has become obvious that such cliches as
'good range management is good wildlife manage-
ment' will no longer suffice."

— Chris Maser and Jack W. Thomas from Wild-
life Habitats in Managed Rangelands — The
Great Basin of Southeastern Oregon

INTRODUCTION

Rangelands, as discussed in this chapter, are
considered to be lands —

• where the potential natural vegetation is mostly
grass, grasslike plants, forbs, or shrubs;

• where grazing and browsing were important
influences during prehistoric times; and

• that are more suitable for management by eco-
logical than by agronomic principles (after
Schwarz et al. 1976).

Western range and Great Plains ecosystems total
over 258 million ha (646 million a.) (Table 1) in 17
western and Great Plains States. Eighty-six percent
is grazed by domestic livestock. More than 1 36 mil-
lion kg ( 300 million t ) of herbage and browse are
produced annually on these rangelands, providing
nearly 149 million Animal Unit Months (AUMs) of
domestic livestock use (Table 1 ). The approximate
location of the potential natural vegetation types that
comprise rangelands is illustrated in Figure 1.

About 65% of the rangeland area listed in Table
1 occurs in 1 1 western states where shrublands,
mountain grassland, and desert grasslands predomi-
nate. These 167 million ha (418 million a.) provide
only about 48% of the herbage and browse and 38%
of the AUMs produced in the western range and
Great Plains ecosystems (Table 1). The Bureau of
Land Management administers about 68.4 million ha
(171 million a.) within 11 western states and author-
izes the use of about 1 2 million AUMs of forage for
3.5 million cattle and 4.9 million sheep (U.S. Depart-
ment of Interior, Bureau of Land Management 1975).

Rangelands are vast areas that provide great
diversity and present significant problems to man-
agers. Many of the western rangelands are fragile, oc-
curring on poorly developed, shallow, and highly
alkaline soils in arid or semiarid climates with a

Rangelands

93

Table 1. Production of forage and Animal Unit Months (AUMs) in 1970 by ecosystem type (after U.S. Forest
Service 1972).

Herbage and

browse produc-

Animal Unit

Total land area

tion by ecosys-

Month production

by ecosystem

tem (thousands

Total land area

by ecosystem

(thousands

of tons per

grazed (thou-

(thousands of

Ecosystems

of acres)

year)

sands of acres)

AUMs)

Sagebrush

94,219

35,840

90,453

10,850

Desert shrub

86,043

9,491

58,762

1,742

Southwestern shrub-steppe

38,601

3,526

35,382

1,958

Chaparral-mountain shrub

32,801

11,559

30,118

1,957

Pinyon-juniper

42,677

4,934

34,163

1,715

Mountain grasslands

79,839

60,858

75,206

21,441

Mountain meadows

4,045

5,482

3,785

4,309

Desert grasslands

26,098

6,481

26,098

5,073

Annual grasslands

6,700

6,886

6,696

7,003

Alpine

8,322

97

162

33

Shinnery

2,004

927

1,982

456

Texas savanna

15,221

8,352

14,870

5,042

Plains grasslands

173,260

91,408

163,709

50,454

Prairie

37,533

54,252

36,539

36,814

Total

646,643

300,092

557,925

148,848

short growing season. Only 1 8% of the western
range acreages are considered in good condition,
whereas 50% are rated fair and 32% are rated poor
(U.S. Forest Service 1972). The U.S. Bureau of Land
Management considers 1 7% of its land in satisfactory
or better condition; 83% produces vegetation at a
rate less than its potential (U.S. Bureau of Land Man-
agement 1975). Forty-eight percent of the ranges
in the more mesic Great Plains are considered in
good condition, 45% in fair condition, and only 7%
in poor condition (U.S. Forest Service 1972). Range-
lands in the 1 1 western states provide a variety of
products in addition to forage. Rangeland surveys
evaluate —

• the current year's production of herbage and
browse;

• the yield of water in acre -feet per acre per year;

• the quantity of sediments deposited in stream
channels per year;

• changes in the fertility, structure, or drainage of
soils;

• carbon and particulate content in the atmos-
phere, as measures of air quality;

• aesthetic quality;

• changes in populations of rare and endangered
wildlife species;

• diversity of habitat for nongame birds;

• resident big game populations;

• outdoor recreation use; and

• cultural values.

Results of these surveys indicate that western
rangelands have poor soil stability; poor to fair non-
hunting recreation opportunities; fair soil quality; fair
to good air quality; fair to good habitat for carni-
vores, raptors, and rare and endangered species of
wildlife and nongame birds; and fair to good hunting
and scenic and cultural values (U.S. Forest Service
1972).

Rangeland habitats are immense and complex.
Some sort of common denominator for rangelands
seems necessary so that range surveys can provide
results that are comparable and interpretable be-
tween areas. The concept of layers of habitat appears
to be a useful denominator for comparing
rangelands.

94

Rangelands

Figure 1. Occurrence of potential natural vegetation types that comprise rangeland habitats in the western
U.S. (after Kuchler 1964).

Rangelands

95

The processes described in this chapter are
based on the assumption that the niches of individ-
ual wildlife species can be positioned within the
vertical structure of habitats. This vertical structure
can be represented in terms of habitat layers. The
abstraction of representing cover types as habitat lay-
ers provides a means for performing a variety of
quantitative assessments of habitat. For example, the
structure of several land units can be compared by
measuring the areas of the different layers of habitat
present on those land units. Monitoring the area of
habitat layers will describe the general direction
of change over time in the quantity of habitat avail-
able for the wildlife community.

The layers of habitat present on a land unit can
be mapped and measured. Wildlife species can be
associated with particular layers of habitat through
application of the species-habitat matrix, where
groups or guilds of wildlife species that share de-
pendencies on the same general structure of habitat
are formed. Maps of layers of habitat and the guilds
associated with those layers of habitat are distribu-
tion maps for wildlife guilds and species. The areas
in the distribution maps can be monitored over time
to describe trends in the quantity of habitat available
for particular wildlife guilds and their component
wildlife species. Land-use changes often can be con-
sidered in terms of impact to layers of habitat. The
impacts of those land use changes on wildlife species
can be predicted because wildlife species have been
associated with habitat layers through the wildlife
guilding process.

A layer of habitat can be thought of as providing
a large array of habitat conditions. Guilds of wildlife
species that have a dependency on a habitat layer
can be disaggregated to describe the dependency of
individual species on specific conditions along this
array or habitat gradient. This disaggregation is done
with species-habitat models (or HSI models) devel-
oped for individual wildlife species. These models
emphasize features that seem important to individual
wildlife species. These features can be measured
during inventories to predict whether or not a habi-
tat apparently is suitable for an individual species.

The absence of a wildlife species in a planning
unit during a wildlife inventory frequently can be
attributed to either the absence of a habitat layer or
inadequate habitat conditions within that layer of
habitat. It is possible, in this way, to predict why a
species is present or absent within a habitat and,
presumably, to develop remedial management prac-
tices to make that habitat more suitable.

CLASSIFICATION SYSTEMS

The vegetation in western rangelands has been
extensively studied. Sagebrush and pinyon-juniper

ecosystems make up about one-third of the area of
western rangelands (Table 1 ); each has been the
subject of recent technical symposia (Utah State Uni-
versity 1975, 1979). These symposia emphasized
the distribution of these major vegetational cover
types in the western U.S., results of autoecological
and synecological studies of the ecosystems, faunal
communities, processes (e.g., fire, mechanical, and
herbicidal techniques) for modifying the structure of
the ecosystems, impacts of ecosystem modification
on other land uses, and procedures for increasing
herbage production in modified habitats. The distri-
bution of shrubs within the U.S., uses of shrub tis-
sues, physiology and nutritive quality of shrubs,
techniques for regenerating shrubs, and the future of
shrubs in arid lands have been emphasized in other
studies (McKell et al. 1972). Additional publications
about rangelands have emphasized floral and faunal
relationships in Great Basin habitats (Harper and
Reveal 1978) and topics such as seedbed preparation
and planting and the management of seeded ranges
(Plummer et al. 1955, 1968), impacts of fire on
shrubs (Wright 1972; Wright et al. 1979), and im-
pacts of drought on grasslands (Albertson and
Weaver 1946).

Studies of rangelands that describe the structure,
composition, distribution, and ecological relation-
ships among vegetative components have been done
on a State (Nichol 1937; Costello 1944) and a re
gional basis (Weaver 1954; Weaver and Albertson
1956). A map of the natural vegetation in the U.S.
and a description of associated cover types was pub-
lished in the Atlas of American Agriculture (U.S. De-
partment of Agriculture 1936). The presumed
distribution of potential range vegetation types in
the U.S. was published in map form by Kuchler
(1964; see Figure 1). Kuchler (1964) also provided
a photographic description of different potential
rangeland types. A more intensive mapping of range
cover types in the southwestern U.S. was done by
Brown and Lowe (1980), and a text description of
range types was provided by Brown ( 1982).

Many other authors have provided specialized
descriptions of rangeland vegetation. These descrip-
tions vary from statewide assessments to intensive
treatments of small range areas. Several range vegeta-
tion classification systems are summarized in Table
2 by area of application, general type of rangeland
vegetation considered, a brief description of the
vegetative classification system, and the apparent
utility of the system in inventories and assessments
of wildlife habitat.

The references in Table 2 will be especially
helpful to persons working in unfamiliar range habi-
tats. The publications listed frequently provide maps
of the distribution of range vegetation, descriptions
of vegetation communities, and lists of wildlife spe-
cies that use different range cover types.

96

Rangelands

Table 2. Range vegetation classification systems developed for grassland, shrubland, and woodland cover
types in the western U.S. and their utility in inventories and assessments of wildlife habitat quality.

Area of
application

Type of
vegetation

Classification system

Utility in inventorying
and monitoring

Reference

Grasslands

1. Southeastern
Washington,
northeastern Or-
egon, west-
central Idaho

Fescue-wheatgrass

Descriptions of vege-
tative associations

Describes a variety
of climax communi-
ties in Pacific North-
west bunch-grass
types

Tisdale (1983)

2. Idaho

Fescue-wheatgrass

Description of vege-
tative associations
similar to those of
Tisdale (1983)

Provides map of
potential vegetation
types in Idaho

Sharp and Sanders
(1978)

3. Snake River
Canyon, Idaho

Fescue-wheatgrass

Dichotomous key to
grasslands in the
Snake River Canyon

Lists specific habitat
criteria for vegetative
association

Tisdale (1979)

4 Southwestern
U.S.

Alpine and subalpine
grasslands

Description of vege-
tative associations

Text, in conjunction
with map (Brown and
Lowe 1980), pro-
vides a description
of the location and
structure of this
vegetative commu-
nity in the Southwest

Brown (1982)

5. Southwestern
U.S.

Plains and Great
Basin grasslands

Same as above

Same as above

Brown (1982)

6. Southwestern
U.S.

Semidesert grass-
land

Same as above

Same as above

Brown (1982)

7. Southwestern
U.S.

California valley
grassland

Same as above

Same as above

Brown (1982)

8. Southwestern
U.S.

Sonoran savanna
grassland

Same as above

Same as above

Brown (1982)

9. California

Grasslands

Dichotomous key to
grasses; illustrations
of common species

Describes appear-
ance, distribution
and use of individual
species

Sampson et al.
(1951)

Shrublands

1 . Southern Califor-
nia

Chamise chaparral
and mixed chaparral

Lists of plant compo-
nents of different
vegetative types

Describes some
vegetative associa-
tions along moisture
and temperature
gradients

Mooney and Parsons
(1973)

2 Southwestern
U.S.

Oak brush (Quercus
sp.)

Description of vege-
tative associations

Text, in conjunction
with map (Brown and
Lowe 1980), pro-
vides a description
of the location and
structure of this
community in the
Southwest

Brown (1982)

3. Southwestern
U.S.

California coastal
scrub

Same as above

Same as above

Brown (1982)

4. Southwestern
U.S.

California coastal
chaparral

Same as above

Same as above

Brown (1982)

5. Southwestern
U.S.

Interior chaparral

Same as above

Same as above

Brown (1982)

6 New Mexico

Saltbush (Atnplex
sp.)

Dichotomous key,
photographs, species
descriptions, range
maps

Describes species of
Atnplex and identi-
fies their range

Wagner and Aldon
(1978)

Rangelands

97

Table 2. Range vegetation classification systems developed for grassland, shrubland, and woodland cover
types in the western U.S. and their utility in inventories and assessments of wildlife habitat quality
(continued).

Area of
application

Type of
vegetation

Classification system

Utility in inventorying
and monitoring

Reference

Shrublands (continued,

7. Intermountain
area

Sagebrush (Artemi-
sia sp), rabbitbrush
(Chrysothamnus
sp), horsebrush (Te-
tradymia sp.), and
matchbrush or
snakewood (Xantho-
cephalum sp.)

Description and
dichotomous key for
species of the four
shrubs

Describes appear-
ance, distribution,
and use of individual
species; provides
illustrations that may
be helpful in species
identification

McArthur et al.
(1979)

8. Western Mon-
tana

Sagebrush-
grasslands

Key based on un-
derstory and mid-
story components

Describes distribu-
tion and vegetative
composition of cover
types, impacts of
grazing on cover
types, and response
of cover types to
range management

Mueggler and Stew-
art (1980)

9. Intermountain
region

Sagebrush-
grasslands

Variety of taxonomic
keys based on vege-
tative characteristics

Describes species of
sagebrush; manage-
ment of sagebrush
for livestock; and the
presumed impacts
of sagebrush man-
agement on livestock

Blaisdell et al. (1982)

10. Columbia River
Basin-Idaho,
Oregon, Wash-
ington

Sagebrush-grass-
lands

Lists of midstory-
understory compo-
nents

Association of sage-
brush-grassland
complexes with
edaphic and climatic
factors

Hironaka (1979)

1 1 . Northern Great
Plains

Sagebrush
species

Lists ecological
relationships of some
common sagebrush
species and associ-
ates

Describes factors
contributing to the
distribution of sage-
brush species in
Montana, Wyoming,
and the Dakotas

Johnson (1979)

12. Blue Mountains
in eastern Ore-
gon and south-
eastern
Washington

Sagebrush-domi-
nated rangelands

Describes environ-
mental and edaphic
conditions character-
istic of range type

Lists species compo-
nents, range appear-
ance, and criteria
for describing range
condition

Hall (1973)

13. Western U.S.

Sagebrush

Dichotomous key

Photographs of
herbarium types and
distribution maps of
12 sagebrush taxa

Beetle (1960)

14 Southern Idaho

Sagebrush-
grasslands

Dichotomous key

Describes wildlife
use of habitat types
and provides sug-
gestion for manage-
ment of habitats

Hironaka et al.
(1983)

15. Oregon

Sagebrush

Diagrammatic key

Describes habitats
where different sage-
brush species occur

Winward (1980)

98

Rangelands

Table 2. Range vegetation classification systems developed for grassland, shrubland, and woodland cover
types in the western U.S. and their utility in inventories and assessments of wildlife habitat quality
(concluded).

Area of
application

Type of
vegetation

Classification system

Utility in inventorying
and monitoring

Reference

Woodlands

1. Southwestern
US

Pinyon-juniper

Descriptions of vege-
tative associations

Text, in conjunction
with map (Brown and
Lowe 1980), pro-
vides a description
of the location and
structure of the
pinyon-juniper com-
munity in the South-
west

Brown (1982)

2 Southwestern
U.S.

Oak-woodlands

Same as above

Same as above

Brown (1982)

MAJOR SPECIES GROUPS

Western range ecosystems provide habitat for at
least 10 rare and endangered terrestrial vertebrate
wildlife species (U.S. Forest Service 1972):

• Utah prairie dog {Cynomys parvidens)

• Grizzly bear (Ursus arctos horribilis)

• Yuma clapper rail (Rallus longirostris yuma-
nensis)

• Light-footed clapper rail (R 1. levipes)

• Masked bobwhite (Colinus virginianus ridg-
wayi)

• California condor {Gymnogyps calif ornianus)

• Sonoran pronghorn (Antilocapra americana
sonoriensis)

• San Joaquin kit fox (Vulpes macrotis mutica)

• Black-footed ferret (M us tela nigripes)

• Whooping crane (Grus americana)

The use of rangeland habitats by terrestrial verte-
brate wildlife species is probably mostly dependent
on the structure of the rangeland vegetative cover.
For example, treeland habitats generally support a
more diverse wildlife community than do shrubland
or grassland habitats, and shrubland habitats fre-
quently support a more diverse wildlife community
than do grassland habitats. Other factors influencing
the diversity of the wildlife community on a range-
land include —

• latitude;

• climate;

• topography;

• distance to water;

• distance to cliff faces, caves, or other features;

• density, height, and vigor of range vegetation;
and

• presence or absence of particular plant species.

Many of these factors are described in greater detail
in the following section.

CRITICAL HABITAT FEATURES

Range surveys are often conducted to describe
soil profiles and to estimate ( 1 ) the potential natural
vegetation that can occur on a range habitat, (2)
the condition of range vegetation, and (3) the wild-
life community that is present.

Range surveys, until World War II, usually were
conducted entirely by ground methods, such as tra-
versing compass lines across landscapes and sketch-
ing in planimetric and vegetation details (Poulton
1970). Cost effectiveness and usefulness of range
surveys were enhanced by the widespread accept-
ance of interpretive aerial photography. Interpreta-
tion of aerial photographs, combined with field
surveys, offers an increased capability to evaluate
large land areas to —

( 1 ) identify many plant species and soil surface
features;

(2) classify surface features into specific habitat
types;

Rangelands

99

(3) measure vegetation characteristics such as
cover, density, height, and spatial distribution;
and

(4) monitor changes in vegetation and soil sur-
face conditions over time.

Samples of current annual growth collected
during range surveys frequently are subjected to
chemical, calorimetric, and digestive analyses to de-
velop estimates of range carrying capacity for domes-
tic livestock and selected wildlife species. Range
conditions often are evaluated by determining if the
plants that characteristically should grow in the area
are present and in good vigor (Stoddard et al. 1975).
Surveys to determine the presence of wildlife in
range cover types usually involve on-site assessments
with standardized trapping techniques, especially
for small mammals and reptiles, and census lines for
avian surveys or for locating pellet group plots for
large mammals. Animal survey techniques sometimes
are supplemented with interpreted aerial photo-
graphs that show, for example, waterfowl on lakes,
deer and elk against snowfield, and domestic live-
stock on range.

Rangelands are complex ecosystems that can be
studied at many levels of detail. Inventories can be
very intensive and applicable to small areas or be
extremely general, providing limited information to
managers. It seems desirable, in this chapter, to pro-
vide a general framework for conducting surveys
to inventory rangelands and monitor their wildlife
values. This framework is a conceptual representa-
tion of the structure of rangeland within which gen-
eral survey information applicable to large land areas
can be gathered and analyzed in a timely manner
and within which very specific information also can
be gathered. I have not included extensive lists of
range variables and techniques for their measure-
ment; this information is available elsewhere. Instead,
I have provided a general inventory process that —

is universally applicable to a wide variety of
terrestrial rangeland types;

considers all the vertebrate wildlife species that
might occur on a rangeland;

allows a manager to predict the impacts that
land use change might have on wildlife species;
and

uses state-of-the-art technologies to provide
quantitative appraisals of rangelands so that
rangeland resources can be monitored through
time.

This inventory process includes a way to abstractly
represent the structure of range vegetation, associate
wildlife species with the structure of range vegeta-
tion, and relate some types of inventory information
to range management.

It is unrealistic to consider the total wildlife
community in inventories and assessments of range-
lands, unless some general criteria can be identified
that determine how wildlife species use habitats and
are easily measured. Layers of habitat, discussed be-
low, is an easily-measured criterion that is a critical
habitat feature for terrestrial vertebrate species.

Any terrestrial habitat, including a rangeland,
can be considered a volume of space that contains
structure, such as vegetation which may be useful to
wildlife species. The volume of space in a mixed
rangeland habitat containing shrubland, grassland,
and treeland components is represented in Part A of
Figure 2. The structure of this rangeland habitat also
can be considered in terms of habitat layers, as in
Part B of Figure 2, which can be used to inventory
and evaluate rangeland habitats. The different habitat
layers are defined in Short and Burnham (1982)
and in Table 3.

B

Midstory

Midstory

Tree bole

Understory

rflftiii&KkA

Terrestrial subsurface

Figure 2. Cover types within a rangeland habitat can be considered in terms of layers of habitat.
100 Rangelands

Table 3. Suggested criteria for determining the presence of different layers of habitat.

Layer

Criteria

Tree canopy or overstory

Vegetation structure is 8 m (25 ft) or more above the terrestrial or aquatic sur-
face and provides at least 5% cover when projected to the surface (500 m2/
ha or 2,200 ft2/a.).

Tree bole

Tree trunks have a dbh 20 cm (8 in.) and occur at a density 12/ha (5/a).

Shrub midstory

Vegetation height from 50 cm (20 in.) up to 8 m (25 ft), which provides at least
5% cover when projected to the surface (500 m2/ha or 2,200 ft2/a.).

Understory

Layer extends from 10 cm (4 in.) below the apparent surface up to, but not
including, 50 cm (20 in.) above the apparent surface and provides at least 5%
cover when projected to the surface (500 m2/ha or 2,200 ft2/a).

Terrestrial subsurface

Extends from more than 10 cm (4 in.) below the apparent surface down.

Surface water layer

Land surface-water interface and shallow water up to 25 cm (10 in.) deep.

The surface water layer includes the shoreline
and shallow water (< 25cm [10 in.] deep) areas and
the terrestrial substrate that is under the shallow
water. Vegetation that emerges through shallow
water and is less than 0.5 m (1.6 ft) high is consid-
ered a product of the surface water layer, as is float-
ing vegetation in shallow water.

The terrestrial subsurface layer extends upward
to within 10 cm (4 in.) of the apparent surface and
includes caves and deep crevices that are critical
habitat features for some wildlife species and a soil
substrate suitable for the construction of burrows
and dens.

The terrestrial surface layer extends from 10 cm
(4 in.) below to 0.5 m (1.6 ft) above the apparent
surface. This layer includes bare ground, talus, cliff
faces, litter, herbaceous vegetation, and dwarf or
supine woody vegetation as habitat features.

The shrub or midstory layer extends from 0.5 to
8.0 m (1.6 to 26.4 ft) above the apparent surface
and includes the canopies of shrubs and dwarf trees,
as well as other vegetative plant parts that extend
into this strata.

The tree bole layer consists of live and dead
tree trunks that provide both a foraging substrate
and a variety of nest substrates, including bark and
cavities. Tree boles comprise a vertical feature that
extends through horizontal layers. Snags are special-
ized tree boles.

The tree canopy extends above 8.0 m (26.4 ft)
and provides large and small branches of living and
dead trees as nest substrates and includes leaves,
buds, and fruit as food for wildlife species.

Considering habitat in terms of layers has sev-
eral advantages. The concept seems intuitively cor-

rect, insofar as the presence of layers is readily
perceivable, even though attempts to define the lay-
ers may seem arbitrary (Table 3). Layers of habitat
can be mapped, wildlife species can be identified
with particular habitat layers, and many management
actions impact habitat layers, which can be used to
predict the impacts of land use change on wildlife.

Wildlife species can be associated with layers of
habitat because the niche of an individual species
occurs within one or more habitat strata. Grasslands
contain only the terrestrial subsurface and terrestrial
surface layers, defined above. Shrublands provide
terrestrial subsurface, terrestrial surface, and mid-
story layers; woodland habitats may also provide tree
boles for cavity makers and users and a tree canopy
layer. Rangeland habitats that provide only a few
layers of habitat have a limited volume of space
within which wildlife species can find niches. More
niches are potentially available as more layers of
habitat occur in cover types, so more wildlife spe-
cies potentially are supported by more structurally
complex habitats.

Rangelands can be described in terms of habitat
layers. For example, Kuchler (1964) listed 27 poten-
tial natural vegetation types in the western U.S. that
can be considered rangelands. These types are ar-
rayed in Figure 3 according to the number of habitat
layers that each type provides for the wildlife com-
munity. Considerable similarity exists in the struc-
ture of the wildlife communities that occur in the 14
grassland-dwarf shrub cover types, which contain
terrestrial subsurface and terrestrial surface layers of
habitat. Great similarity also is evident in the struc-
ture of the wildlife communities that occur in the
seven shrubland cover types, which contain terres-
trial subsurface, terrestrial surface, and midstory
layers of habitat. The similarities in the wildlife com-
munities within grasslands and the communities
within shrublands occur even when the vegetation

Rangelands

101

Sandsage-bluestem prairie

Bluestem-grama prairie

Wheatgrass-needlegrass

Grama-buffalograss

Grama-needlegrass-wheatgrass

Foothills prairie

Galleta-three awn shrubsteppe

Wheatgrass-needlegrass-
shrubsteppe

Grama-galleta steppe

Wheatgrass-bluegrass

Fescue-wheatgrass

California steppe

Blackbrush

Fescue-oatgrass

Juniper steppe woodland
Sagebrush steppe
Saltbush-greasewood
Great Basin sagebrush
Mountain mahogany-oak scrub
Coastal sagebrush
Chaparral

Oak-juniper woodland
Juniper-pinyon woodland
Mesquite-savanna
Mesquite-acacia savanna

Mesquite-live oak savanna
Montane chaparral

Figure 3. Rangeland cover types provide different habitat layers for the wildlife community.

1 02

Rangelands

species differ among grasslands or among shrublands.
Certain broad habitat changes within grasslands or
within shrublands are expected to affect similar func-
tional segments of the wildlife community in a pre-
dictable manner. The capability to predict impacts
on the wildlife community resulting from changes in
a cover type is important to resource managers.

Layers of habitat can be mapped and used to
describe potential wildlife habitat values in a plan-
ning unit. As the quantity of a habitat layer changes,
the quantity of habitat available for species depend-
ent on that habitat layer also changes. Thus, a meas-
ure of the structure of a rangeland provides a first
estimate of the utility of that rangeland as habitat for
wildlife species.

Habitat Layer Index (HLI)

The interpretation of aerial photography has
become an important tool in range surveys because
it is the best way to represent the physical relation-
ships between surface cover features. For example,
aerial photographs have often been used to describe
where shrubland, grassland, and treeland cover
types, depicted in Figure 2A, occur within a study
area. An aerial photograph that has been interpreted
to describe cover types is provided in Figure 4,
which indicates the location and identification of
surface cover types on a selected land area within
the Piceance Basin of northwestern Colorado. The
photograph has been interpreted according to the

423PJBS

212

522SG

523SGHR

522SG

523SGHR

Figure 4. The cover type polygons in this figure were traced from an interpreted aerial photograph of a

small portion of the Piceance Basin in northwestern Colorado. Identification of the vegetation within indi-
vidual polygons can be made from vegetation classes listed in Table 5. Segments A and B are described
in detail in the text.

Rangelands

103

hierarchical classification system listed in Table 4,
coupled with ground truthing, which is the on-site
examination of photo points to correlate images on
the photograph with cover features on the ground.

Maps of cover types, like that in Figure 4, can
provide useful information to persons trying to in-
ventory and assess wildlife habitat on rangelands
because some wildlife species, within their distribu-
tion range, are closely associated with the habitat
structure provided by specific cover types. For ex-
ample, species like sage grouse (Centrocercus uro-
phasianus), Brewer's sparrows (Spizella breweri),
sage sparrows (Amphispiza belli), and sagebrush
voles (Lagurus curtatus) may be dependent on the
presence of sagebrush (Artemesia sp. ) habitat,
whereas bushtits (Psaltriparus minimus), ash-
throated flycatchers (Myiarchus cinerascens), gray
vireos (Vireo vicinior), black-throated gray warblers
(Dendroica nigrescens), and pinyon mice (Peromys-
cus truei) may be dependent on the presence of
pinyon-juniper (Pinus-Juniperus sp.) and oak brush
(Quercus sp. ). Lark buntings (Calamospiza melano-
corys), horned larks (Eremophila alpestris), chest-
nut-collared longspurs (Calcarius ornatus),
McCown's longspurs (Calcarius mccownii), western

meadowlarks (Sturnella neglecta), Baird's sparrows
(Ammodramus bairdii), swift foxes (Vulpes velox),
northern grasshopper mice (Onchomys leucogaster),
plains harvest mice (Reithrodontomys montanus),
Ord's kangaroo rats (Dipodomys ordii), thirteen-
lined ground squirrels (Spermophilus tridecemlinea-
tus), and pronghorns (Antilocapra americana),
may be dependent on the presence of prairie grasses.

Interpreted aerial photographs can be used to
determine the presence of other important habitat
features like high cliffs and cliff faces, which may be
habitat for peregrine falcons (Falco peregrinus),
prairie falcons {Falco mexicanus), black swifts (Cyp-
seloides niger), white-throated swifts (Aeronautes
saxatalis), cliff swallows (Hirundo pyrrhonota), and
canyon wrens (Catherpes mexicanus); boulder
fields, talus slopes, and rock slides in mountainous
areas, which may be habitat for pikas (Ochotona
princeps); caves, which may be habitat for a variety
of bats; and prairie dog towns, which may be habitat
for black-tailed prairie dogs (Cynomys ludovici-
anus), black-footed ferrets, and burrowing owls
(Athene cunicularia); and water sources essential to
the reproduction of amphibians and a life requisite
for numerous other species. The absence of these

Table 4. Hierarchical classification system for interpreting aerial photographs, developed for the Piceance
Basin of Northeastern Colorado.

Level 1

Level 2

Level 3

Level 4

Level 5

1. Urban or

built-up land

1. Commercial

2. Residential

1. Transportation
(update of
USGS
maps only)

2. Agricultural/
reclaimed
land

1. Agricultural

2. Reclaimed land

1. Irrigated crops

2. Nonirrigated

crops

3. Orchards,

groves,
nurseries

4. Other

1. Reclaimed mine

land

2. Other reclaimed

land

3. Chained land

4. Other

3, Wetlands

(interpreted

according to

National

Wetland

Inventory

Standards)

Percent cover of
dominant cover
type

Dominant cover

Subdominant
cover

104

Rangelands

Table 4. Hierarchical classification system for interpreting aerial photographs, developed for the Piceance
Basin of Northeastern Colorado (continued).

Level 1

Level 2

Level 3

Level 4

Level 5

4. Forest (>20%

1. Deciduous

1.

Closed 80 + %

AS Aspen

BS Bare soil

trees)

2.

Open 50 to 80%

UD Upland

BR Bare rock

3.

Sparse 20 to

deciduous

HR Herbaceous

50%

OT Other

SG Sagebrush
OB Oakbrush
US Upland shrub
HA Halophytic

shrub
OT Other

2. Evergreen

1.

Closed 80 + %

PD Ponderosa

BS Bare soil

2.

Open 50 to 80%

pine

BR Bare rock

3.

Sparse 20 to

JN Juniper

HR Herbaceous

50%

PY Pinyon pine
DF Douglas fir
SP Spruce sp.
SA Subalpine fir
LP Lodgepole

pine
PJ Pinyon pine/

juniper assn.
SF Spruce/fir assn.
CN Coniferous
OT Other

SG Sagebrush
OB Oakbrush
US Upland shrub
HA Halophytic

shrub
OT Other

3. Mixed

1.

Closed 80 + %

AS Aspen

BS Bare soil

deciduous/

2.

Open 50 to 80%

PD Ponderosa

BR Bare rock

evergreen

3.

Sparse 20 to

pine

HR Herbaceous

(>33% de-

50%

JN Juniper

SG Sagebrush

ciduous and

PY Pinyon pine

OB Oakbrush

>33% ever-

DF Douglas fir

US Upland shrub

green)

SP Spruce sp.
SA Subalpine fir
LP Lodgepole

pine
PJ Pinyon pine/

juniper assn.
SF Spruce/fir assn.
UD Upland

deciduous
CN Coniferous
OT Other

HA Halophytic

shrub
OT Other

5. Shrubland

1. Deciduous

1.

Closed 80 + %

OB Oakbrush

BS Bare soil

(>20%

2.

Open 50 to 80%

US Upland shrub

BR Bare rock

shrubs

3.

Sparse 20 to

OT Other

HR Herbaceous

and <20%

50%

OT Other

trees)

2. Evergreen

1.

Closed 80 + %

SG Sagebrush

BS Bare soil

2

Open 50 to 80%

SL Saltbrush

BR Bare rock

3.

Sparse 20 to

GR Greasewood

HR Herbaceous

50%

HA Halophytic

shrub
OT Other

OT Other

3. Mixed

1.

Closed 80 + %

SG Sagebrush

BS Bare soil

deciduous/

2.

Open 50 to 80%

OB Oakbrush

BR Bare rock

evergreen

3.

Sparse 20 to

SL Saltbrush

HR Herbaceous

(>33% de-

50%

GR Greasewood

OT Other

ciduous and

US Upland shrub

>33% ever-

HA Halophytic

green)

shrub

^__

OT Other

Rangelands

105

Table 4. Hierarchical classification system for interpreting aerial photographs, developed for the Piceance
Basin of Northeastern Colorado (concluded).

Level 1

Level 2

Level 3

Level 4

Level 5

6. Grassland

(>20% grass
and <20%
trees and
<20% shrubs)

1. Plains

grassland

2. Mountain

grassland

1 . Closed 80%

2. Open 50 to 80%

3. Sparse 20 to

50%

1. Closed 80 + %

2. Open 50 to 80%

3. Sparse 20 to

50%

TL Tallgrass
SH Shortgrass
BN Bunchgrass
DS Desert grass
AN Annual grass
OT Other

TL Tallgrass
SH Shortgrass
BN Bunchgrass
DS Desert grass
AN Annual grass
OT Other

BS Bare soil
BR Bare rock
OT Other

BS Bare Soil
BR Bare rock
OT Other

7. Barren land
(>40% bare
ground and
<20% trees
and <20%
shrubs and
<20%
grasses)

1. Natural

2. Man-made

1. Bare exposed
rock

2 Bare exposed

soil

3. Spires/cliffs

1. Quarries

2. Strip mine

3 Gravel pit

4. Mine spoils

5. Nonvegetated

HR Herbaceous
SG Sagebrush
US Upland shrub
HA Halophytic

shrub
PJ Pinyon pine/

juniper assn.
SF Spruce/fir assn.
AS Aspen
CN Coniferous
OT Other

8. Tundra

1. Shrub and

brush tundra

2. Herbaceous

tundra

3. Bare ground

tundra

4. Wet tundra

5. Mixed tundra

9. Perennial snow
or ice

1 . Perennial

snowfield

2. Glacier

important habitat features in a planning unit or study
site suggests that wildlife species dependent on the
structure of these physical features will be uncom-
mon or absent.

Many wildlife species seem to utilize a habitat
structure more general than that provided by a sin-
gle plant species. Consequently, it may be easier
to produce maps like that in Figure 4 than to deter-
mine what the maps really mean in terms of species
presence. For example, it is difficult to list the wild-
life species expected to use each of the polygons
in Figure 4 or to compare the structure of the habi-
tats in segments A and B. However, the concept of
habitat layers provides a framework for interpreting

the usefulness of the structure of such habitats to the
wildlife community.

HLI provides a way to quantitatively describe
the relative structural complexity of the vegetative
cover occurring in a study area. It provides a num-
ber between 0.0 (no structural diversity) and 1.0
(high structural diversity) that can be used to char-
acterize vegetative diversity on a parcel of land,
compare vegetative diversity on several parcels of
land, or provide a basis for statistical assessments
that show direction and rate of change in habitat
structure over time. The index does not predict the
suitability of habitats for individual wildlife species,
although wildlife species that tend to occur in struc-

106

Range-lands

;.

turally complex habitats may occur where HLIs are
high, whereas other wildlife species requiring simple
habitats may occur where HLIs are low. High HLIs
should not be equated to "good" habitat and low
HLIs to "poor." A grassland, for example, might pro-
vide excellent habitat for grassland wildlife species
and still receive a low HLI because of its limited
structural diversity. Disturbance to a grassland could
result in the invasion of shrub species, which might
produce increased structural diversity and a high
HLI, but result in a poor quality habitat for grassland

wildlife species. The HLI would be considered sim-
ply as a way to numerically represent change in
structure in order to enhance habitat assessments.

The cover types depicted in Figure 2A were
represented in terms of layers of habitat in Figure
2B. Likewise, the aerial photograph of the land area
within the Piceance Basin can be represented in
terms of cover types ( Figure 4 ) or in terms of layers
of habitat. This interpretation (Figure 5) is based

Two layers of habitat

Three layers of habitat

Four layers of habitat

Figure 5. Aerial photograph of the land area represented in Figure 4 has been reinterpreted to indicate the
number of layers of habitat present in different polygons.

Rangelands

107

on the assumption that the niche of a wildlife spe-
cies can be associated with the structure of habitat,
represented in terms of layers of habitat. Figure 5
identifies the number of habitat layers present in
each polygon, but not the identity of the individual
layers. This information is sufficient to calculate the
HLI.

The determination of the areas of the tree can-
opy, tree bole, shrub midstory, understory, subsur-
face layer suitable for the establishment of burrows,
and the water surface (if it exists) on a study area
constitutes an inventory of habitat resources on that
study area. Results of such an inventory can be ex-
pressed in terms of hectares (acres) of each habitat
layer within the study area.

Results of this type of inventory for segments A
and B of the habitat illustrated in Figure 5 are pre-
sented in Table 5. The inventory summarized in
Table 5 was accomplished by determining the layers
of habitat present in the different cover types.

The formula for calculating the HLI for a study
area is —

HLI =

HLI =

i=l

(6) (5)

n
i = i

where:

X = the number of layers of habitat present
within some bounded area,

Ai = the area of layer of habitat i within the
bounded area,

Aj = the surface area of cover type j within the
bounded area,

n = the number of different cover types present
within the bounded area,

6 = the maximum number of habitat layers that
could occur in a unit of structurally com-
plex terrestrial habitat, and

5 = the maximum number of units of area of

habitat layers that could occur within a unit
of structurally complex terrestrial habitat.

The calculation of HLIs for segments A and B of
Figure 5 is presented below. There are four layers
of habitat present in segment A (terrestrial subsur-
face, understory, midstory, and tree bole); the sum
of the area of the layers of habitat equals 836.7 ha
(2091.8 a.) (Table 5). The denominator equals six
layers of habitat and 5 x 281.1 ha (702.8 a.) of habi-
tat layers.

number of actual layers actually present x

actual area of those habitat layers

number of habitat layers potentially present

x

potential area of those habitat layers

4 x 2091.8
6 x 5 X 702.8

= 0.397 or 39.7%

Segment A contains a structural diversity measure of
39.7%. This HLI can be compared with that of other
land units to compare the structural diversity of
different study areas.

The HLI for segment B in Figure 5 is calculated
in a similar manner. The same four layers of habitat
are present in B as in A; the area of the habitat layers
in B equals 732.7 ha (18317 a.) (Table 5). The de-
nominator equals six layers of habitat and 5 x 269.4 ha
(673.6 a.).

HLI =

4 x 1831-7
6 x 5 x 673.6

0.362 or 36.2%

The HLI for segment B is reduced because a
substantial block of sagebrush habitat was cleared to
produce a dry land pasture (polygon 212 in section
B of Figure 4 ). The loss of sagebrush, which pro-
vided a midstory layer on this land unit, made this
polygon unsuitable for species with obligate ties
to sagebrush and for species requiring a midstory
structure.

Habitat Layers — Wildlife Guilds

The inventory and assessment of the structure
of habitat is one level of range analysis; the determi-
nation of the quantity of particular habitats suitable
for particular wildlife guilds is a second range evalua-
tion process. The map in Figure 5 shows where two,
three, or four habitat layers occur on the study area
in the Piceance Basin, although the particular habitat
layers present in any one polygon are not identified.
Five combinations of habitat layers actually occur
in the mapped area. Polygons identified as 212 in
Figure 4 provide subsurface and understory layers of
habitat (Table 5). Polygons identified as 522 SG pro-
vide subsurface and midstory layers; polygons identi-
fied as 523 SGHR and 522 SGHR provide subsurface,
understory, and midstory layers. Polygons identified
as 423 PJBS provide subsurface, midstory, and tree
bole layers; and polygons identified as 423 PJSG pro-
vide subsurface, understory, and midstory, and tree
bole layers of habitat. These different habitat struc-
tures may vary in their suitability for wildlife species.

The location of each of the different habitat
structures can be represented on a map. Data for
drawing such maps can be obtained from terrestrial

108

Rangelands

Table 5. Areas (acres) of each cover type and individual layers of habitat present in habitat segments A and
B of Figure 5.

Segment

Cover type
(Figure 4)

Individual layers of habitat present

Total

Sub-
surface

Under-
story

Mid-
story

Tree
bole

Total area
of layers

A

212

523 SGHR
423 PJSG
522 SGHR
522 SG
423 PJBS

5.3
352.9
144.9

22.0
156.2

21.5

5.3
352.9
144.9

22.0
156.2

21.5

5.3

352.9

144.9

22.0

352.9
144.9

22.0
156.2

21.5

144.9
21.5

10.6

1058.7

579.7

66.0
312.4

64.5

Total

702.8

702.8

525.1

697.5

166.4

2091.8

B

212

523 SGHR
423 PJSG
522 SGHR
522 SG
423 PJBS

139.6

328.9

48.5

0.0

98.0

58.6

139.6

328.9
48.5

98.0
58.6

139.6

328.9

48.5

328.9
48.5

98.0
58.6

48.5
58.6

279.2
986.7
194.0
0.0
196.0
175.8

Total

673.6

673.6

517.0

534.0

107.1

1831.7

The numerator of the Habitat Layer Index is calculated by determining the product of the number of habitat layers actually
present in a study area, like that in Figure 5, and the area of the individual habitat layers present on that area. The denominator is the
hypothetical value that would be realized if all six layers of habitat (water surface, terrestrial subsurface, terrestrial surface, midstory,
tree bole, and tree canopy) occurred on the study area and five of these layers extended throughout the study area

surveys of the structure of cover types or from in-
terpretive aerial photography with ground truthing.
Each polygon in the map of habitat structure within
the Piceance Basin study area ( Figure 6 ) provides
subsurface, surface, midstory, and tree bole layers of
habitat. Wildlife species can be associated with the
structure of habitats like that represented in Figure 6
through the formation of wildlife guilds (described
below). Maps of habitat structure and lists of wildlife
guilds dependent on that habitat structure will show,
for example, that the notated polygons in Figure 6
constitute a distribution map of suitable habitat for
cavity-using species of wildlife in this study area. The
habitat layer-wildlife guild maps help explain the
distribution of wildlife species throughout the range-
land and provide a rationale for inventorying habitat
conditions suitable for groups of wildlife species.

The association of wildlife species with polygons
of habitat can be accomplished with the guilding
procedure designed by Short and Burnham ( 1982),
as modified by Short (1983). The vertical layers of
habitat are represented in that procedure as axes of a
two-dimensional matrix. The x-axis of the matrix
represents layers of habitat where nesting, birthing,
or hatching occur; the y-axis of the matrix repre-
sents habitat layers where wildlife species forage for
food. This type of "species-habitat" matrix is illus-
trated in Figure 7. The y-axis of the matrix is divided
so that layers where primary consumption (plant
materials) occurs can be differentiated from habitat

layers where secondary consumption ( animal mate-
rials) occurs. The process described below produces
both guilds of primary consumers and guilds of sec-
ondary consumers. Omnivorous species occupy two
guilds, one as a primary consumer and one as a sec-
ondary consumer.

The blocks formed by the intersection of the
lines that differentiate the layers of habitat on the x-
and y-axes of the matrix are called "guild blocks."
The species-habitat matrices for the variety of range-
land habitats are different because the number of
guild blocks actually available for use by members of
the wildlife community varies among range cover
types. The guild blocks available to wildlife in the
three basic range cover types are indicated in Figure
7 for grassland, shrubland, and woodland habitats.
Woodland habitats obviously provide more guild
blocks than a shrubland habitat which, in turn, pro-
vides more guild blocks than a grassland habitat.

Guild blocks represent the different ways in
which wildlife species utilize the resources of habi-
tats. For example, wildlife species in grassland habi-
tats (Figure 7) can breed in the terrestrial subsurface
layer and forage in the subsurface layer, on the ter-
restrial surface, or in the air. Or they can breed on
the terrestrial surface and feed underground, on the
terrestrial surface, or in the air. Wildlife species can
breed in other habitats and forage in the subsurface
layer, on the terrestrial surface, or in the air above

Rangelands

109

Figure 6. Map of segments A and B of the Piceance Basin study area showing polygons containing subsur-
face, surface, midstory, and tree bole layers of habitat.

the grassland. Shrubland and woodland communities
are more complex because more layers of habitat are
available for the location of niche spaces for wildlife
species.

Groups or guilds of wildlife species are devel-
oped from a species-habitat data base. The data base
describes the layers of habitat where breeding (nest-
ing, birthing, or hatching) occurs, as well as the lay-
ers of habitat where foraging occurs, for each
wildlife species potentially occurring in a study area.
The information within the wildlife species data base
is processed (Short 1983) as follows: the numerical
codes that indicate the layers of habitat used for
breeding and for feeding (listed along the x- and y-
axes of Figure 7) are entered into a computer for
each vertebrate wildlife species inhabiting a cover
type. These coded data are subjected to standard
"merge-sort" routines that produce lists of species
with dependencies on the same guild block or the

same group of guild blocks. An example of the wild-
life guilds produced with this technique is provided
in Table 6, which lists the guilds of primary con-
sumers occupying chaparral rangelands habitats in
the Hualapai-Aquarius Planning Units in west-central
Arizona. This list indicates the dependency of the
wildlife community on the structure of the chaparral
habitat in this study area.

Wildlife guilds occurring in pinyon-juniper habi-
tats in the Piceance Basin could be developed in
the same way that guilds were developed in Table 6
and associated with a map of specific layers of habi-
tat like that in Figure 6. Wildlife species requiring
understory and midstory vegetation could occur in
the polygons in Figure 6, as well as in the polygons
with understory and midstory vegetation. The poly-
gons listed in Figure 6, however, would be the only
habitat suitable for cavity-making and cavity-using

110

Rangelands

10. Feeds elsewhere

G W

G W

W

w

G W

9. Air

S

S

s

S

8. Tree canopy

W

w

W

w

W

7. Tree bole

E

CO

w

w

w

w

w

c
o

6. Shrub layer

s

s

s

s

G W

G W

w

W

G W

c
o

5. Terr, surface

S

S

s

S

o

CD
C/3

4. Terr, subsurface

G W
S

G W
S

G

s

3. Water surface

2. Water column

o
o

1 . Bottom water column

c

T3

LL

8. Tree canopy

w

W

w

w

w

7. Tree bole

w

W

w

w

w

CO

6. Shrub layer

s

S

s

s

CD

E

G W

G W

w

w

G W

CO

5. Terr, surface

c

S

S

s

s

o
o

CO

4. Terr, subsurface

G W
S

G W
S

G
S

E

a.

3. Water surface

2. Water column

1 . Bottom water column

CO

CD

o

C

E

CD

o

3

o

CO

O

o

m

CD

CD

Si
co

S

CO

c
E

CD
O
CO

3

CO
CO

O

ca

CD

Q.

CD

5

CD

CO

CO

o

o

=J

CO

CO

_co
n

3

CD

O

CD

o

Q.

E
o

o

CD

CO
CD

to
CD

to
CD

o

CD

CO

o

CD

CO

"D
CD

CD

I—

o

CD

CO

5

CO

CD

1—

CD

1—

co

CD

CD

h1

CD
CO

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

Breedi

ng loc

i

Figure 7. Form of the species-habitat matrix used in the development of wildlife guilds, and use of guild
blocks by wildlife species in grassland (G), shrubland (S), and woodland (W) habitats.

Rangelands

111

Table 6. Wildlife guilds of primary consumers in chaparral habitats of west-central Arizona (from Short
1983).

Guild
No.

Guild
members

Feeding loci

Breeding loci

CD
O

re

=3
CO

n

z>

CO

H

to
CD

53
1—

CD

o
m

D

go
to

to
CD

CD

CD
>.

_C0

>.
O
CO

"O

E
o
n

SZ

C/)

CD
O

n

CD
CD

h1

CL

o

c

CO

o

CD
CD

<

CD
CD

_C

S

CD
CO
CD
00

"O
CD
CD

m

CD
O

ra

=i

CO

n

CO

to

CD

CD

H

CD
O
cfl

3
CO

to
to

CD

CD

1—

CD

a
>.

o

to
"D

E
o
n

sz
CO

CD
O

n

CD
CD

h1

>s

CL

o

c

CO

o

CD
CD

1

Black bear
Collared peccary
Deer mouse
Cactus mouse
Coyote

•
•
•
•
•

2

Striped skunk
Hog-nosed skunk

•
•

3

Botta's pocket
gopher

•

•

4

White-throated
wood rat

•

•

•

•

5

Bighorn sheep
Gray fox

Stephen's woodrat
Ringtail
Brush mouse

•
•
•
•
•

6

Harris' antelope
squirrel

•

7

Brown-headed

cowbird
House finch
Mockingbird
Cactus wren
Costa's hummingbird
Western harvest

mouse
Crissal thrasher

•

8

Mule deer
Rock squirrel
Cattle

9

Scrub jay
Bushtit

•
•

10

Scott's oriole
American robin
Lesser goldfinch
Townsend's solitaire
Anna's hummingbird
Lewis' woodpecker
Blue grosbeak
Black-chinned
hummingbird
Acorn woodpecker

*

•

112

Rangelands

Table 6. Wildlife guilds of primary consumers in chaparral habitats of west-central Arizona (from Short
1983) (continued).

Guild
No.

Guild
members

Feeding loci

Breeding loci

CD
O

as

CO
-O

13
CO

to
CD

q3

o
to

CO

to
CD

CD
h-

CD
03
>.
O

00

"O

E

o
n

co

CD
O
JD

CD
CD

h1

o.
o

c

<T3
O

CD
CD

h1

<

CD

q3
.c
5

CD

oo

CD
c/)
"O
CD
CD

m

<D
O

m

=j

CO

n

00

to

CD

CD

(—

CD
O
CD

=3
00

CO
CD

o3

1—

CD
>.

a

>,

o

to
"D

E
o

-Q

=3

.C

CO

CD

o
n

CD
CD

h1

Q.

o

c

o

CD
CD

H

10
(cont.)

Broad-tailed

hummingbird
Cassin's kingbird
Mountain bluebird
Rufous hummingbird
Starling
Pinyon jay

Calliope hummingbird
Green-tailed towhee
Black-headed

grosbeak
American goldfinch
Common flicker
Steller's jay
Ladder-backed

woodpecker
Lazuli bunting
Yellow-rumped warbler
Western kingbird
Swainson's thrush
Elk

Hermit thrush
White-winged dove
Western bluebird

11

Desert spiny lizard
Canyon mouse
Western spotted

skunk
Cliff chipmunk
Arizona woodrat

12

Southern grass-
hopper mouse

Ord's kangaroo rat

Northern grass-
hopper mouse

Desert tortoise

13

Mourning dove

•

14

Rufous-sided
towhee

Eastern cottontail

Black-tailed jack-
rabbit

Common raven

Rock dove

Gambel's quail

House mouse

Rangelands

113

Table 6. Wildlife guilds of primary consumers in chaparral habitats of west-central Arizona (from Short
1983) (concluded).

Guild
No.

Guild
members

Feeding loci

Breeding loci

CD
O

ro

CO

n

CD

q3

(—

CD
O

CO

to

CD

CD

I—

CD
>.

ro
>~
o

CO

"O
E

o
n

sz

CD
O

n

CD
CD

H

D.
O
C
CO
o

CD
CD

H

<

CD
CD

sz

CD
CO
CD
CO

-a
o

CD

CD
O
03

(/>

n

zj

CO

03

CO

CD

CD

1—

CD
U

■2

CO

to

CD

q3

l—

CD

m
>.
o

CO

E

o
n

x:

CD
O
_Q

CD
CD

H1

>.

Q.
O

c

CO

o

CD
CD

h1

14
(cont.)

Desert cottontail
Rofous-crowned
sparrow

•
•

•
•

15

Brown towhee
Black-chinned

sparrow
Black-throated

sparrow

•
•

•

•
•

•

16

Western meadowlark
Lark sparrow
Chipping sparrow
Fox sparrow
Purple finch
Brewer's sparrow
Dark-eyed junco
White-crowned sparrow
Water pipit

•
•
•
•
•
•
•
•
•

17

Say's phoebe

•

•

18

Verdin

•

19

Hooded oriole
Warbling vireo
Phainopepla
Ruby-crowned kinglet
Plain titmouse
Ash-throated flycatcher
Wied's crested

flycatcher
Northern oriole
Yellow-bellied

sapsucker
Hermit warbler

114

Rangelands

species that require the tree bole layer habitat. The
map in Figure 6 would identify the distribution of
habitat favorable to this group of species. The land
units containing particular layers of habitat can be
surveyed to determine the quantity of habitat that
exists for specialized wildlife species or groups of
wildlife species. For example, 66.6 ha ( 166.4 a.) of
habitat containing tree boles suitable for cavity nest-
ing species occurred in Segment A, and 42.8 ha
(107.1 a.) of this habitat occurred in Segment B of
Figure 6 (Table 5). The chaining of pinyon- juniper
on Segments A or B would reduce the quantity of
the tree bole layer and adversely affect wildlife
guilds having a dependency on the tree bole layer.

Use of Habitat Layers by Individual Species

Maps of habitat layers, coupled with lists of the
wildlife guilds that use those habitat layers, provide
one way to evaluate a habitat cover type. Approxi-
mating where a niche space of a species occurs in a
layer of habitat is a second appraisal technique, al-
though a more arduous one. This technique, how-
ever, yields an estimate of the utility of a range
habitat for a particular wildlife species.

A layer of habitat can vary greatly in its compo-
sition and utility to wildlife. For example, the terres-
trial surface layer can provide a variety of structural
features for wildlife (listed under code 6 in Table 7).
The understory can ( 1 ) be devoid of cover on salt
playas and some beaches, (2) have the appearance of
desert pavement, (3) contain rocks or be a rock
surface in talus fields and on cliffs, (4) be covered
with tree litter under a forest overstory, and ( 5 ) be
covered with grasses, forbs, and dwarf shrubs of
different densities and heights. The large array of
structured conditions along the terrestrial surface
layer can be considered as constituting a gradient of
conditions for wildlife species having a dependency
on this layer of habitat. Range surveys that describe

surface cover, plant cover type, and plant vigor pro-
vide information that helps position a study area
along such a terrestrial surface gradient.

The terrestrial surface layer in the native grass-
lands of North America provides a variety of struc-
tures corresponding to those listed in subunits 6.14
to 6.19 of Table 7. The grasslands might vary from
bunchgrasses, providing low, scattered herbaceous
cover, to sod grasses, providing a tall, dense herba-
ceous cover. Wildlife species may generally select a
particular habitat layer (e.g., terrestrial surface layer)
and a particular structure of that habitat layer (e.g.,
shortgrass prairie with sparse cover) and then fine
tune their habitat selection on criteria like percent-
age bare ground or density and height of particular
plant species.

Thus, the greatest abundance of wildlife species
A may occur at point X along the terrestrial surface
habitat gradient, whereas the greatest abundance
of wildlife species B may occur at point Y along the
gradient.

A hypothetical habitat layer gradient for grass-
lands is presented in Figure 8. Horned larks (Eremo-
phila alpestris) occupy a position along the habitat
gradient characterized as open shortgrass prairie;
western meadowlarks occupy a position character-
ized by medium-dense mid to tall grasses; and bobo-
links (Dolichonyx oryzivorus ) occupy a position
characterized by dense, tall stands of grasses, weeds,
or alfalfa. These three species might belong to the
same grassland wildlife guild, even though their indi-
vidual presence in a survey would be dependent on
the nature of the herbaceous cover in the grassland
being inventoried.

Individual wildlife species can be approximately
positioned within a gradient developed for a layer
of habitat, if simple word models or HSI models (U.S.
Fish and Wildlife Service 1980) are constructed.

Low, scattered, herbaceous cover •*■

-► High, dense, herbaceous cover

Figure 8. Hypothetical habitat gradient that illustrates how a variety of birds select grasslands with slightly
different structures.

Rangelands

115

Table 7. Habitat conditions that help describe gradients for the terrestrial subsurface, terrestrial surface,

midstory, tree bole, and tree canopy layers of habitat. The codes 5—9 cross-reference to the identification
of habitat layers listed along the x-axes of the species-habitat matrices in Figure 7.

5. Terrestrial subsurface (to 10 cm [4 in.] below apparent surface)

5 01 Rocky substrate unsuitable for burrow or tunnel construction

5.02 Hydric or heavy soils unsuitable for burrow or tunnel construction

5.03 Powdery soils unsuitable for burrow or tunnel construction

5.04 Soils suitable for burrow or tunnel construction

5.05 Cave and deep crevice

5.06 Artificial (man-made) structures, such as mine shafts and out buildings, where interior use is similar to that of burrows or
caves

6. Terrestrial surface — Understory layer (from 10 cm [4 in] below to 0.15 m [6 in] above apparent surface)

6.01 Salt playas or flats with hydric soils

6.02 Beaches (mud, sand, or rock) without hydrophytes

6.03 Marshy areas with hydrophytes but not hydric soils

6.04 Bare ground (sand to rubble, up to 305 mm [12 in] particles)

6.05 Boulder (> 305 mm [12 in.]) — covered surface

6.06 Talus — unvegetated

6.07 Talus — vegetated

6.08 Cliff — on ledge near valley floor

6.09 Cliff — in cavity near valley floor

6.10 Cliff — on ledge near mesa or mountaintop
6 1 1 Cliff — in cavity near mesa or mountaintop
6 12 Herbaceous litter

6 13 Woody litter (includes shrub branches, tree branches, and stumps)

6.14 Shortgrass prairie vegetation with sparse (< 50%) cover

6.15 Shortgrass prairie vegetation with open or closed (> 50%) cover
6 16 Midgrass prairie vegetation with sparse (< 50%) cover

6 17 Midgrass prairie vegatation with open or closed (> 50%) cover
6 18 Tallgrass prairie vegetation with sparse (< 50%) cover

6.19 Tallgrass prairie vegetation with open or closed (> 50%) cover

6.20 Supine or dwarf woody vegetation or woody vegetation within 0.5 m (19 in.) of apparent surface

6.21 Cactus stems and pads

6.22 Artificial (man-made) structures — ground debris and artifacts, bridges, trestles, and rooftops where external use is analo-
gous to that of the horizontal and vertical surface of natural objects

7. Midstory layer (from 0.5 to 8.0 m [ 19 to 26 ft] above apparent surface)

7 01 Canopy of deciduous shrubs or small trees (< 50% cover)

7.02 Canopy of deciduous shrubs or small trees (> 50% cover)

7.03 Canopy of evergreen shrubs or small trees (< 50% cover)

7.04 Canopy of evergreen shrubs or small trees (> 50% cover)

7.05 Grass and grasslike vegetation (includes bamboo) extending into midstory layer

7.06 Forb vegetation extending into midstory layer

7.07 Cactus stems and pads extending into midstory layer

7.08 Artificial (man-made) structures extending into midstory layer and used in a manner analogous to that of natural vegeta-
tion

8. Tree bole layer

8.01 Snag — trunk of dead deciduous tree

8.02 Snag — trunk of dead evergreen tree

8.03 Trunk — living deciduous tree

8.04 Trunk — living evergreen tree
8 05 Cactus bole or stem

8.06 Artificial (man-made) structure — telephone and power poles, chimneys, nest boxes

9. Tree canopy

9.01 Small branches— live deciduous trees (< 50% cover)

9.02 Small branches — live deciduous trees (> 50% cover)

9.03 Small branches — live evergreen trees (< 50% cover)

9.04 Small branches — live evergreen trees (> 50% cover)

9.05 Large branches — live deciduous trees (< 50% cover)

9.06 Large branches — live deciduous trees (> 50% cover)

9.07 Large branches — live evergreen trees (< 50% cover)

9.08 Large branches — live evergreen trees (> 50% cover)

9.09 Large branches — dead deciduous trees

9.10 Large branches — dead evergreen trees

116 Rangelands

Screech owl (Otus kennicottii)

1. Cover types inhabited

Northern floodplain forest
Pinyon-juniper woodland
Eastern ponderosa forest
Black Hills pine forest
Pine-Douglas fir forest
Grama- buffalograss
Wheatgrass-needlegrass
Bluestem-grama prairie
Sandsage-bluestem prairie
Grama-needlegrass-wheatgrass

2. Habitat layers where nesting, birthing, or hatching occur

Tree bole layer

Substrate 8.01, 8.02, 8.03, 8.04, 8.06 (from Table 8). Nest is nearly always in a woodpecker/flicker hole in a Cottonwood or
other large deciduous tree. Occasionally found in cavity in pinyons and junipers.

3. Habitat layers where foraging occurs

Secondary consumers

Terrestrial surface layer: feeds on rodents, small birds, amphibians, and reptiles

4. Species-habitat model (references)

Requires dense cottonwoods or other deciduous
woodlots along watercourses from lower moun-
tains throughout plains. May select other trees,
including isolated cottonwoods, in a variety of
range habitats.

Cottonwoods or other trees are at least 20 cm
(8 in ) dbh.

■No-

Yes

Trees are of negligible value as nesting habitat
for the screech owl.

Trees possess woodpecker/flicker holes suitable
for use as nest sites.

■No-

Yes

Trees may represent useful nesting habitat for
the screech owl.

Figure 9. Information to approximate conditions along a layer of habitat useful to the screech owl (Otus
kennicottii) in Wyoming and Colorado (after U.S. Forest Service 1981 ).

Rangelands

117

Least chipmunk (Tamias minimus)

1. Cover types inhabited

Mountain mahogany-oak scrub
Sagebrush steppe
Eastern ponderosa pine
Black Hill pine forest
Pine-Douglas fir forest
Pinyon-juniper woodland
Great Basin sagebrush
Douglas fir forest
Saltbush-greasewood

2. Habitat layers where nesting, birthing, or hatching occur

Terrestrial subsurface layer — Substrate 5.04

Terrestrial surface layer — Substrate-6 06, 6.07, 6.13, 6.22 (from Table 8) Nests are found under stumps, logs, or rocks.

3. Habitat layers where foraging occurs

Primary consumer

Terrestrial surface layer: consumes fruit, nuts, seeds, berries, mushrooms, leaves, stems.

Secondary consumer

Terrestrial surface layer: consumes invertebrates, especially insects

4. Species-habitat model

Uses surface layer within a variety of cover
types, including semiarid sagebrush and grease-
wood cover types, browse slopes, brushy
ravines, and coniferous mountain forests.

Yes

_L

Terrestrial surface has sparse (20 to 50%) cover.

■No

Yes

Terrestrial surface provides stumps, logs, rocks,
and roots under which chipmunks can burrow.

Yes

Understory, midstory, and/or overstory plant
species provide a source of fruit, nuts, berries,
seeds, and leaves and the terrestrial surface
layer seems capable of providing a source of
invertebrates as food sources.

Habitat is of negligible value to the least chip-
munk.

■No

Yes

Cover type may represent useful habitat to the
least chipmunk.

Figure 10. Information to approximate conditions along a layer of habitat useful to the least chipmunk
(Tamias minimus) in Wyoming and Colorado (after U.S. Forest Service 1981).

118

Rangelands

Brewer's sparrow (SpizeUa brewer!)

1. Cover types inhabited

Grama-buffalo grass
Sandsage-bluestem prairie
Mountain mahogany-oak scrub
Great Basin sagebrush
Wheatgrass-needlegrass shrub steppe
Wheatgrass-needlegrass
Grama-needlegrass-wheatgrass
Saltbush-greasewood
Sagebrush steppe

2. Habitat layers where nesting, birthing, or hatching occur

Terrestrial surface layers

Substrate 6.20, 7.01 , 7.02, 7.03, 7.04 (from Table 8). Brewer's sparrows frequently construct their nest about 15 to 25 cm (6
to 10 in.) above ground in dense woody shrubs, especially sagebrush.

3. Habitat layers where foraging occurs

Primary consumer: terrestrial surface layer. The Brewer's sparrow feeds on grass and seeds on the ground.

Secondary consumer: terrestrial surface layer. The Brewer's sparrow gleans insects from the ground and from low shrubs.
Midstory layer: the Brewer's sparrow gleans insects from large shrubs.

4. Species-habitat model (references)

Living woody shrub vegetation within the breed-
ing range of the Brewer's sparrow is present in a
habitat block greater than or equal to 0.2 ha
(0.5 a.) in area

No

Yes
t

Living woody shrub vegetation occurs on terrain
whose slope does not exceed 30°.

Habitat block is of negligible value as breeding
habitat for the Brewer's sparrow.

■No

Yes

Living woody shrub vegetation occurs on gener-
ally well-drained soils rather than on rock out-
crops.

■No

Yes

Habitat may be of high quality if woody shrub
species are 50-70 cm (20-28 in.) in height, pro-
vide a dense limb structure 25 cm (10 in.) above
ground, and provide a shrub canopy 30%. Habi-
tat may be of low to moderate quality if these
conditions do not occur.

Yes

Habitat may be of high quality as nesting habitat
to the Brewer's sparrow if woody shrubs are
sagebrush; habitat may be of medium quality if
woody shrubs are hawthorn, plum, serviceberry,
bitterbrush, or rabbitbrush, habitat may be of low
quality if shrubs are saltbush, greasewood, hop-
sage, ceanothus, manzanita, currant, or other.

Figure 11. Information to approximate conditions along a layer of habitat useful to the Brewer's sparrow in
(Spizella breweri) Wyoming and Colorado (after U.S. Forest Service 1981).

Rangelands

119

The intent of either model is to describe habitat con-
ditions favorable to the species. Word models are
listed in Figures 9 through 1 1 for the screech owl
{Otis kennicottii), least chipmunk (Tamias mini-
mus), and Brewer's sparrow, respectively. The
models describe the cover types inhabited by the
species. This information can be compiled from
many sources, including field guides, range maps,
museum records, and other published literature. The
best situation is where such a list already exists. For
example, U.S. Forest Service personnel already have
associated wildlife species with range cover types in
Colorado, Wyoming, Kansas, Nebraska, and South
Dakota (U.S. Forest Service 1981); for Arizona and
New Mexico on the computerized RUNWILD format
(Patton 1978); and for the western Sierra Nevadas
(Verner and Boss 1980).

The models also identify the layers of habitat
where nesting, hatching, or birthing of the species

occur and where foraging takes place and special
structural features or plant species apparently se-
lected by the species within a habitat layer. The
numerical designation of the major structural fea-
tures within layers of habitat corresponds to the
coded subunits in Table 8. Sound, professional judg-
ment is needed when developing word models for
placing wildlife species along habitat gradients be-
cause species-habitat information usually is not well-
synthesized in the literature.

Models like those in Figures 9 to 1 1 try to de-
fine the niche space of a species in sufficient detail
so that the usefulness of habitat for particular species
can be qualitatively estimated in inventories. Field
surveys can be used to estimate the relative quantity
of habitat for particular species. It may also be possi-
ble to determine, from the word model, those life-
requisite needs of the species that need to be pro-
vided or modified by remedial management actions.

A typical sagebrush rangeland in the western United States.

120

Rangelands

LITERATURE CITED

ALBERTSON, F.W. and J.E. WEAVER. 1946. Reduction of
ungrazed mixed prairie to short grass as a result of
drought and dust. Ecol. Monogr. l6(4):449-463-
BEETLE, A.A. I960. A study of sagebrush. The section
tridentatae of Artemisia. Wyoming Agric. Exp. Sta.
Bull. 368. Laramie. 83pp.
BLAISDELL, J.P, R.B. MURRAY, and ED. MCARTHUR.
1982. Managing intermountain rangelands — sage-
brush-grass ranges. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. INT- 134. 4 lpp.
BROWN, D.E. 1982. Biotic communities of the American
Southwest — United States and Mexico. Boyce Thomp-
son Southwestern Arboretum. Superior, AZ. 342pp.

and C.H. LOWE. 1980. Biotic communities of the

Southwest. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-78. lpp. (map).
COSTELLO, D.F. 1944. Important species of the major

forage types in Colorado and Wyoming. Ecol. Monogr.
14(1):107-134.
HALL, F.C. 1973- Plant communities of the Blue Mountains
in eastern Oregon and southwestern Washington.
U.S. Dep. Agric, For. Serv. PNW-R6 Area Guide 3-1.
62pp.
HARPER, K.T. and J.L. REVEAL (symp. organizers). 1978.
Intermountain biography: a symposium. Great Basin
Naturalist Memoirs. 2. Brigham Young Univ. Provo,
UT. 268pp.
HIRONAKA, M. 1979. Basic synecological relationships of
the Columbia River sagebrush type. Pages 27-32 in
The Sagebrush Ecosystem — A Symposium. Utah State
Univ., College of Natural Resources. Logan. 251pp.
, MA. FOSBERG, and A.H. WINWARD. 1983. Sage-
brush-grass habitat types of southern Idaho. Forest,
Wildlife, and Range Exp. Sta. Bull. 35. Univ. of Idaho,
Moscow. 44pp.
JOHNSON, K.L. 1979- Basic synecological relationships of
the sagebrush types on the high plains of Montana,
Wyoming, and the Dakotas. Pages 42-49 in the Sage-
brush Ecosystem — A Symposium. Utah State Univ.,
College of Natural Resources. Logan. 251pp.
KUCHLER, AW. 1964. Manual to accompany the map.
Potential natural vegetation of the conterminous
United States. Am. Geogr. Soc Spec. Publ. 36. 39pp.
+ 116 plates, (map).
MCARTHUR, ED., AC. BLAUER, A.P. PLUMMER, and R.
STEVENS. 1979. Characteristics and hybridization of
important intermountain shrubs. III. Sunflower family.
U.S. Dep. Agric, For. Serv. Res. Pap. INT-220. 82pp.
MCKELL, CM., J.P. BLAISDELL, and JR. GOODIN, tech.
eds. 1972. Wildland shrubs — their biology and utiliza-
tion. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. INT-
1. 494pp.
MOONEY, H.A. and D.J. PARSONS. 1973. Structure and

function of the California chaparral — an example from
San Dimas. Pages 83-1 12 in di Castra, F. and H.A.
Mooney, eds. Mediterranean Type Ecosystems: Origin
and Structure. Springer-Verlag. New York, NY.
MUEGGLER, W.F. and W.L. STEWART. 1980. Grassland
and shrubland habitat types of western Montana. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. INT-66. 154pp.
NICHOL, A.A. 1937. The natural vegetation of Arizona.
Ariz. Agric. Exp. Sta. College of Agric, Univ. Ariz.
Tucson. Tech. Bull. 68:181-222.

PATTON, DR. 1978. RUNWILD: A storage and retrieval
system for wildlife habitat information. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. RM 51. 8pp.

PLUMMER, A.P, DR. CHRISTENSEN, and SB. MONSEN.
1968. Restoring big-game range in Utah. Publ. No. 68-
3. Utah Div. Fish and Game. Salt Lake City. 183pp.

, AC. HULL, Jr., G. STEWART, and J.H. ROBERTSON.

1955. Seeding rangelands in Utah, Nevada, southern
Idaho, and western Wyoming. U.S. Dep. Agric, For.
Serv. Agric. Handb. 71. 73pp

POULTON, C.E. 1970. Practical applications of remote
sensing in range resources development and manage-
ment. Pages 179-189 in Range and Wildlife Habitat
Evaluation — A Research Symposium. U.S. Dep. Agric,
For. Serv. Misc. Publ. 1147. Washington, DC.

SAMPSON, AW, A. CHASE, and D.W. HEDRICK. 1951.
California grasslands and range forage grasses. Calif.
Agric Exp. Sta, College of Agric. Univ. Calif. Bull.
724. 130pp.

SCHWARZ, C.F., C.E. THOR, and G.H. ELSNER. 1976.

Wildlife planning glossary. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. PSW-13. 252pp.

SHARP, LA. and K.D. SANDERS. 1978. Rangeland re-
sources of Idaho: A basis for development and im-
provement. Idaho Rangeland Comm. and College of
FWR. FWR Exp. Sta. Misc. Publ. 6. Moscow, ID. 74pp.

SHORT, H.L. 1983. Wildlife guilds in Arizona desert habi-
tats. U.S. Dep. Inter, Bur. Land Manage. Tech. Note
362. 258pp.

and K.P. BURNHAM. 1982. Technique for structur-
ing wildlife guilds to evaluate impacts on wildlife
communities. U.S. Dep. Inter, Fish and Wildl. Serv.
Spec. Sci. Rep. - Wildl. 244. 34pp.

STODDARD, LA, AD. SMITH, and T.W. BOX. 1975.

Range management. Third ed. McGraw-Hill Book Co.
New York, NY. 532pp.

TISDALE, E.W. 1979. A preliminary classification of Snake
River Canyon grasslands in Idaho. Univ. Idaho FWR
Exp. Sta. Note. 32. Moscow.

. 1983. Grasslands of western North America: The

Pacific Northwest bunchgrass type. Pages 223-245 in
Nicholson, A.C, A. McLean, and T.E. Baker, eds. Grass-
land Ecology and Classification Symposium Proceed-
ings. B.C. Min. For. Victoria, B.C.
DEPARTMENT OF AGRICULTURE. 1936. Atlas of
American agriculture. U.S. Govt. Printing Office.
Washington, DC.

DEPARTMENT OF INTERIOR, BUREAU OF LAND
MANAGEMENT. 1975. Range condition report pre-
pared for the Senate Committee on appropriations.
U.S. Dep. Inter.. Bur. Land Manage. Washington, DC.
(unpubl).

FISH AND WILDLIFE SERVICE. 1980. Habitat Evalu-
ation Procedures (HEP), ESM 102. U.S. Dep. Inter,
Fish and Wildl. Serv. Div. of Ecological Services.
Washington, DC.

FOREST SERVICE. 1972. The Nation's range re-
sources— a forest-range environmental study. For.
Resour. Rep. 19. U.S. Dep. Agric, For. Serv. Washing-
ton, DC. 147pp.

. 1981. Wildlife and fish habitat relationships. Vol. I

Narratives. Vol. II Matrices. U.S. Dep. Agric, For. Serv.
Rocky Mtn. Region. Denver, CO.
UTAH STATE UNrVERSITY. 1975. The pinyon- juniper

ecosystem: a symposium. Utah State Univ., College of
Natural Resources, Utah Agric. Exp. Sta, Logan.
194pp.

U.S.

U.S.

U.S

U.S.

Rangelands

121

.1979. The sagebrush ecosystem: a symposium. Utah

State Univ., College of Natural Resources, Logan.

251pp.
VERNER, J. and A.S. BOSS. 1980. California wildlife and

their habitats: Western Sierra Nevada. U.S. Dep. Agric,

For. Serv. Gen. Tech. Rep. PSW-37. 439pp.
WAGNER, W.L. and E.F. ALDON. 1978. Manual of the

saltbushes (Atriplex sp.) in New Mexico. U.S. Dep.

Agric, For. Serv. Gen. Tech. Rep. RM-57. 50pp.
WEAVER, J.E. 1954. North American prairie. Johnson Publ.

Co. Lincoln, NE. 348pp.
and F.W. ALBERTSON. 1956. Grasslands of the

Great Plains. Johnson Publ. Co. Lincoln, NE. 395pp.

WINWARD, A.H. 1980. Taxonomy and ecology of sage-
brush in Oregon. Agric. Exp. Sta. Bull. 642. Oregon
State Univ., Corvallis. 15pp.

WRIGHT, HA. 1972. Shrub response to fire. Pages 204-
217 in McKell, C.M.J.P Blaisdell, and JR. Goodin,
tech eds. Wildland Shrubs — Their Biology and Utiliza-
tion. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. INT-
1. 494pp.

, L.F. NEUENSCHWANDER, and CM. BRITTON.

1979. The role and use of fire in sagebrush-grass and
pinyon-juniper plant communities. A state-of-the-art
review. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
INT- 58. 48pp.

122

Rangelands

DESERTS

K. Bruce Jones1

U.S. Bureau of Land Management
Phoenix Training Center
Phoenix, AZ 85015

"But the desert is much more than merely warm.
It is a consistent world with a special landscape, a
special geography, and to go with them, a special
flora and fauna adapted to that geography and that
climate."

—Joseph Wood Krutch, from The Voice of the
Desert

Editors Note: Deserts compose a large portion of
the western U.S. Wildlife are one of their most im-
portant resources. Animals of the desert have
unique adaptations to the extremes of heat and
precipitation that characterize such regions. As the
author emphasizes, effective inventory and moni-
toring studies must take into account these ex-
tremes and desert animals' adaptations to them.

INTRODUCTION

Deserts are characterized by low, erratic precip-
itation and highly variable temperatures, and extend
from southeastern Oregon and southern Idaho
through Nevada and Utah, except at higher eleva-
tions, continuing south through southern California
and Arizona, and eastward through central and
southern New Mexico (Oosting 1956). Desert habi-
tats range from homogeneous stands of sagebrush
{Artemisia sp. ) in the Great Basin Desert to highly
diverse, structurally rich vegetation of the Sonoran
Desert. To a large degree, vegetation structure
within deserts reflects the areas' precipitation pattern
and temperature regimes. For example, structural
simplicity within the Great Basin Desert reflects the
area's short growing season, low precipitation, and
precipitation pattern (60% of the precipitation is in
the form of snow ). Conversely, parts of the Sonoran
Desert are structurally rich because of a year-round
growing season (few freezing temperatures), bian-
nual precipitation patterns, and precipitation in the
form of rain.

Northern, cooler desert regions, such as the
Great Basin Desert, support far fewer wildlife species
than more southern, warmer regions such as the So-
noran Desert (Bender 1982; Brown 1982b). Certain
groups of wildlife, such as large herbivores, are gen-
erally more numerous in northern deserts. Similar to
their effect on vegetation, shorter growing seasons
affect the diversity of wildlife, especially small birds,
mammals, reptiles, and amphibians by reducing the
diversity and availability of insect prey. Thermal re-
gimes in northern deserts limit ectothermic (cold-
blooded) and small homeothermic (warm-blooded)
wildlife activity to short periods (generally May to
early September).

Desert habitats possess some of the most unu-
sual wildlife in North America. Many desert animals
are physiologically and morphologically adapted to
survive under extreme environmental conditions
(low, erratic rainfall and highly variable tempera-
tures). Many small mammals require no free water;
these animals survive on their own metabolic water
and through water conservation strategies such as

•Current Address: U.S. Fish and Wildlife Service. Office of Endan-
gered Species, Washington, DC 20240.

Deserts

123

nocturnal activity patterns (Golley et al. 1975).
Many desert animals such as lizards conserve water
by excreting uric acid rather than urea (Porter
1972).

Perhaps the single most important adaptation
employed by desert wildlife is behavior. Many small
mammals and snakes are entirely nocturnal and fos-
sorial. They avoid high temperatures by remaining
inactive in burrows, under rock, or in caves (e.g.,
bats) during the day, and perform feeding and repro-
ductive activities at night. Even diurnal animals such
as some birds, big game, and lizards are inactive dur-
ing hot parts of the day.

In desert regions, most animals hibernate or re-
main inactive during cold winter months. This is es-
pecially true in northern sections of the Great Basin
Desert where winter temperatures drop below 0° F.
In the Sonoran Desert, where temperatures rarely
reach the freezing point, many small mammals and
small lizards remain active year-round.

as vegetation patches, rock, soil, and surface debris
(e.g., logs). The arrangement and abundance of these
microhabitats determine the diversity and abundance
of desert wildlife; small changes in these microhabi-
tats can alter species abundance and diversity (see
Heatwole 1982 and Short 1983 for examples). Be-
cause animals are closely associated with microhabi-
tats, man can easily affect the composition and abun-
dance of wildlife in deserts.

With rapid expansion of man into desert habi-
tats, many habitat components crucial to species ex-
istence are being altered, especially those important
to both man and wildlife. For example, dewatering of
perennial streams and springs for domestic and live-
stock water has drastic effects on wildlife, especially
aquatic organisms. Recreation activities have become
very popular in deserts near metropolitan areas, and
these activities have both direct and indirect effects
on wildlife and their habitats (physical disturbance
and habitat alteration, respectively). In addition,
many regions, especially within the Great Basin Des-
ert, have large coal, oil, and gas deposits. The devel-
opment of these resources can have significant ef-
fects on desert habitats and their faunas.

<-.*■:,;. '.-•$•' ->■< «*-?m. *j?r

Great Basin Desert in the winter.

Response (return to original state) of desert
habitats after man-caused physical changes is slow,
and in severe instances, where soils are lost, habitats
may never return to their original state.

Inventories and monitoring studies of wildlife
communities in desert habitats help biologists and
managers understand the sensitive and complex in-
terrelationships between wildlife species and com-
munities, and resources of desert habitats. With
these data, one can protect, and in some situations,
enhance the integrity of our deserts. This chapter de-
scribes desert habitats, including habitat compo-
nents, and provides general guidance for conducting
inventories and monitoring studies in deserts.

Although many desert wildlife require no free-
standing water (see Golley et al. 1975 for examples),
permanent and temporary water sources are critical
to certain wildlife. For example, big game species
such as bighorn sheep (Ovis canadensis) occupy
areas with scattered, but reliable free-standing water
(Wilson et al. 1980). Many habitat management strat-
egies are developed around this concept; bighorn
sheep populations can be increased by increasing the
distribution and reliability of free-standing water.
Small streams transecting through desert habitats are
inhabited by several small fish and amphibians. These
animals are entirely dependent on permanent, run-
ning (lotic) water within these regions.

Because of extreme environmental conditions,
desert wildlife are highly dependent on microhabi-
tats, especially those that provide thermal cover such

Desert bighorn sheep at man-made water supply.

124

Deserts

CLASSIFICATION AND DESCRIPTIONS OF
MAJOR DESERTS

Although deserts cover large sections of western
North America, they possess certain similar environ-
mental characteristics throughout (Oosting 1956). All
deserts share the following general characteristics:

1. low, erratic precipitation, generally less than
300 mm (12 in. ) per year

2. high, daily air and soil temperatures that drop
rapidly at night

3. highly variable, seasonal temperatures

4. low atmospheric moisture throughout most of
the year

5. few overcast or cloudy days

Shreve ( 1942) and Oosting ( 1956) characterized des-
erts into two general groups based on climate: cold
formation and warm formation. Lowe and Brown
( 1982) further subdivided warm formation deserts
into warm-temperate and tropical-subtropical based
on climate and plant species affinities and origin:

1. Cold-temperate Deserts
a. Great Basin Desert

2. Warm- temperate Deserts

a. Mojave Desert

b. Chihuahuan Desert

3. Tropical-subtropical Deserts
a. Sonoran Desert

Because most authors agree on the four major
deserts listed above, I will base my discussions
around this classification. Finer delineation of these
four major deserts is available. Perhaps the most ex-

tensive delineation of North America's four major
deserts is that of Brown ( 1982b). In his publication,
deserts are broken down to association and commu-
nity (series) levels, following the digitized format of
Brown et al. (1979). Other delineations of deserts
include Province and Region descriptions by Bailey
(1978), and potential natural vegetation descriptions
to the association level by Kuchler ( 1 964 ).

The following are brief descriptions of each des-
ert (see Figure 1 for locations).

Great Basin Desert

The Great Basin Desert is the most extensive
desert in the U.S., stretching from southeastern Ore-
gon and Wyoming, south to northern Arizona and
New Mexico, east to northern New Mexico, and
west into extreme eastern California. Topography of
the Great Basin Desert varies, but generally consists
of wide valley floors between 1,200 and 1,800 m
(4,000 and 6,000 ft), interrupted by mountains.
Temperatures drop much lower than any other U.S.
desert, with a short frost-free season and very cold
winters (Oosting 1956), and precipitation ranges
from 76 to 300 mm (4 to 12 in.). Two major vegeta-
tion communities occur within this desert, both of
which are structurally and floristically simple: sage-
brush associations dominated by sagebrush, Artemi-
sia sp., and shadscale or saltbush associations domi-
nated by saltbushes, Atriplex sp. (Oosting 1956;
Turner 1982). Species with evolutionary affinities to
warmer climates such as rabbitbush {Chrysothamnas
sp.) and blackbrush (Coleogyne sp. ) are also present
in the Great Basin Desert (Turner 1982a).

Great Basin Desert in Utah, dominated by sand sagebrush.

Deserts

125

Great Basin

Mojave

Sonoran

Chihuahuan

Figure 1. Locations of major North American deserts.

126 Deserts

Mojave (Mohave) Desert

The Mojave Desert is the smallest of the desert
formations, occurring in southeastern California,
southern Nevada, northwestern Arizona, and extreme
southwestern Utah at elevations of 300-1,200 m
(1,000-4,000 ft; Oosting 1956). Summers are hot and
dry; precipitation throughout this desert is generally
meager (less than 120 mm [5 in.] per year) although
some areas receive up to 22.5 cm (9 in.) per year
(Thome et al. 1981). Floristically, faunally, and geo-
graphically, the Mojave Desert is an intermediate be-
tween Great Basin and Sonoran Deserts, although its
fauna more closely resembles that of the Sonoran
Desert (Turner 1982b). The Mojave Desert is per-
haps the most structurally and floristically variable of
all the deserts although most regions consist of low-
height (<1.5 m [5 ft]) shrubs consisting of homoge-
neous stands of creosotebush (Larrea sp.) and black-
brush; other regions include structurally-diverse flora
such as the Joshua tree ( Yucca brevifolia; Turner
1982a). There are few Mojave Desert endemics
(plant species) that distinguish this desert from oth-
ers (with the exception of the Joshua tree). Classifi-
cation of the Mojave Desert as a distinct desert is
primarily based on geographic position and climate
rather than plant or animal associates (Turner
1982b).

' - ■ : ' ' j*

Yucca in bloom.

A Mojave Desert Joshua tree "woodland" community in Arizona.

Deserts

127

Chihuahuan Desert, with mesquite (New Mexico).

Chihuahuan Desert

The Chihuahuan Desert is floristically and struc-
turally variable and has great regional differences in
plant species composition (Brown 1982a). Most of
this desert occurs in Mexico (80% ) with northern
boundaries extending into southeastern Arizona,
southern and central New Mexico, and western
Texas at elevations of 900-1,800 m (3,000-6,000 ft;
Brown 1982a). Similar to the Great Basin Desert, this
desert is interrupted by high mountains (Oosting
1956). Precipitation is variable and ranges from
76-300 mm (4-12 in.) per year. The Chihuahuan Des-
ert has expanded in the past 200 years at the ex-
pense of semidesert grassland, primarily due to man's
activities (York and Dick-Peddie 1969; Brown
1982a). Although present in the Mojave and Sonoran
Deserts, mesquite (Prosopis sp.) and creosotebush
dominate most sites within the U.S. Chihuahuan Des-
ert. Whereas most areas are structurally homogene-
ous and simple, whitethorn acacia (Acacia con-
stricta), ocotillo (Fouquieria splendens), yuccas
(Yucca sp. ), agaves (Agave sp.), and cactus (Opuntia
sp.) add to structural diversity in mountainous areas
and canyons (Brown 1982a). Solid mesquite dune
stands are common in southern New Mexico, pri-
marily on sites previously occupied by semidesert
grassland (Brown 1982a).

Sonoran Desert

The Sonoran Desert, occupying areas in south-
ern Arizona, southeastern California, and northwest-
ern Mexico, is perhaps the most structurally and flo-
ristically diverse desert in the world. It is also the
only U.S. desert with most of its flora derived from

tropical climatic zones, and appears to have formed
less than 10,000 years ago (Turner and Brown
1982). There are two major subdivisions of the So-
noran Desert within the U.S.: Lower Colorado River
Valley and Arizona Upland (Turner and Brown
1982). The Lower Colorado subdivision occurs pri-
marily in southwestern Arizona and southeastern Cal-
ifornia at elevations of 0-600 m (0-2,000 ft). Precipi-
tation is low (less than 175 mm, 7 in. per year) and
temperatures moderate (mostly frost-free). Creosote-
bush and bursage (Franseria sp. ) dominate most of
this subdivision. Structural and floristic diversity is
achieved in these areas along intermittent washes
consisting of ironwood (Olneya tesota), blue palo
verde (Cercidium floridum), and mesquite (Turner
and Brown 1982).

The Arizona Upland subdivision of the Sonoran
Desert is the most structurally diverse of all U.S. des-
erts. It occurs in southern Arizona and northern So-
nora at elevations of 300-1,200 m (1,000-4,000 ft).
Similar to the Lower Colorado subdivision, this sub-
division receives rain biannually (summer and win-
ter), but unlike its other Sonoran counterpart, has
greater moisture (due to higher elevations and rug-
ged, broken terrain; Turner and Brown 1982).
Turner and Brown ( 1982) attributed dominance of
arboreal and succulent vegetation within this region
to available moisture. Overstory structure is pro-
vided by ironwoods, foothill palo verde (Cercidium
microphyllum), saguaro cactus (Carnegiea gigan-
tea ), and mesquite. A wide variety of understory
components may also be present: cactus (Opuntia
sp. ), bursage, ratany (Krameria sp. ), buckwheat (Er-
iogonum sp. ), brittlebush (Encelia sp. ), and jojoba
(Simmondsia chinensis; Turner and Brown 1982).

128

Deserts

Sonoran Desert in Arizona.

MAJOR WILDLIFE SPECIES GROUPS

Wildlife species assemblages of North American
deserts range from relatively undiverse in homogene-
ous, northern sections of the Great Basin Desert to
highly diverse (Table 1 ) in the Arizona Upland sub-
division of the Sonoran Desert (see Bender 1982 and
Brown 1982b). Great differences in faunas of north-
ern portions of the Great Basin Desert and southern
deserts such as the Sonoran Desert primarily result
from differences in temperature, moisture, and struc-
tural characteristics associated with each. As dis-
cussed in the previous section, the Great Basin Des-
ert is extremely homogeneous, structurally simple,
cold, and precipitation falls primarily in the winter

(60% ) when plants are dormant. Conversely, the So-
noran Desert is structurally diverse, warm, and re-
ceives precipitation during both winter and summer.
Moderate temperatures allow plants to use winter
and summer precipitation for growth (Oosting 1956;
Daubenmire 1974). Lack of available moisture and
short growing seasons may preclude existence of
many arboreal plants from the Great Basin Desert,
common in southern deserts.

The Great Basin Desert has fewer wildlife species
than any of the other deserts. However, this desert
supports larger populations of pronghorn (Antiloca-
pra americana), mule deer {Odocoilens hemionus),
and elk {Cetvus elapbus ) than other deserts (see

Table 1. Qualitative assessment of species diversity for the four major U.S. deserts.
£ = highly diverse, Q = moderately diverse, Q = low diversity.

Desert

Total
Diversity

Raptor

Fish

Amphibians

Reptiles

Small
Birds

Mammals
(Non-game)

Upland
Game

Water-
fowl

Big
Game

Great Basin

o

Q

O

o

O

O

Q

Q

9

•

Mojave

Q

Q

O

Q

Q

9

Q

Q

O

o

Chihuahuan

Q

Q

O

•

•

9

Q

9

O

*

9

Sonoran

•

•

Q

•

•

•

•

9

o

9

'Moderate diversity results primarily from introduced species.

Deserts

129

Schmidt and Gilbert 1978). It also supports large up-
land game populations (Johnsgard 1973) such as chu-
kars (Alectoris chukar) and sage grouse (Centrocer-
cus urophasianus ). Several bird species move into
the Great Basin Desert during spring and summer to
breed, although breeding bird diversity and abun-
dance is generally less than any of the other three des-
erts (see Brown and Gibson 1983)- Breeding and res-
ident raptor and bird species are greatest within the
Great Basin Desert in areas adjacent to more structur-
ally diverse habitats, such as riparian areas and cliffs.
For example, riparian sites provide structural diversity
necessary for nesting, whereas the adjacent Great
Basin Desert provides feeding grounds. Areas with
cliffs increase diversity in this desert by increasing
structural habitat diversity.

Reptile, amphibian, and small mammal diversity
is relatively low in the Great Basin Desert when
compared with the three other deserts (see Brown
and Gibson 1983). As previously discussed, low di-
versity of small nongame wildlife results primarily
from a short growing season and low environmental
temperatures.

The Mojavc Desert supports an array of wildlife.
Whereas this desert's big and upland game diversity
and abundance is not as great as the Great Basin De-
sert, it does support a diverse small mammal and
reptile fauna. Similar to the Great Basin Desert, small
bird and raptor diversity and abundance are greatly
enhanced in areas with added structural diversity
(cliffs, riparian habitats, canyons, and Joshua trees).

The Chihuahuan and Sonoran Deserts have
highly diverse faunas, including large numbers of am-
phibians, reptiles, small birds, mammals, and raptors.
High species diversity results from topographical and
vegetational structural diversity, a long growing sea-
son, and influence of Mexican and southcentral U.S.

faunas. In southern New Mexico, Ibex (Capra aega-
grus ) and Oryx ( Oryx gazella ) have been intro-
duced into parts of the Chihuahuan Desert. Javelina
( Tayassu tajacn ) have also been introduced into
several parts of both the Chihuahuan and Sonoran
Deserts. These introductions, combined with native
big game and upland game species, give these des-
erts relatively diverse game faunas.

Siberian ibex in southern New Mexico habitat. Released in
1975 in the Canadian River Canyon.

Oryx, commonly called Gemsbok, in southern New Mex-
ico habitat. Released in 1969 on White Sands Missile
Range.

The Sonoran Desert has the highest wildlife di-
versity of any North American desert ( see Brown
and Gibson 1983), primarily resulting from a year-
round growing season and great vegetation structural
diversity.

Bender ( 1982 ) and Brown ( 1982b) provided
lists of wildlife and plants associated with North
American deserts. Table 1 provides a general assess-
ment of species richness in the four major U.S.
deserts.

CRITICAL HABITAT FEATURES

Deserts consist of several physical habitat com-
ponents that contribute to wildlife species diversity.
These components are used by wildlife to perform a
number of ecological functions such as mating, feed-
ing, nesting, and thermoregulation, and they also
provide resting substrate and escape cover. Habitat
components or resources are partitioned in time and
space by desert wildlife (Creusere and Whitford
1982). Because deserts arc environmentally extreme,
many wildlife species use specific habitat compo-
nents, especially those components that provide
greater moisture and lower temperatures. For exam-
ple, the desert night lizard (Xantusia vigilis) is re-
stricted to desert regions with downed litter of
yucca and agave plants (Zweifel and Lowe 1966;
Jones 1981a).

130

Deserts

Several other factors determine wildlife commu-
nity composition and affect availability of habitat
components. These include slope, aspect, precipita-
tion, ecotones, geographic, and man-caused factors.

Microhabitat components, such as soil, rock,
vegetation, and water, vary greatly between and

within our four U.S. deserts. The following discusses
each habitat component with examples of their rela-
tions to various wildlife. This will help determine
which microhabitats should be sampled when sur-
veying and monitoring certain desert wildlife. Tables
2 through 4 summarize important habitat compo-
nents in deserts.

Major Vertebrate Associations and Relationship
of Species' Ecology to Components

Table 2. Important habitat components of desert habitats.

CD
03

to
Q

CD

>

o
O

CD
Q.
CO

o
en
111

><

CO

a.
o

o
o

LL

03

to

3

(D
O)

c

T3
CD

CD
LL

cn

c
!c

t:

in .cd
i- 73

O i:

4_ </3

C/3 J3

CO 3

z c/>

CO

o
cn
o
o

03
>-

-C

Q-

c
o

o

3
"O

o

a.

CO

<r

03

to

3

CO

cn

c

u5

CO
LT

o
to

3

cn

CD CD
O CO

CO -Q

SI 3

h in

Microhabitat
Components

Description

Variables/Factors

Water
(Lotic)

Rivers and
streams,
flowing
springs

Riffle/run/pool water temperature,
turbidity, DO, organic content,
siltation, pollutants, interface with
other habitats, substrate

Fl

•

•

•

•

•

•

RE

•

•

•

•

•

•

BG

•

RA

Water

(Permanent

Lentic)

Ponds,

marshes,

lakes,

potholes,

bogs,

seeps,

springs,

agricultural

runoff,

natural and

man-made

catchments

Water temperature, DO, organic
content, siltation, pollutants,
interface with other habitats (e.g.,
nesting islands), emergent
vegetation, shoreline vegetation,
substrate, water distributors

Fl

•

•

•

•

•

•

AM

•

•

•

•

•

•

RE

•

•

•

•

•

RA

WA

•

•

BA

•

BG

•

Water

(Temporary

Lentic)

Temporary
rain pools,
irrigation
ditches

Water temperatures, duration,
frequency, siltation, pollutants,
DO, emergent vegetation,
substrate interface with other
habitats, organic content, water
distribution

AM

•

RE

•

•

•

•

•

WA

•

•

BG

•

Rock

Talus

slopes,

caves, cliffs,

boulders,

substrate,

outcorps,

instream

structure

Rock size, heterogeneity,
interfaces with other habitats,
origin, vertical and horizontal
structure

Fl

•

•

AM

•

•

RE

•

•

SM

•

BA

•

BG

•

•

RA

•

SB

•

•

Soil

Surface and
subsurface
soil types

Soil types, depth, heterogeneity,
horizontal and vertical structure,
interfaces with other habitats

AM

•

RE

SB

RA

SM

Vegetation-
Litter/
Debris

Leaves,
logs, limbs
and other
persistent
and non-
persistent
litter/debris

Litter size, depth, heterogeneity,
horizontal and vertical structure,
type, moisture retention,
temperature.

Fl

AM

•

RE

SB

SM

•

UG

•

Deserts

131

vlajor Vertebrate Associations and Relationship
Df Species' Ecology to Components

Table 2. Important habitat components of desert habitats
(concluded).

CD
f>

C

a>

"cd

Q

CD

>

o
O

CD
Q.
03
O

</>
LU

CD
CL

O
TD

O
O
U_

to
.a

3

CO

a>

c:

T3
CD

CD
U.

en

c
jz
tr
in a>

>- 03

— in
a> 3

CO

o
a)
o
o

</)

>%

-C
0-

g
o

o

Q.
CD

<r

£
o5

3
CO
en

c

o5

CD
CL

>.
O

re

3
05

CD CD
O CO

E^

CD -Q

SI 3

H CO

Microhabitat
Components

Description

Variables/Factors

Vegetation-
Dead

Standing

vegetation,

roots

Size, interface with other
habitats, heterogeneity, vertical
and horizontal structure, soils
(roots only).

Fl

AM

BA

UG

RE

•

SB

•

RA

•

SM

•

A

Vegetation-
Live

All

vegetation
including
roots

Horizontal and vertical structure,
interfaces with abiotic habitats,
heterogeneity

AL

•

•

•

•

•

Vegetation-
Plant
Species

Individual
plant
species
associations

Individual plant species
occurrence and abundance

RE

•

•

•

SB

•

RA

UG

Animal-
Created

Burrows
and other
cover sites

Cover site size, shape, animal
species activity patterns, animal
species size

AM

•

•

RE

•

•

RA

•

•

SM

•

•

CA

Man-made
Structures
(Other than
water and
agricultural)

Bridges,

towers,

electrical

lines,

buildings,

mine shafts,

posts,

fences

Size, type, frequency.

AL

•

•

•

•

•

•

Codes for major vertebrate associations:

AL — all wildlife
AM — amphibians
BA — bats
BG — big game
CA — carnivores
Fl —fish
RA — raptors
RE — reptiles
SB — small birds
SM — small mammals
UG— upland game
WA — waterfowl

132

Deserts

Table 3. Important macrohabitat components/factors in deserts.

Macrohabitat
Components/Factors

Description

Variables/Factors

Microhabitats Affected

Slope

% angle of area from
horizontal

% slope, moisture avail.,
thermal regimes, vegetation
structure

Water, soils, vegetation
debris, vegetation (live and
dead)

Aspect

South, north, east, and west
facing

Direction, moisture avail.,
temperature regimes,
vegetation structure

Water, soils, vegetation
debris, vegetation (live and
dead)

Elevation

Vertical above or below sea
level

Vertical distance, moisture
avail., thermal regimes,
vegetation structure

Water, vegetation debris,
vegetation (live and dead)

Precipitation

All forms

Quantity, type, duration,
frequency, moisture avail.,
thermal regimes, vegetation
structure

Water, soils, vegetation
debris, vegetation (live and
dead)

Ecotones/Habitat
Juxtaposition

Habitat interfaces and
locations

Heterogeneity, interface size,
quantity, position

Vegetation debris and
vegetation (live and dead)

Geographic Location

Major geographic boundaries
and barriers, habitat location

Size, location, and frequency,
habitat size, disjunction,
influence of major faunal
groups in adjacent areas

Direct effect on species
reproductive capabilities

Human Influences

Disturbance and alterations of
habitat due to man

Habitat loss and structural
change

Loss of habitat necessary to
meet ecological needs. Less
diverse structure associated
with habitat alterations

Table 4. Abundance and diversity of habitat components in four deserts.

Code for abundance and diversity: # = high Q = moderate Q = low

Habitat Component

Great Basin

Mojave

Chihuahuan

Sonoran

WATER
Lotic

Permanent Lentic
Temporary Lentic

abundance diversity

o o
o o

© ©

abundance diversity

o o
o o
o o

abundance diversity

o o
o ©
o ©

abundance diversity

• ©

• •
© ©

ROCK

© ©

• •

© •

• •

SOIL

• ©

• ©

• •

• •

VEGETATION
Litter/Debris
Dead
Live
Plant Species

oooo
oooo

o ©

© ©
© ©
o o

o ©

© ©
© ©
o o

© •

• •

• •

o o

ANIMAL-CREATED

© ©

© ©

© ©

© ©

Deserts

133

Water (Lotic)

Running, permanent water (lotic water) is an
extremely important component in desert ecosys-
tems (Naiman and Soltz 1981). Several species, espe-
cially aquatic wildlife such as fish, amphibians, and
some reptiles, are physically dependent on lotic
water sources (Jones 1981a; Lowe 1964; Minckley
1973; Naiman and Soltz 1981). Without this compo-
nent, these types of wildlife cannot exist within des-
erts. Aquatic amphibians, turtles, and snakes require
running water for physiological functions such as
thermoregulation, water balance, escape cover, and
food (Stebbins 1966).

Lotic (running-water) habitat.

Substrate of running streams often determine
the presence and abundance of certain fish species.
For example, the desert sucker (Catostomus clarki)
feeds on rock-bound algae that is present only in
streams and rivers with cobble bottoms (Minckley
1973).

Aquatic sites and riparian areas also provide an
abundant food source. For example, certain raptor

species such as black hawks (Buteogallus anthraci-
nus) and bald eagles (Haliaeetus leucocephalus) fre-
quent deserts only where aquatic food is available
(see Bent 1961 and Millsap 1981). Generally, these
raptors require large fish such as desert suckers, and
these types of prey are restricted to permanent
streams, rivers, and lakes. For these raptors to suc-
cessfully capture prey and nest along streams, rivers,
and lakes, vertical structures such as snags, cliffs, and
trees must be present. This is an example of more
than one habitat component that must be available
for a species to use an area. It also demonstrates im-
portance of habitat edges and land/water ecotones
(see Thomas et al. 1979).

Running streams also provide drinking water for
many species. Many big game species inhabiting des-
erts are dependent on flowing streams for physiologi-
cal water, although lentic (still) water distribution is
usually more important. Yoakum ( 1980) and Kind-
schy et al. ( 1982) stated that American pronghorns,
within the Great Basin Desert, require permanent
water at less than 8-km (5-mi.) intervals.

The quality of water in desert streams also limits
species' occurrence and abundance. Naiman and
Soltz ( 1981 ) discussed the importance of physio-
chemical, heavy metal, stream structure, and macro-
invertebrate ( benthic ) characteristics in determining
fish diversity and abundance. Ratios of riffles, runs,
pools, and backwater subhabitats also determine fish
occurrence in streams much like vertical and hori-
zontal structures in terrestrial habitats (see discus-
sions in Naiman and Soltz 1981 ).

Water (Permanent Lentic)

Permanent, lentic habitats are important to large
numbers of desert-dwelling wildlife. Lentic dwelling
fish, including introduced fishes such as bass (Per-

»'_.

ssfif

Kr

flfinP

'

; . '

Lentic (still-water) habitat.

134

Deserts

cichthyidae sp. ), catfish (Ictaluridae sp. ), and sunfish
(Centrarchidae sp.), are totally restricted to and are
dependent on lakes, stock ponds, and natural ponds
such as cienegas within deserts (Minckley 1973; Nai-
man and Soltz 1981 ). Cienegas, springs, bogs, and
potholes provide the only habitat within deserts for
fish such as desert pupfish (Cyprinodon macularius)
and Gila top minnows (Poeciliopsis occidentalis).

Aquatic amphibians (e.g., Rio Grande leopard
frogs, Rana berlatideri), turtles (e.g., mud turtles,
Kinosternon sp. ), and snakes (e.g., checkered garter
snake, Thamnophis marcianus) are totally depen-
dent on lentic waters within deserts (when lotic
habitats are not present ) for reproduction, food, es-
cape cover, physiological processes such as thermo-
regulation and water regulation, and egg and tadpole
development (frogs and toads).

Lentic waters, like running streams, also provide
an abundant and diverse prey base. For example, cer-
tain bats (Chiroptera) require large amounts of in-
sect prey, which in many desert regions are pro-
vided directly above lentic water sources such as
reservoirs and lakes (see Barbour and Davis 1969 for
Chiroptera requirements ). Raptors such as ospreys
(Pandion haliaetus) and bald eagles are highly de-
pendent on fish (Bent 1961; Millsap 1981), and lakes
are often the only source of these prey in desert
regions.

Most waterfowl are limited to permanent, lentic
habitats within deserts (see Fleming 1959 for exam-
ples). Lentic habitats provide waterfowl with food,
escape cover, and resting sites.

Interfaces of lentic habitat with adjacent habitat,
especially shoreline and emergent vegetation, en-
hance and increase use of lentic habitats by fish,
frogs, toads, snakes, waterfowl, and small birds.
Emergent and shoreline vegetation increase nesting
opportunities for frogs, toads, and waterfowl, and
provide excellent escape cover and feeding substrate
for all wildlife. Floating logs and islands also provide
excellent in-water structures for basking turtles
(thermoregulation and escape cover) and waterfowl
(nesting substrate and isolation from predators),
respectively.

Permanent natural or man-made lentic habitats
(e.g., stock tanks and rainwater catchments) are ex-
tremely important drinking waters for big game spe-
cies inhabiting deserts. Although relatively drought
tolerant, most desert bighorn sheep and pronghorn
require free-standing water, especially during the
breeding season (Wilson et al. 1980; Yoakum 1980).

Water quality, as discussed for lotic habitat, is
equally important in lentic habitats. Dissolved oxy-
gen, temperature, pollutant concentrations, habitat
structure (water depth and horizontal arrangement),

and water availability and distribution determine suc-
cess of many wildlife species in desert habitats (see
Naiman and Soltz 1981 for examples).

Turtle basking on log.

Man-made water catchment.

Water (Temporary Lentic)

Temporary surface water accumulation is com-
mon throughout U.S. deserts, especially on clay soils
( slow percolation ) during summer convectional
storms. In particular, the Chihuahuan Desert is domi-
nated by clay soils, providing large areas of tempo-
rary surface water. Large amounts of surface water
are reflected in this desert's abundant and diverse
amphibian fauna (see Conant 1978). Toad and sala-
mander (e.g., tiger salamander, Ambystoma ti-
grimim ) breeding is tied to summer rain and agri-
cultural pools. Water is necessary for embryonic
development of these animals (see Stebbins 1966 for
species' examples ). These waters also provide escape

Deserts

135

cover and food for many forms of wildlife. For exam-
ple, yellow mud turtles (Kinosternon flavescens) and
waterfowl use temporary pools created by summer
rains or agricultural runoff for escape and resting
cover, and for food (see Fleming 1959 and Stebbins
1966).

Rocky and boulder-strewn canyons also retain
intermittent surface water. Red-spotted toads (Bufo
punctatus) use these canyon pools for breeding (see
Jones 1981a). These natural temporary catchments
are also important physiological water sources for
big game such as bighorn sheep, especially during
warm summer months (Wilson et al. 1980).

Temporary lentic habitat.

Rock

Rock arrangement, heterogeneity, and proximity
to other habitat components affect nearly every spe-
cies of wildlife in deserts (see discussions of Thomas
et al. 1979 and Maser et al. 1979a).

Talus slopes, cliffs, and rock outcrops provide
nesting and feeding substrate, thermal and escape
cover, and resting sites for many forms of wildlife in-
cluding small birds, lizards, snakes, small mammals,
big game, bats, predatory mammals, and raptors.
Rock structure is an extremely important component
for desert bighorn sheep. Deep, rugged cliffs are nec-
essary for lambing, escape, and thermal cover (Wil-
son et al. 1980). Raptors also use these sites as
perches for locating and capturing prey (Call 1978;
Bent 1961). Caves provide excellent resting sites for
bats. Several million bats have been documented in a
single large cave (Connor, personal comm.) because
caves provide stable, moderate temperatures
throughout the year (see Barbour and Davis 1969).

Type and origin of rock also determine species'
use of rock structures (Maser et al. 1979a). For ex-
ample, chuckwallas (Sauromalus obesus) use rock

structures that flake and crack (mostly granitic and
volcanic rock ) as escape cover, bloating themselves
between cracked walls making it extremely hard for
predators to remove them (Stebbins 1966).

Arrangement of different size rock and depth be-
low the surface affect wildlife use in deserts. Burrow-
ing wildlife are especially affected by these factors.
Generally, small burrowing mammals such as kanga-
roo rats (Dipodomys sp. ) inhabit only areas where
surface and subsurface rock is broken or soils are
deep. Other small mammals such as pocket mice
{Perognathas sp. ) and certain ground squirrels (Ci-
tellus sp. ) exploit rocky hills and mountain slopes
because rocks are broken (e.g., volcanic soils), pro-
viding excellent escape, resting, and nesting cover
(see Burt and Grossenheider 1976 and Golley et al.
1975).

Boulder-strewn surfaces (large boulders) pro-
vide excellent feeding, nesting, and resting sites for

Color variations in collared lizards can be linked to differ-
ent habitats.

Another color variation of collared lizard.

136

Deserts

many wildlife. Desert tortoises (Gopherus agassizii)
are common in boulder-strewn areas of west-central
Arizona (Burge 1979). Boulders seem to provide tor-
toises with excellent cover sites (located below
rocks) and a cryptic substrate. Large boulders also
provide small birds and predatory lizards (e.g., col-
lared lizards, Crotaphytus collaris ) with perching
sites for obtaining food.

Rocks are also extremely important for thermo-
regulation in reptiles, especially lizards. Simon and
Middendorf ( 1976) showed that Yarrow's spiny liz-
ards (Sceloporus jarrovi) were able to attain core
temperatures necessary for activity during winter
months because rocky, south-facing slopes acted as
heat sinks. Rocky hillsides also provide caverns in
which rattlesnakes (Crotalns sp. ) and other snakes
hibernate (see Hirth et al. 1969).

Rocky slopes and hillsides also contain many
small, moist microhabitats created by water buildup
and shading. Cumulatively, this phenomenon creates
a high degree of heterogeneity in vegetation struc-
ture and composition which in turn increases inver-
tebrate production and diversity (see Maser et al.
1979a).

Soil depth, type, heterogeneity, and interface
with other habitat components such as rock and veg-
etation, determine subsurface space for wildlife.
Some soils such as clay, and in areas with bedrock,
are impenetrable to certain burrowing wildlife (e.g.,
green toad, Bufo debilis). On these sites, roots of
vegetation are important in breaking up rock and
soil. In parts of the Chihuahuan and Sonoran Deserts,
creosotebush breaks up clay soils and bedrock, pro-
viding wildlife access to otherwise unavailable space
(Mares and Hulse 1977; Barbour et al. 1977).

Similar to surface and arboreal edges created by
landform and vegetation, greater soil heterogeneity
and interfaces with other habitat components (e.g.,
rock) provide greater numbers of spatial niches, result-
ing in higher wildlife biomass and diversity, especially
amphibians, reptiles, and small mammals (see Thomas
et al. 1979 for surface and arboreal examples).

Wildlife use subsurface niches for a wide variety
of ecological functions, and several species depend
on soil characteristics for exploitation of desert re-
gions. For example, spadefoot toads (Scaphiopus sp. )
are dependent on deep, loose soils to burrow and es-
tivate (McClanahan 1967). Spadefoot toads remain

Soil

The number of subsurface (fossorial) niches in
deserts is largely dependent on soil characteristics.
Lack of vegetation structure in deserts is often offset
by subsurface space created by deep and diverse
soils, especially in warm deserts (see discussions of
Bender 1982, Brown 1982b, and Attenborough
1976). Creosotebush stands of the Chihuahuan and
Sonoran Deserts have little vegetation structure, but
have a rich fauna, especially amphibians, lizards,
snakes, and small birds and mammals (Bender 1982;
Brown 1982b), primarily because of soil diversity
and access to subsurface space.

Spadefoot toad.

Badger at burrow.

burrowed in these soils for the entire year, except
during warm-season convectional storms when they
surface to feed and breed. Similarly, small mammals
are affected by soil type, depth, and heterogeneity
(Golley et al. 1975). For example, many kangaroo
rats are generally incapable of burrowing into hard,
coarse, shallow soil, except along deeply rooted
plants (see Burt and Grossenheider 1976 for exam-
ples), and are limited to deserts with deep, loose,
and gravel substrate.

Deserts

137

In addition to providing resting sites, certain
soils provide nesting, thermoregulatory substrate,
and escape cover. For example, fringe-toed lizards
{Uma sp.) exploit deserts with fine sandy soils. Be-
sides supplying nesting substrate, fine textured sand
provides fringe-toed lizards cryptic background and
warm thermal regimes necessary for activity (see
Norris 1958).

Water movement through different soils (perco-
lation) also affects species' existence in deserts.
Jones et al. ( 1983) and Conant ( 1978) showed that
deserts with clay soils accumulated more surface
water than areas with loose, gravelly, and rocky soils,
resulting in larger desert amphibian faunas.

Soil depth, origin (e.g., volcanic), and heteroge-
neity also affect vegetation composition and struc-
ture within deserts (Davis 1976). Vegetation compo-
sition and structure are discussed later in this
chapter.

Vegetation — Litter/Debris

Ground debris or litter provides excellent feed-
ing, thermoregulatory, nesting substrate, and escape
cover for wildlife inhabiting desert regions. Vegeta-
tion debris consists of dead plant material (e.g., tree
limbs and leaves) that accumulate on the surface,
and it is generally correlated with vegetation compo-
sition and abundance (Daubenmire 1974). Litter
abundance and diversity are probably highest in the
Sonoran Desert and least in sections of the Great
Basin Desert.

In deserts, large numbers of invertebrate species
inhabit ground litter, providing abundant prey for
ground-foraging amphibians, lizards, and small birds
(see Pianka 1970 and Tomoff 1974 for examples).

Downed limbs and logs are important nest sub-
strate for many lizards, snakes, and small mammals,
providing cool, moist, well-concealed sites within
hot deserts. Certain species such as Gilbert's skink
(Eumeces gilberti) are highly dependent on these
microhabitats for maintaining relatively low pre-
ferred body temperatures within deserts (Jones and
Glinski 1985). Pack rats (Neotoma sp. ) require var-
ious forms of litter to build nests, and their occur-
rence within desert regions seems to be related to
occurrence of vegetation debris (Burt and Grossen-
heider 1976).

Vegetation debris is extremely important in
aquatic desert habitats. Downed logs, limbs, and
other litter provide excellent thermal and resting
cover for amphibians, turtles, snakes and fishes, and
contribute organic nutrients to aquatic systems (Nai-
man and Soltz 1981).

Because wildlife respond to several physical

Pack rat nest.

attributes of litter in deserts, it is important to re-
cord the type of litter, depth, width, and frequency.

Vegetation — Standing — Dead

Standing, dead vegetation or snags provide ex-
cellent vertical structure for nesting, feeding, resting,
thermoregulatory substrate, and escape cover within
deserts. Many cavity nesting birds (e.g., Gila wood-
peckers, Melanerpes uropygialis) and raptors (e.g.,
screech owl, Otus asio) use standing, dead vegeta-
tion for nest substrate, roosting, and resting sites
(see Knopf 1977 for examples). Small lizards such as
tree lizards (Urosaurus ornatus) use standing, dead
vegetation for foraging, defense, and thermoregula-
tion substrate (Vitt et al. 1981 ).

Similar to downed litter, standing, dead vegeta-
tion abundance and diversity are positively associ-
ated with vegetation composition and structure (see
Daubenmire 1974). Therefore, Sonoran Desert re-
gions, especially the Arizona Upland subdivision,
probably possess the highest abundance and diver-
sity of snags among major U.S. deserts.

Vegetation — Live

Live vegetation structure is perhaps the most
important habitat component in desert regions, al-
though soils and rock may be more significant where
vegetation is low and homogeneous. A number of
vegetation variables should be sampled when con-
ducting inventories and monitoring in deserts.

Arrangement of vertical and horizontal vegeta-
tion structure in deserts, and its relations to wildlife,
has been thoroughly studied (see Rottenberry and
Wiens 1980 for an example). Most authors agree
that high wildlife abundance and diversity accom-
pany deserts with high horizontal and vertical vege-

138

Deserts

tation structure, especially birds. Tomoff (1974) cor-
related avian species diversity increases to increases
in vegetation structure within the Sonoran Desert.
Greater vertical structure increases the number of
avian nesting, feeding, and breeding niches within
these habitats. Conversely, deserts with little vertical
vegetation structure provide few spatial niches. Rap-
tors are also positively associated with vegetation
structure (Millsap 1981).

Certain lizards and bats also benefit from verti-
cal vegetation structure (Pianka 1966; Barbour and
Davis 1969). Vitt et al. (1981) demonstrated that
woody vegetation was important reproductive and
feeding substrate for arboreal lizards (tree lizard,
long-tailed brush lizard [Urosaurus graciosus], and
the desert spiny lizard [Sceloporus magister]). Verti-
cal vegetation structure provides bats with greater
amounts of roosting and resting space (Barbour and
Davis 1969), and it also increases abundance of prey
(flying insects) for insectivorous lizards, small birds,
and bats. Vertical vegetation structure is also used by
lizards for thermoregulation. Norris (1953) de-
scribed climbing behavior of desert iguanas (Dipso-
saurus dorsal is) into creosote to escape rapidly in-
creasing surface temperatures; temperatures 1-2 m
(3-6 ft) above the surface are considerably lower.

Horizontal vegetation structure, in conjunction
with vertical structure, also affects species abun-
dance and diversity (Rottenberry and Wiens 1980).
Generally, greater varieties of horizontal patchiness
are associated with greater wildlife diversity. Rotten-
berry and Wiens ( 1980) suggested that certain birds
prefer either open or closed habitats. By providing
highly variable horizontal structure, desert habitats
support species that prefer both open and closed
habitats. Deserts possessing homogeneous, horizontal
arrangement, and horizontal structure that is ex-
tremely opened or closed probably have lower avian
diversity than those that are structurally variable.

Similarly, lizards and snakes respond to horizon-
tal patchiness, although their relationship to patchi-
ness is related to foraging and thermoregulatory re-
quirements. Certain lizard species require alternating
open and closed habitat configurations ( Pianka
1966). These lizards forage in open spaces and re-
treat to large, dense bushes between foraging bouts
to lower their body temperatures. Deserts transected
by intermittent washes provide these alternative
structures.

Horizontal patchiness also conceals nests of up-
land game such as Gambel's quail {Lophortyx gam-
belli) and sage grouse, and provides escape cover
for individual birds (see Goodwin and Hungerford
1977 for Gambel's quail and Eng and Schladweiler
1972 for sage grouse).

Composition and availability of vegetation in des-

erts affects species' distribution and diversity. Pres-
ence and abundance of certain plant species affect
horizontal and vertical structure (as previously dis-
cussed), and also provide food for commercially im-
portant species such as bighorn sheep, Gambel's
quail, herbivorous lizards, birds, and mammals.

Availability of perennial and annual vegetation
affects bird species abundance and diversity in desert
regions (Serventy 1971; Tomoff 1974). Generally,
bird species' diversity increases with greater mix-
tures of annual and perennial vegetation because of
great variety in avian food preferences (Martin et al.
1951). Similarly, abundance, type, and diversity of
vegetation affects food available to herbivorous, des-
ert rodents; greatest herbivorous rodent diversity ex-
ists on sites with high annual and perennial diversity
and mixture (Black 1968; Brown 1973).

Big game use of desert regions are affected by
plant species composition and abundance. For exam-
ple, desert bighorn sheep require a high diversity of
forbs, perennial grass, and shrubs in desert habitats
to meet basic metabolic needs (Wilson et al. 1980).

Mi*53*' ■ ■' ***&*

i$

Desert sheep select areas near rocky escape cover.

Although wildlife diversity generally increases
with increases in plant species diversity, increased
plant diversity may reduce numbers of important
game species. For example, Eng and Schladweiler
( 1972) found that sage grouse preferred closed
stands of sagebrush.

Extensive root systems of certain plants, such as
creosotebush, provide access to subsurface space for
toads, salamanders, lizards, snakes, and small mam-
mals. Species diversity is often increased in areas
with extensive plant root systems.

Deserts

139

Vegetation — Plant Species Associations

Certain wildlife species have evolved morpho-
logical, physiological, and behavioral characteristics
that are adaptive to certain plant species attributes.
These should be carefully considered when sampling
desert wildlife. There are several good examples of
these relationships in deserts. Desert iguanas forage
heavily on creosotebush buds, especially in the
spring, and their distribution is closely tied to occur-
rence of creosotebush (Norris 1953). Sage grouse
rely heavily on sagebrush buds, especially in the
spring, and this species is highly associated with
dense stands of sagebrush throughout its range ( Eng
and Schladweiler 1972). Although both the desert
iguana and sage grouse are dependent on these plant
species for food, they also benefit from escape and
thermoregulatory cover provided by these plants.

Wintering sage grouse feed extensively on sagebrush buds
and seedheads.

Cactus wren nest in cholia cactus.

Certain plant species also provide nesting struc-
ture for desert wildlife. Occurrence of elf owls (Mi-
crathene whitneyi) in the Sonoran Desert is associ-
ated with nesting cavities presence in saguaro cactus
(Knopf 1977; Bent 1961 ). Similarly, cactus wrens
(Campylorhynchus brunneicapillus) are associated
with cactus and yuccas, primarily because these
plants provide nesting structure (Knopf 1977).

Animal-Created Structures

Structures created by certain wildlife provide
other desert animals with habitat necessary for
breeding, feeding, escape, resting and thermoregula-
tory substrate, and escape cover.

Fossorial mammals construct vast networks of
subsurface burrows that provide resting and nesting
habitat structure for a number of desert dwelling
wildlife such as toads, lizards, snakes, and raptors
(burrowing owl {Athene cunicularia]). Subsurface
space contributes greatly to species diversity in des-
erts with little vegetation structure.

Primary cavity nesting species such as wood-
peckers provide nesting habitat in cactus and woody
vegetation (e.g., mesquite) for a number of second-
ary cavity nesting birds such as elf owls (see Knopf
1977 for examples).

Beavers modify flow of desert streams, providing
pool and backwater habitat, and enhance populations
of fish, toads, frogs, aquatic turtles, shorebirds, and
waterfowl (Burt and Grossenheider 1976). It is im-
portant to consider animal-created structures when
attempting to sample and predict wildlife occurrence
within deserts.

Man-Made Structures

Man-made structures have both positive and
negative effects on wildlife in desert environments.
Although power lines provide nesting, resting, and
roosting structure for birds and raptors (e.g., red-
tailed hawk, Buteo jamaicensis), they also cause
mortality due to electrocution and obstruction (see
Olendorff et al. 1981).

Man-made structures such as buildings, barns,
mine shafts, and fences provide a wide range of nest-
ing, feeding and thermoregulatory substrate, and es-
cape cover for wildlife such as barn owls (Tyto
alba), fence lizards (Sceloporus sp. ), bats, mourning
doves (Zenaida macroura), and house mice (Mus
mmculus). Vertical mine shafts and fences have
some negative effects on desert wildlife. Many small
wildlife species are trapped in shafts, and fences pre-
sent physical hazards and movement restriction to
big game species (see Yoakum 1980 for pronghorn
examples).

140

Deserts

A number of man-made structures such as water
catchments and nesting platforms have been used in
desert regions to enhance wildlife. Maser et al.
( 1979b) provided an excellent synopsis of man-
made habitats and their effect on wildlife. All man-
made habitats should be considered when sampling
deserts. They often affect the presence or absence of
certain species within deserts.

Slope

Because of diverse topography, slopes vary
greatly within desert habitats. Slope variability and
mixture in deserts contribute greatly to overall habi-
tat variability and diversity of wildlife faunas (see
Maser et al. 1979a for examples). Wind currents, di-
rect and indirect heating of substrate, and moisture
retention associated with different slope angles di-
rectly (animal heat and water exchange, and ease of
movement) and indirectly (vegetation differences)
affect wildlife (see Maser et al. 1979a). The interrela-
tionship between slope and soils, vegetation, aspect,
and elevation determines the physical structure of
desert ecosystems. Therefore, effects of slope on
physical structures in deserts will vary, depending on
interfaces with these other habitat components. Dif-
ferences in slopes often contribute to presence, ab-
sence, and abundance of certain wildlife species.
Whereas steep, rugged slopes are generally a require-
ment of bighorn sheep (Wilson et al. 1980), they
preclude occurrence of many other species such as
horned lizards (Phtynosoma sp. ). Slopes should be
measured and described when conducting invento-
ries and monitoring in deserts.

The Sonoran Desert exhibits many diverse habitat features.

Aspect

Aspect involves the compass direction of land-
scape. Similar to slope, aspect affects moisture availa-
bility, air and soil temperatures, which in turn affect
vegetation composition and structure, and wildlife
thermal and moisture regimes (see previous discus-
sions on these microhabitat components for effects
on desert wildlife). Effects of aspect on desert envi-
ronments also depend on interfaces with microhabi-
tat components, especially soil and vegetation, and
macrohabitat factors such as slope, precipitation, and
elevation. Generally, vegetation structure is greatest
along north- and east-facing slopes because of lower
temperatures and higher moisture regimes. However,
soil type, rock, and elevation may produce contrary
results. Aspect should be considered when selecting
sample locations within deserts.

Elevation

Elevation affects moisture availability and temper-
atures that in turn affect vegetation structure and com-
position (see Lowe 1964 for examples). It also affects
thermal and moisture regimes of air, soil, and litter,
which can affect wildlife (see earlier discussions).

Generally, higher elevations have cool weather,
greater soil and litter moisture, and more precipita-
tion. Increased vegetation structure resulting from
greater moisture and cooler weather can be offset by
low temperatures associated with higher elevations.
For example, high desert regions of Wyoming have
very few species of amphibians, reptiles, and small
mammals because environmental temperatures are
too low for species' thermal maintenance. Food nec-
essary for birds and other herbivorous and insectivo-
rous wildlife are too infrequent (short growing sea-
son) for physiological maintenance in these cold
desert regions.

Similar to slope and aspect, elevational attributes
are affected by association with soil, vegetation,
slope, aspect, and degree of mountain rain shadow.

Precipitation

Precipitation varies greatly within deserts (see
Bender 1982), and this directly affects vegetation
structure and moisture available to wildlife in air,
soils, and litter (see previous discussions on these
microhabitat components; Oosting 1956). Precipita-

Deserts

141

tion within deserts is primarily affected by mountain
rain shadow and elevation (see Bender 1982).

Frequency and type of precipitation also affect
vegetation structure and moisture available to wild-
life. Northern deserts (Great Basin) receive large
amounts of snow (Bender 1982). Moisture is gener-
ally not available during warm spring and summer
months for plant growth and wildlife consumption.
Conversely, southern deserts that receive large pro-
portions of their precipitation as rain (e.g. Chihua-
huan and Sonoran Deserts) in summer months (see
Bender 1982), provide water for plant growth, insect
production, amphibian breeding, and wildlife con-
sumption. Large snow packs at higher elevations in
the Great Basin Desert (e.g., Wyoming) also reduce
plant food available to wildlife, except on south-fac-
ing winter ranges (see Skovlin 1982 for an example).

Because precipitation quantity type and fre-
quency can radically affect wildlife activity, year-to-
year fluctuations in precipitation are important when
interpreting inventory and monitoring results.

Ecotones/Habitat Juxtaposition

Throughout, I have mentioned how edges and
position of habitats affect habitat structure and wild-
life diversity. Generally, wildlife diversity is highest
in deserts with high mixtures of different microhabi-
tat components, and favorable thermal and moisture
regimes (warm and moist). The Sonoran Desert fully
demonstrates this relationship. Great mixtures of to-
pography, soils, rock and vegetation structure, warm
thermal regimes, and moisture provide many differ-
ent spatial and temporal niches to wildlife (reflected
in one of the world's most diverse wildlife faunas ).

Ecotones consisting of habitat elements from des-
erts and adjacent habitats (e.g., pinyon-juniper) cre-
ate distinct physical elements (e.g., structure) that
enhance species diversity (Thomas et al. 1979). For
example, breeding bird diversity is increased in eco-
tones between sagebrush and pinyon-juniper habitats
because of greater habitat structural diversity (both
tree and shrub life forms).

Geographic Location

Geographic location and past geological history
account for present species distribution and diversity
of some wildlife within desert regions. For example,
species diversity in the Sonoran Desert is greatly en-
hanced by contributions of adjacent subtropical and
tropical avifauna ( Brown 1982b; Bender 1982). The
evolution of the Sonoran Desert from mostly subtropi-
cal vegetation also accounts for occurrence of other
wildlife with more southerly ties. Similarly, Great Basin
wildlife diversity is increased by species from adjacent
pinyon-juniper and grassland communities.

Geographic boundaries associated with deserts
are also responsible for present distributions. Abrupt
mountain ranges and canyons (e.g., Grand Canyon)
have isolated certain desert wildlife (Golley et al.
1975). Current scattered fish distributions within
many desert regions reflect a drying trend since the
last ice age (10,000 years ago; Naiman and Soltz
1981 ). Many streams have dried up, leaving disjunct
fish populations.

Past and present weather and associated vegeta-
tion patterns have also affected species' occurrences
throughout deserts. Van Devender and Spaulding
(1979) demonstrated recent drying trends in the
West (past 10,000 years), resulting from recession of
the last ice age. This drying trend left relict and dis-
junct woodland on mountains throughout southwest-
ern deserts. Several woodland vertebrates remain on
these relict sites, with diversity dependent on habitat
size and distance from major stands (see Martin
1980, 1981; Jones et al. 1985).

Human Influences

Man has drastically affected plant and animal
composition and diversity throughout U.S. deserts. Ir-
rigation, urbanization, and other developments have
severely reduced natural desert habitats. Artificial, ur-
ban, and agricultural habitats support fewer species
than do native surrounding deserts (Davis 1973;
Landcaster and Rees 1979). Other land uses have
also reduced quality of desert habitats such as off-
road vehicles, livestock grazing, and mineral develop-
ment (Webb and Wilshire 1983; Jones 1981b; and
Dunaway 1971). Fire prevention and channelization
of desert streams and rivers have also reduced natu-
ral characteristics in desert ecosystems. See the fol-
lowing section for greater detail on impacts to desert
habitats.

MAJOR IMPACTS ON DESERT HABITATS

There are two major types of impacts on desert
habitats, natural and man-caused. I have already dis-
cussed, with the exception of fire and flooding, how
natural factors affect wildlife distributions and com-
position in deserts.

Fire affects wildlife within deserts by altering
habitat structure (Wright and Bailey 1982). Gener-
ally, vegetation on a site is limited to annual grasses
and sprouts from perennial shrubs and grasses up to
4 years after a fire. This favors species that prefer
open sites and areas with annual grasses and forbs,
and is less favorable to species requiring shrubs and
trees. However, fires in deserts do not generally
carry over large areas; burns are usually patchy
(Wright and Bailey 1982). This result tends to in-
crease species diversity in an area due to increased
horizontal and vertical vegetation structural diver-

142

Deserts

sity. For example, in deserts such as the sagebrush
communities of the Great Basin Desert, fire opens up
homogenous vegetation stands.

Fire does not seem to result in high direct mor-
tality to wildlife (Vogl 1977). Most animals are able
to escape rapidly burning, low intensity desert fires.
In summer months following a wet winter or spring,
annual grasses are more abundant, resulting in larger
and more frequent fires. However, since fuels in des-
erts are mostly annual grass, fires burn very quickly
over a site, with little direct wildlife mortality.

Flooding is another natural phenomena that af-
fects wildlife habitat, especially in desert riparian sys-
tems. For the most part, flooding is beneficial to
wildlife habitat. In cottonwood-willow (Populns-
Salix sp. ) communities of the Southwest, flooding is
required for establishment of deciduous tree seed-
lings. When flooding is reduced or eliminated by in-
stream structures such as dams, reproduction of
these important tree species are reduced or elimi-
nated. Reduction or loss of trees will reduce wildlife
diversity on the site.

Flooding occurs as a result of short, intense summer thun-
derstorms. Most of this precipitation runs off without pen-
etrating the soil.

Similar to fire, species that occupy riparian habi-
tats evolved with these natural perturbations. There
is probably little direct mortality of wildlife during
most floods, although 100- and 500-year floods may
affect sedentary wildlife such as small mammals and
reptiles.

Like natural perturbations of desert habitats,
man-caused alterations affect wildlife both directly
and indirectly. Noise and activity from off-road vehi-
cle use and mining activity such as drilling have
been shown to affect reproductive success of wildlife
within an area. For example, drilling activities con-
ducted in a bighorn sheep lambing ground may

cause ewes to abort their young (Wilson et al.
1980). Similarly, Bury and Luckenbach ( 1983) found
that continual off-road vehicle use in an area reduces
the number of breeding birds.

Fences and roads also have a direct effect on
wildlife within desert habitats. Both types of devel-
opment reduce or eliminate wildlife movement
within an area. For example, the fitness of a prong-
horn population in a desert area can be reduced by
cutting off natural migrational routes by fencing
(Yoakum 1980).

The most extensive and severe impacts to wild-
life are those that occur from loss of habitat and hab-
itat quality. Of all habitats, deserts are probably the
most severely affected by domestic livestock grazing.
Low, erratic precipitation and extreme environmen-
tal temperatures reduce the ability of most desert
plants to handle persistent livestock grazing, espe-
cially around water developments where livestock
tend to congregate. In addition, the Mojave, Sonoran,
and Chihuahuan Deserts historically had sparse wild
ungulate populations (Schmidt and Gilbert 1978).
Therefore, most desert plants have not evolved with
grazing pressures. An indicator of excessive livestock
use in most deserts is reduced amounts of forbs and
perennial grasses. In some parts of the Sonoran Des-
ert, perennial grasses are absent, primarily due to
long-term, year-round livestock use. Other condi-
tions such as reduced shrub and tree cover (Sonoran
Desert) and plant species diversity can indicate ex-
cessive livestock use.

The direct impact of livestock on desert vegeta-
tion results in reduced wildlife diversity primarily
due to habitat structural losses. Excessive livestock
grazing also reduces food available to herbivorous
wildlife, often resulting in reduced wildlife abun-
dance and diversity.

Other impacts that indirectly affect wildlife
through habitat disturbance or loss include mining,
road construction, off-road vehicle use, and other ac-
tivities that cause surface disturbance.

In addition to the type of land use, the severity
of the impact depends on the size of the develop-
ment relative to the total area of habitat and the du-
ration of the project. Generally, a short-duration pro-
ject that disturbs only a small portion of the total
habitat does not significantly reduce the area's over-
all species diversity. However, even the smallest de-
velopment can significantly reduce or eliminate a
species that historically depends on a specific area
(e.g., bighorn sheep lambing ground within the So-
noran Desert or elk winter range on a south slope
within the Great Basin Desert).

When a land use modifies or eliminates a large
portion of a habitat, even for a short period of time,

Deserts

143

species diversity can be drastically reduced. Recent
studies in animal biogeography indicate that species
require some minimum habitat size (see Martin
1980, 1 981; Jones et al. 1985).

When a large portion of a habitat is lost, certain
species are extirpated because habitat size is re-
duced below that required to maintain minimal via-
ble populations. This type of impact is most acute in
small, isolated habitats where loss of only a few
hundred acres represents a significant portion of the

Off-road vehicle damage to wildlife habitats.

total habitat. For example, disturbance and loss of
sand dune habitat in the Algodones dunes of south-
eastern California from intensive off-road vehicle use
has significantly reduced fringe-toed lizard popula-
tions (Bury and Luckenbach 1983). Other small, iso-
lated habitats within North American deserts whose
faunas can be drastically reduced by small man-
caused perturbations include woodland on top of
mountains and riparian sites. Jones et al. ( 1985)
found that reptile diversity on island woodlands of
the Sonoran and Mojave Deserts were highly depen-
dent on the size of the island.

In summary, impacts of land use on desert habi-
tats are generally longer-lived than in other ecosys-
tems, primarily because low, erratic precipitation
precludes rapid recovery.

SUMMARY OF FACTORS AFFECTING
SAMPLING IN DESERTS

Several factors affect sampling of desert wildlife
faunas and their habitats. First and most important,
biologists must be aware that most desert wildlife re-
spond quickly to environmental changes, especially
amphibians, reptiles, and small mammals. The life-his-
tory strategies of most of these animals are "boom

and bust," taking advantage of favorable environmen-
tal conditions (Attenborough 1976). Because of
these strategies, results of surveys and studies may
vary directly with seasonal and yearly climatic condi-
tions under which samples were taken. Therefore, bi-
ologists should be cautious in interpreting inventory
and monitoring studies in deserts. A one-time inven-
tory of an area, preceded by a dry year, may fail to
verify several species. Similarly, results of a monitor-
ing study may be affected by fluctuations in precipi-
tation. For example, a biologist attributes reduced
abundance of small mammals on a site to increases
in off-road vehicle use, when in fact, reduction in
small mammal numbers is caused by a previously dry
year. It is very important for one to consider re-
sponses of animals to environmental fluctuations
when designing wildlife studies in deserts.

In collecting data on distribution and habitat as-
sociations of desert wildlife, biologists must be aware
of the diverse ways in which these animals use des-
ert resources, especially when deciding which habi-
tat components to measure. For example, soil types,
vegetation debris, and water availability play large
roles in desert species diversities.

Microhabitat components should be considered
when selecting sample site location. For example,
the Arizona Upland subdivision of the Sonoran Des-
ert has animals arranged in several microhabitats.
Therefore, to adequately sample the faunas of this
area, the biologist should select sample sites so that
all microhabitats are represented (see earlier discus-
sion for important components).

Because animal abundance and diversity also
vary with aspect and slope within deserts, it is im-
portant to select sample sites that will represent vari-
ation in these factors.

Sample size for inventories and monitoring stud-
ies within deserts vary with habitat size and hetero-
geneity. For example, fewer samples are needed to
represent large, homogeneous stands of sagebrush in
the Great Basin Desert or creosote in the Chihua-
huan Desert, than in relatively diverse stands of the
Sonoran Desert. Generally, larger areas require more
samples than smaller areas to compensate for area-re-
lated sample variability.

Temporal ( time ) partitioning of desert resources
also plays a large role in the composition of desert
wildlife (see Creusere and Whitford 1982). There
are many different activity phases of desert wildlife,
especially nocturnal ones. One must become familiar
with different activity periods to obtain complete
lists of desert wildlife, or to study responses of cer-
tain species to land management.

144

Deserts

Certain sampling techniques are more effective
in deserts than others. Because many reptiles, am-
phibians, and small mammals are nocturnal, these an-
imals are most effectively sampled by live and kill
trapping, although roadriding is also a popular way
of sampling these animals. Trapping also allows the
biologist to capture secretive, burrowing forms.

Although effective in woodland and grassland
habitats with large big game herds, pellet group tran-
sects are not effective for sampling the smaller herds
of desert big game species. For example, in an area
with 10-15 deer per section, several hundred 1/100-
acre circular plots would be needed to adequately
sample pellet groups. Aerial count methods may pro-
vide a better estimate of big game density.

Because most deserts are dominated by shrubs,
and to a lesser extent trees and grasses, I recom-
mend using a line-intercept transect method, or
point-intercept method along a line to sample wild-
life habitat variables within deserts. This will give a
good estimate of habitat frequency, abundance, and
horizontal and vertical cover. When monitoring, sam-
ple points should be permanently marked. If grass
and forb estimates are needed, a combination of a
point-intercept along a line and plots is useful. If
shrub density is desired (e.g., for sage grouse), a
point-center quarter method is recommended. See
Chapter 5 for greater detail of habitat measurement
techniques.

Great Basin Desert inventories and monitoring
should be conducted between May and September,
except when working with wintering big game pop-
ulations. In the Mojave, Chihuahuan, and Sonoran
Deserts, sampling is most effective between April
and June, and again in September and October, al-
though certain raptors nest in the late winter and
early spring within these deserts. Summer months
are generally too warm to sample within southern
U.S. deserts.

Summer rains within deserts also provide an ex-
cellent opportunity to sample many small nongame
wildlife. For example, many burrowing snakes and
toads are active immediately after rain. Toads com-
monly breed in temporary rainwater within deserts.
Temporary rainwater pools should be checked at
night to verify certain toad species.

Deserts are extremely warm during summer
months and can be bitter cold during winter, espe-
cially in the Great Basin Desert. Generally, most
wildlife greatly reduce their activity during these pe-
riods. Surveys and studies should be scheduled
around extreme environmental conditions.

Most desert regions are remote with limited ac-
cess and few towns. A wide variety of transportation
and sampling schemes may be necessary to counter-
act limited access. For example, helicopters provide
rapid and efficient means of mapping desert habitats.
One might also consider placing several samples
closely together rather than spreading them out, pro-
vided random sampling criteria are considered.

Rattlesnakes are very common within deserts,
especially in southern deserts during the spring. Care
should be taken when moving through densely vege-
tated areas while walking bird transects in the early
morning or while setting small mammal trap lines in
the early evening.

As in other habitats, inventories and monitoring
studies of deserts are affected by objectives. Chapter
1 provides a discussion of how survey and study ob-
jectives affect sampling. Objectives developed for
studies and surveys will also affect species and habi-
tats to be sampled.

In summary, deserts provide tremendous oppor-
tunities to study relationships between animals and
their environment. But careful consideration must be
given if inventories and monitoring are to answer
specific management questions.

Deserts

145

LITERATURE CITED

ATTENBOROUGH, D., ed. 1976. Deserts and Grasslands.
Doubleday and Company, Inc. New York, NY.

BAILEY, R.G. 1978. Description of ecoregions of the
United States. U.S. Dep. Agric, For. Serv. Ogden, UT.
77pp.

BARBOUR, M.G., J.A. MAC MAHON, S.A. BAMBERG, and
J.A. LUDWIG. 1977. The structure and distribution of
Larrea communities. Pages 227-251 in Mabry, T.J., J.H.
Hunziker, and DR. DiFreo, Jr., eds. Creosote Bush.
U.S./IBP Synthesis Series 6. Dowden, Hutchison and
Ross, Inc. Stroudsburg.

BARBOUR, R.W. and W.H. DAVIS. 1969. Bats of America.
Univ. Kentucky Press. Lexington. 286pp.

BENDER, G.L., ed. 1982. Reference handbook on the de-
serts of North America. Greenwood Press. Westport,
CT. 594pp.

BENT, AC. 1961. Life histories of North American birds of
prey. Vol. 1 398pp. Vol. 2 466pp. Dover Publ., Inc.
New York, NY.

BLACK, H.L. 1968. Populations of small rodents in relation
to grazing by cattle on foothill rangeland. Masters
Thesis, Univ. of Utah. Provo.

BROWN, D.E. 1982. Introduction. Pages 8-16 in Brown,
D.E., ed. Desert Plants: Biotic Communities of the
American Southwest — United States and Mexico. Univ.
of Arizona Press. Tucson.

. 1982a. Chihuahuan desert scrub. Pages 169-179 in

Brown, D.E., ed. Desert Plants: Biotic Communities of
the American Southwest — United States and Mexico.
Univ. of Arizona Press. Tucson.

. 1982b. Desert plants: biotic communities of the

American Southwest — United States and Mexico. Univ.
of Arizona Press. Tucson. 342pp.

, C.H. LOWE, and C.P. PASE. 1979. A digitized clas-
sification system for the biotic communities of North
America, with community (series) and association ex-
amples for the Southwest. J. Arizona Acad. Sci. 14
(Suppl):l-16.

BROWN, J.A. 1973- Species diversity of seed-eating desert
rodents in sand dune habitats. Ecology 54:775-787.

BROWN, J.H. and AC. GIBSON. 1983- Biogeography. C.V.
Mosby Co. St. Louis, MO 492523pp.

BURGE, B.L 1979. Survey of the present distribution of
the desert tortoise, Gophenis agassizii, in Arizona.
U.S. Dep. Inter., Bur. Land Manage. Contract YA-5 12-
CT8-108. Denver, CO.

BURT, W.H. and R.P. GROSSENHEIDER. 1976. A field

guide to the mammals. Houghton-Mifflin. Boston, MA.
289pp.

BURY, R.B. and R.A. LUCKENBACH. 1983. Vehicular recre-
ation in arid land drives, biotic responses and manage-
ment alternatives. Pages 217-221 in Webb, R.H. and
H.G. Wilshire, eds. Environmental Effects of Off-Road
Vehicles. Impacts and Management in Arid Regions.
Springer-Verlag. New York, NY.

CALL, M.W. 1978. Nesting habitats and surveying tech-
niques for common western raptors. Tech. Note 316.
U.S. Dep. Inter., Bur. Land Manage. BLM Service Cen-
ter. Denver, CO.

CONANT, R. 1978. Semiaquatic reptiles and amphibians of
the Chihuahuan Desert and their relationships to
drainage patterns of the region. Pages 445-491 in
Waver, R.H. and D.W. Riskind, eds. Trans. Symp. Biol.
Resour. Chihuahuan Desert Region, U.S. and Mexico.
U.S. Dep. Inter., Nat. Park Serv. Proc. Trans. Ser. 3-

CREUSERE, F.M. and W.G. WHITFORD. 1982. Use of time
and space by lizards, in Scott, N., ed. Herpetological
communities, U.S. Dep. Inter., Fish and Wildl. Serv.
Wildl. Res. Rep. 1 3.

DAUBENMIRE, R.F. 1974. Plants and environment: a text-
book of autecology. Wiley and Sons, Inc. New York, NY.

DAVIS, G.P. Jr. 1973- Man and wildlife in Arizona: the pre-
settlement era, 1823-1864. MS Thesis. Univ. of Ari-
zona. Tucson.

DAVIS, R.M., ed. 1976. National range handbook. U.S. Dep.
Agric, Soil Cons. Serv. Washington, DC.

DUNAWAY, D. 1971. Human disturbances as a limiting
factor of Sierra Nevada bighorn sheep. N. A. Wild.
Sheep Conf. Trans. 1:165-173.

ENG, R.L. and P. SCHLADWEILER. 1972. Sage grouse win-
ter movements and habitat use in central Montana. J.
Wildl. Manage. 36( 1 ):14 1-146.

FLEMING, W.B. 1959. Migratory waterfowl in Arizona.
Ariz. Game and Fish Dep. Wildl. Bull. 5. Phoenix.

GOLLEY, F.B., K PETRUSEWICZ, and L. RYSZKOWSKI.
1975. Small mammals: their productivity and popula-
tion dynamics. Cambridge Univ. Press. Cambridge.

GOODWIN, J.G. Jr. and C.R. HUNGERFORD. 1977. Habitat
use by native Gambels and scaled quail and released
masked bobwhite quail in southern Arizona. U.S. Dep.
Agric. For. Serv. Res. Paper RM-197.

HEATWOLE, H. 1982. A review of structuring in herpeto-
logical assemblages. Pages 1-19 in Scott, N., ed. Her-
petological Communities. U.S. Dep. Inter., Fish and
Wildl. Serv. Wildl. Res. Rep. 1 3.

HIRTH, HE, R.C. PENDLETON, AC. KING, and T.R.

DOWNARD. 1969. Dispersal of snakes from a hiberna-
culum in northwestern Utah. Ecology 50(2):332-339.

JOHNSGARD, PA. 1973. Grouse and quails of North
America. Univ. of Nebraska Press. Lincoln. 553pp.

JONES, KB. 1981a. Distribution, ecology, and habitat
management of the reptiles and amphibians of the
Hualapai Aquarius planning area, Mojave and Yavapai
Counties, Arizona. U.S. Dep. Inter., Bur. Land Manage.
Tech. Note 353. Denver, CO.

. 1 98 1 b. Effects of grazing on lizard abundance and

diversity in western Arizona. Southwest Nat. 26( 2 ):
107-115.

and PC. GLINSKI. 1985. Microhabitats of lizards in

a Southwestern riparian community in North Ameri-
can Riparian Communities: Resolving Conflicting Uses.
U.S. Dep. Agric, For. Serv. Gen. Tech. Note. In press.

, L.M. KEPNER, and W.G. KEPNER. 1983. Anurans of

Vekol Valley, Arizona. Southwest Nat. 28(4):469-470.

, LP. KEPNER, and T.E. MARTIN. 1985. Species of

reptiles occupying habitat islands in western Arizona:
a deterministic assemblage. Oecologia In press.

KINDSCHY, R.R., C. SUNDSTROM, and J.D. YOAKUM.
1982. Wildlife habitats in managed rangelands — the
Great Basin of southeastern Oregon. Pronghorns. U.S.
Dep. Agric, For. Serv. Gen. Tech. Report PNW-145.

KNOPF, A.A. 1977. The Audubon Society field guide to
North American birds — western region. Alfred A.
Knopf, Inc. New York, NY.

KUCHLER, AW. 1964. The potential natural vegetation of
the conterminous United States. American Geographic
Society. Special Publ. 361. 116 pp. map.

LANDCASTER, R.K and WE. REES. 1979. Bird communi-
ties and the structure of urban habitat. Can. J. Zool.
57:2358-2368.

LOWE, C.H., ed. 1964. The vertebrates of Arizona. Univ. of
Arizona Press. Tucson. 342pp.

146

Deserts

MARES, MA. and AC. HULSE. 1977. Patterns of some ver-
tebrate communities in creosote bush deserts. Pages
209-226 in Mabry, T.J., J.H. Hunziker, and DR. DiFreo
Jr., eds. Creosote Bush. U.S./IBP Synthesis Series 6,
Dowden, Hutchison and Ross, Inc. Stroudsburg.
MARTIN, A.C., H.S. ZIM, and A.L. NELSON. 1951. American
wildlife and plants: a guide to wildlife food habits.
Dover Publ., Inc. New York, NY. 500pp.
MARTIN, T.E. 1980. Diversity and abundance of spring mi-
gratory birds using habitat islands on the Great Plains.
Condor 82:430-439.

. 1981. Limitation in small habitat islands: chance or

competition? Auk 98:715-734.
MASER, C, J.M. GEIST, DM. CONCANNON, R. ANDER-
SON, and B. LOVELL. 1979a. Wildlife habitats in man-
aged rangelands — the Great Basin of southeastern Ore-
gon— geomorphic and edaphic habitats. U.S. Dep.
Agric, For. Serv. Gen. Tech. Report PNW-99.

, J.W. THOMAS, ID. LUMAN, and R. ANDERSON.

1979b. Wildlife habitats in managed rangelands — the
Great Basin of southeastern Oregon — man-made habi-
tats. U.S. Dep. Agric, For. Serv. Gen. Tech. Report
PNW-86.
MC CLANAHAN, L. 196^. Adaptations of the spadefoot
toad, Scaphiopus conchi, to desert environments.
Comp. Bio. Chem. Physiol. 20:73-99.
MILLSAP, B.A. 1981. Distributional status of Falconiformes
in west- central Arizona: with notes on ecology, repro-
ductive success, and management. U.S. Dep. Inter.,
Bur. Land Manage. Tech. Note 355. Denver, CO.
MINCKLEY, W.L. 1973. Fishes of Arizona. Ariz. Game and

Fish Dept. Phoenix. 293pp
NAIMAN, RJ. and D.L. SOLTZ, eds. 1981. Fishes in North
American deserts. Wiley and Sons, Inc. New York, NY.
552pp.
NORRIS, K.S. 1953- The ecology of the desert iguana, Dip-
sosaurus dorsalis. Ecology 34:265-287.

. 1958. The evolution and systematica of the iguanid

genus Uma and its relation to the evolution of other
North American desert reptiles. Bull. Am. Mus. Nat.
Hist. H4(3):247-326.
OLENDORFF, R.R., AD. MILLER, and R.N. LEHIV .. 1981.
Suggested practices for raptor protection on power
lines: the state of the art in 1981. Raptor Res. Rep. 4,
Raptor Research Foundation. Dep. Vet. Biol., Univ. of
Minnesota, St. Paul.
OOSTING, H.J. 1956. The study of plant communities.

W. H. Freeman and Company. San Francisco, CA.4-i0pp.
PIANKA, E.R. 1966. Convexity, desert lizards, and spatial
heterogeneity. Ecology -t7( 6): 1055- 1059.

. 1970. Comparative autecology of the lizard Cnem-

idophorus tigris in different parts of its geographic
range. Ecology 51:703-720.
PORTER, KR. 1972. Herpetology. W. B. Saunders Co. Phil-
adelphia. PA. 281pp.
ROTTENBERRY, J.T. and J.A. WIENS. 1980. Habitat struc-
ture, patchiness, and avian communities in North
American steppe vegetation: a multivariate analysis.
Ecology 61:1228-1250.
SCHMIDT, J.L. and D.L. GILBERT, eds. 1978. Big game of
North America: ecology and management. Stackpole
Books. Harrisburg, PA. 512pp.
SERVENTY, D.L. 1971. Biology of desert birds. Pages 287-
339 in Farmer, D.S., and JR. King, eds. Avian Biology:
Vol. I Academic Press. New York, NY.
SHORT, ILL 1983. Wildlife guilds in Arizona desert habi-
tats. BLM Tech Note 302. Denver, CO. 258pp.

SHREVE, F. 1942. The desert vegetation of North America.
Bot. Rev. 8:195-246.

SIMON, C.A. and G.A. MIDDENDORF. 1976. Resource par-
titioning by an iguanid lizard: temporal and microhabi-
tat aspects. Ecology 57<6):1 317-1 320.

SKOVLIN, J.M. 1982. Habitat requirements and evalua-
tions. Pages 369-413 in Thomas, J.W. and D.E. Toweill,
eds. Elk of North America: Ecology and Management.
Stackpole Books, Harrisburg, PA.

STEBBINS, R.C. 1966. A field guide to western reptiles
and amphibians. Houghton-Mifflin Co. Boston, MA.
279pp.

THOMAS, J.W., C. MASER, andJ.E. RODIEK. 1979. Wildlife
habitats in managed rangelands — the Great Basin of
southeastern Oregon — edges. U.S. Dep. Agric, For.
Serv. Gen. Tech. Rep. PNW-85. 17pp.

THORNE, R.F., B.A. PRIGGE, and J. HENDR1CKSON. 1981. A
flora of the higher ranges and the Kelso Dunes of the
eastern Mohave Desert in California. Aliso 10:F- 1-186.

TOMOFF, C.S. 1974. Avian species diversity in desert
scrub. Ecology 55(2):396-403.

TURNER, R.M. 1982a. Great Basin desert scrub. Pages 145-
155 in Brown, D.E., ed. Desert Plants: Biotic Commu-
nities of the American Southwest — United States and
Mexico. Univ. of Arizona Press. Tucson.

. 1982b. Mohave desert scrub. Pages 157-168 in

Brown, D.E., ed. Desert Plants: Biotic Communities of
the American Southwest — United States and Mexico.
Univ. of Arizona Press. Tucson.

and D.E. BROWN. 1982. Sonoran desert scrub, in

Brown, D.E., ed. Desert Plants: Biotic Communities of
the American Southwest — United States and Mexico.
Univ. of Arizona Press. Tucson. 342pp.

VAN DEVENDER, T.R and W.G. SPAULDING. 1979. De-
velopment of vegetation and climate in the southwest-
ern United States. Science 204:701-710.

VITT, L.J., R.C. VAN LOBEN SELS, and R.D. OHMART.

1981. Ecological relationships among arboreal desert
lizards. Ecology 62(2): 398-4 10.

VOGL, RJ. 1977. Fire: A destructive menace or a natural
process? Pages 261-289 in Cairns, J. Jr., KL. Dickson,
and E.E. Herricks, eds. Recovery and Restoration of
Damaged Ecosystems. Univ. Press of Virginia. Char-
lottesville.

WEBB, R.H and H.G. WILSHIRE, eds. 1983. Environmental
effects of off-road vehicles: impacts and management
in arid regions. Springer-Verlag. New York, NY. 534pp.

WILSON, L.O., J. BLAISDELL, G. WELSH, R. WEAVER, R.
BR1GHAM, W. KELLY, J. YOAKUM, M. HINKES, J.
TURNER, and J. DE FORGE. 1980. Desert bighorn hab-
itat requirements and management recommendations.
Pages 1-7 in Des. Bighorn Coun. Trans. 24.

WRIGHT, HA. and AW. BAILEY. 1982. Fire ecology. Wiley -
Interscience, New York, NY. 501pp.

YOAKUM, J. 1980. Habitat management guides for the
American pronghorn antelope. U.S. Dep. Inter., Bur.
Land Manage. Tech. Note 347. Denver, CO.

YORKJ.C. and W.A. DICK-PEDDIE. 1969. Vegetation
changes in southern New Mexico during the past
hundred years. Pages 157-166 in McGinnies, W.G. and
B.J. Goldman, eds. Arid Lands in Perspective. Univ. of
Arizona Press. Tucson.
ZWEIFEL, R.G. and C.H. LOWE. 1966. The ecology of a
population of Xantusia vigilis, the desert night lizard.
Am. Mus. Novit 2247;1 S~

Deserts

147

8

TUNDRA

Peter C. Lent

Bureau of land Management
920 Valley Road
Reno, NV 89512

"When we reach the Arctic regions, or snow-capped
summits, the struggle for life is almost exclusively
with the elements."

— Charles Darwin, from On the Origin of
Species

Editor's Note: Animals of tundra habitat must
struggle for life. Because of the harsh working con-
ditions and inaccessibility of many tundra areas,
few biologists have conducted systematic habitat
inventories or surveys of general vertebrate habitat
relationships in these severe environments. Most
of the work in such areas has concentrated on a
few large, economically important wildlife species.
Nevertheless, the conditions are so different from
those in more temperate areas that the subject de-
serves a separate chapter. This chapter describes the
state-of-the-art for both arctic and alpine tundra;
the biologist working in such areas will find much
room for innovation and development of new
systems.

INTRODUCTION

The origin of the English word "tundra" is said
to come from the Finnish "tunturi" then from the
Russian "tundra," meaning a marshy, treeless plain.
In a general sense the term refers to any cold cli-
mate landscape having vegetation without trees.
Such landscapes occur predominantly in the north-
ern hemisphere north of the tree line (arctic tundra)
and in mountainous areas of both hemispheres above
the tree line (alpine tundra). The vegetation commu-
nities are characterized by short-stemmed perennial
herbaceous plants, stunted or prostrate shrubs, li-
chens, and mosses. Similar vegetation, where trees
are absent due to low-growing season temperatures,
also occurs in mid- to high-latitude maritime envi-
ronments such as the Aleutian Islands and on certain
islands in the southern hemisphere, such as South
Georgia and the Macquaries (Jenkins and Ashton
1970).

Tundra areas constitute slightly over 8% of the
total U.S. land area and about 5% of the earth's land
surface. Of the U.S. total, about 74 million ha (183
million a.) are in Alaska and 3 million ha (7.5 million
a. ) of alpine tundra are in the contiguous 48 states
(Figure la). Of the Alaskan tundra area, Brown et al.
( 1978) estimated that 52% is arctic and 48% is al-
pine tundra (Figure lb).

The manager or biologist new to arctic ecosys-
tems will find a detailed and comprehensive body of
literature available. The purpose of this chapter is
to help the manager get started but is by no means
exhaustive. In contrast, the managers concerned
with alpine areas will find a more scattered body of
literature and will need to make a greater effort to
seek out material specific to their area. Two general
treatments of North American alpine areas are worth
noting: Zwinger and Willard (1972) and Price (1981).
A glossary of terms broadly relevant to arctic and
alpine environments has been prepared by Gabriel
and Talbot ( 1984 ). The biologist or manager working
with alpine areas will find some useful material in

Tundra

149

1 . Sierra Nevada

2. Southern Cascades, Siskiyou and Klamath Mountains

3. Olympic Mountains

4. Northern Cascade Mountains

5. Northern Rocky Mountains

6. Great Basin Ranges

a) Blue and Wallowa Mountains: Oregon

b) Ruby, Wheeler, Charleston Mountains: Nevada

c) Uinta and Wasatch Mountains: Utah

d) San Francisco Peak and White Mountains: Arizona

e) Mounts San Jacinto and San Gorgonio: California

f) White, Inyo and Panamint Mountains: California

7. Southern Rocky Mountains

Figure la. Alpine tundra areas in the contiguous 48 states (adapted from Zwinger and Willard 1972).

150

Tundra

Moist tundra
Wet tundra

Alpine tundra,
ice and snow

Figure lb. Arctic and alpine tundra areas in Alaska.

the arctic literature, but must use caution in its ap-
plication; there are some fundamental differences
between arctic and alpine ecologies.

All tundra areas are characterized by short
growing seasons, usually fewer than 100 days, and
low summer temperatures. In extreme situations the
daily temperature may average only 2° to 5° C (36°
to 41° F). Although all share this characteristic of
"cold-dominated," the vegetation types encompassed
by tundra are exceptionally diverse. At one extreme
is the high arctic polar desert, where vegetative
cover may be less than 10%, composed primarily of
cushion plant forms and mosses with a few other
annuals scattered sporadically. Polar desert and semi-
desert are rare within the U.S. They are best repre-
sented in the arctic-alpine environment of the
Brooks Range in northern Alaska. At the other end of
the spectrum, over most of North America, the tun-
dra grades into forest in a forest-tundra ecotone.
Here vegetative biomass can be relatively high with
several well-developed canopy layers.

Discussion in this chapter will encompass these
transition zones because of the unique qualities they
have in common with tundra, such as prevalence
and importance of lichens and potential presence of
permafrost, and because important migratory species,
such as elk (Cervus elaphus) and caribou (Rangifer
tarandus), depend on a single system composed
of both tundra and forest-tundra zones.

Some writers have contended that it is better to
reserve the term "tundra" to refer to areas with
permafrost, either continuous or discontinuous. Bill-
ings ( 1973), for example, preferred the use of the
terms "meadow" and "fell-field" for the mesic and
xeric alpine equivalents of tundra, respectively. As
he noted, important differences exist, such as the
more intense solar radiation in alpine environments
coupled with extreme diurnal temperature fluctua-
tions, as opposed to the continuous summer daylight
in the high latitude tundra environment. But Billings'
own work revealed that 50% of the vascular plant
species in the alpine zone of the Beartooth Plateau

Tundra

151

of Montana were species also present in arctic tun-
dra. A similar large overlap of species does not occur
between arctic and alpine faunas, however
(Hoffmann 1974, 1984).

Aside from the similarities in low growth forms,
which suggest that similar vegetation sampling tech-
niques are applicable in alpine and arctic situations,
there are some basic similarities arising from the
cold-dominated environment of both, such as low
primary productivity, low decomposition rates, low
stress tolerance, and high susceptibility to physical
disturbance — even where permafrost is not present.
The complex spatial heterogeneity of arctic and al-
pine vegetation, a "patchiness" that is based on re-
sponses to microrelief features, also dictates that
similar problems are to be encountered in inventory
and monitoring of arctic and alpine habitats.

The presence of permafrost is certainly charac-
teristic of most arctic tundra areas. Discontinuous
permafrost is also found in about 100,000 knT
(40,000 mi. ) of alpine areas in the western U.S.
exclusive of Alaska (Pewe 1983). However, perma-
frost is not universally present in tundra nor does
the presence of permafrost, if it is at sufficient depth,
necessarily preclude tree growth. Furthermore, even
in the absence of permafrost, per se, cryopedogenic

forces (frost action) play a significant ecological role
in arctic and alpine habitats. Thus the position taken
by Bliss ( 1981 ) and the one taken in this chapter is
that the presence, absence, and the characteristics of
the permafrost layer are among the major influences
on the nature of a given tundra area and are major
considerations in inventory and management. They
are not, however, decisive in determining whether
an area is considered to be tundra.

A current tendency exists to shift tundra to
lower levels in hierarchical vegetation classification
systems. For example, Viereck and Dyrness (1980)
distinguished between tundra and grass/herbaceous
formations in the first dichotomy (Level I) of their
hierarchical system. However, in a later version of
their classification system (Viereck et al. 1982), all
tundra types were reduced to subcategories under
two of their Level I classes (herbaceous and shrub).
They classified communities of the Aleutian Islands
and other maritime environments that are dominated
by Calamagrostis canadensis and Elymus arenarius
as grasslands and not tundra. They also referred to
shrubland's several types that have been included
with tundra or shrub-tundra by other workers. Dris-
coll et al. (1983) also included tundra as a subdivi-
sion in their grassland/herbaceous class.

MMfaT 'l^ftg

V# ';r;v^;--

*..*v>*

»43ur* l , Eft * s $fa i*<m *T m . . - . ^ ■

/*»■

^^ -'<•*<

• •*

< ■

rW

Caribou on tundra riparian area in July.

152

Tundra

GENERAL HABITAT CLASSIFICATION
SYSTEMS

Various wildlife habitat classification systems
have been employed in both arctic and alpine areas.
There are no generally or widely used systems and
none of them has been systematically tested as a
general system. There is, however, some general cor-
respondence among some of these, particularly for
terrestrial birds and small mammals (Table 1). Most
of these habitat classes represent various subdivi-
sions of the soil moisture gradient from flooded and
saturated through desiccated and rocky (Pruitt 1966;
Hoffmann 1974). This reflects the dominant role of
soil moisture as a niche variable both directly and
indirectly through its interrelations with the perma-
frost regime and vegetation. Winter habitats of small
mammals also reflect similar snow-cover gradients
(Pruitt 1966). These general habitat classes and
other habitat features and their relationships to one
another are illustrated in Figures 2 and 3-

Moitoret et al. (1985) performed the most de-
tailed analysis of arctic tundra habitats to date. Using
the clustering and ordination analysis program
TWINSPAN, they found flooded and riparian to be
the most distinct of their seven coastal tundra terres-
trial bird habitats. Drier types, other than riparian,
had lower and less diverse bird use. Their analysis
suggested that habitat factors relating to microrelief
and interspersion characteristics may be important in
defining habitats. Thus, habitat classes so defined do
not necessarily show one-to-one correspondence
with the broad vegetation-based classes ascertained
from satellite imagery such as that of Landsat.

Several more specialized tundra habitat classifi-
cations have been employed. One relating specifi-
cally to wetlands is described in the discussion
regarding migratory waterfowl. A system used for
impact analysis in arctic and subarctic Alaska is de-
scribed in the section treating inventory and moni-
toring for impact assessment.

Table 1. Habitat classes used in some studies of arctic and alpine terrestrial birds and mammals.

Locale

Habitat Classes

Reference

Birds and small

Moist

Sedge tussock

Willow sedge

Dry

Krumm-

Fellfield

Cliff, talus

Hoffmann

mammals,

meadow

meadow1

holz2

(1974)

Beartooth

Plateau, MT

Breeding habi-

Moist, wet

Heath tussock

Riparian

Dry

Combined

Cliffs, talus

Williamson

tats of birds,

meadow

meadow

with dry

(1961)

Cape Thomp-

Fellfield

meadow

son, Alaska

Bird habitats,

Sedgegrass

Tussock heath

Tall shrub

Dwarf

Dry tundra

Kessel and

Upper Sheen-

marsh

shrub2

Schaller

jek R. (Subarc-

(1960)

tic-alpine)

Alaska

Bird habitats,

5 hydric

Tussock

Riparian

Moitoret et

coastal Arctic

types3

al. 1985

National Wild-

life Refuge,

Alaska

Selected

Long-tailed,

Lapland, long-

Arctic warblers,

Golden

Willow

Horned

Marmots,

Characteristic

parasitic

spur, Calcarius

Dendroica sp.

plover, Plu-

ptarmigan

lark, Ere-

Marmota

Species

jaegers,

lapponicus

vialis dom-

mophila

sp.

Sterconarius

Red-backed

Water shrew,

inica

Chip-

alpestris

sp.

vole, Clethrion-

Sorex palustris

munks,

Mountain

omys rutilus

Pocket

Eutamias

Dall's

goat

Tundra vole,

Moose, Alces

gophers,

sp.

sheep

Microtus

alces

Thomomys

Golden

Oeconomus

sp.

Dall's
sheep

eagle,
Aquila
chrysaetos

Hoffmann (1974) also distinguishes rock polygons, these being dry meadow with rock stripes or patterns. Characteristic species are

pikas, Ochotona sp.

2Krummholz is a coniferous ecotone below the alpine tundra. These two are only crudely structural counterparts.
3Five hydric types identified: Flooded. Wet sedge, Moist sedge, Mosaic, and Moist-sedge shrub.

Tundra

153

MAJOR WILDLIFE GROUPS

It is useful, particularly for wildlife species in
arctic tundra habitats, to consider three categories of
species: ( 1 ) resident species that remain active year-
round, (2) resident species hibernating in winter,
and (3) migratory species present for only a portion
of the year. Tundra areas tend to have a larger pro-
portion of migratory and otherwise highly mobile
species than most other broad habitat classes.

Resident arctic mammal species include some
medium-sized hibernators, such as the arctic ground
squirrel {Spermophilns parry ii) and hoary marmot
(Marmota caligata), or large ones, such as grizzly
bears (Ursus arctos), but include a larger variety of
species that remain active year-round.

The smaller species, rodents and shrews primar-
ily, tend to be "chionophiles" (Pruitt 1978), that is,
they require snow cover to ameliorate an otherwise
unsurvivable winter environment. Others are "chion-
ophobes," such as the musk-ox (Ovibos moschatus),

requiring habitats of restricted snow cover for win-
ter survival. The importance of snow in the tundra
environment is treated in a separate section of this
chapter.

The number of resident arctic bird species is
extremely small, but may include ptarmigan (Lago-
pus lagopus and L mutus), ravens (Corvus corax),
snowy owls (Nyctea scandiaca), and gyrfalcons
(Falco rusticolus). The great majority of the 97 or
so bird species using the northern Alaska tundra are
migratory (Pitelka 1979). Their dependency on tun-
dra habitats, however critical, is a brief one and
timed precisely according to the adaptations of each
species. Thus there is a complex and compressed
pattern of arrivals and departures that makes the
correct timing of inventories critical to their success.
In northern Alaska the peak of arrival dates is in late
May; some early breeding species may leave as early
as late July.

Waterfowl and shorebirds are almost entirely
migratory except for a few species that may associ-
ate with open water marine environments in winter

GLACIER

BEACH

DE-ROCK,
TALUS

FORB MEADOW

SANDY BLUFF

POLYGON

THAW LAKE
r MEADOW

WOODED
TUNDRA 4

> A>* 4

Figure 2. Schematic representation of arctic and arctic-alpine tundra habitat types and features (after Hoff-
mann 1984).

154

Tundra

Musk-ox winter tundra habitat with low snow-cover.

in some arctic areas. Most waterbird species are
clearly tied to wet tundra habitats. However, there
are some shorebird species clearly associated with
dry, upland, or even arctic-alpine habitats (Kessel

and Schaller I960; King 1979). The arctic-alpine tun-
dra supports a surprisingly diverse array of summer
breeding birds, estimated by Pitelka (1979) to be 50
species, of which nearly half are passerines.

Because all migratory species cannot be sampled
at an optimal time in a single inventory, it is neces-
sary to have a clear statement of objectives and clear
familiarity with the regional literature to strive for
the optimal timing for specific needs or specific spe-
cies. Moulting concentrations are common after
breeding in several waterfowl species, and moulting
areas may be used only briefly, yet are of critical
importance to entire populations (Pitelka 1979; NPR-
A Task Force 1978; Derksen et al. 1982).

Systematic surveys of arctic tundra waterfowl
with small aircraft began as early as 1948 (Smith and
Allen 1948) and have been greatly improved since
then (King 1970). Bergman et al. (1977) devised
a system for classifying those tundra wetlands that
support large breeding bird populations. The system
uses data on size of wetland, vegetation, water depth,
and water chemistry to identify eight key habitat

Alpine Stand Types

IGNEOUS ROCK AREA

SEDIMENTARY ROCK AREA

SOLIF

Figure 3. Schematic representation of alpine habitat types and features (after Hoffmann 1984).

Tundra

155

classes of importance to waterbirds (Table 2). Pre-
liminary studies suggest that identification of the
class of wetland associated with a given tundra area
will be an acceptable predictor of the waterbird
species, using such an area without a direct survey
of the bird populations themselves (U.S. Fish and
Wildlife Service 1982).

Some species may have both migratory and resi-
dent populations, even within the same general area,
i.e., the caribou. In other situations, a population
may shift from resident to migratory status depend-
ing on certain environmental conditions. In the arc-
tic, the gyrfalcon is one such species in which prey
availability can shift behavior.

As with the arctic, alpine tundra areas show low
species diversity. A major difference is that, espe-
cially with resident species, alpine populations often
exist in small islands of habitat. Habitats formerly
occupied by extirpated populations will not be
quickly recolonized and genetic conservation is
more likely to be a management concern. For exam-
ple, an endemic species of butterfly (Boloria acrone-
mid) has been found only in three alpine meadows
in the San Juan Mountains of Colorado (Gall 1984).

Thirteen species of herbivorous mammals in-
habit the alpine area of the Beartooth Mountains,
Montana, and only six species of herbivorous or

Table 2. Wetland classification system of Bergman et al. (1977) with two additional types from Derksen
(1979).

Wetland Class

Description

Class 1: Flooded tundra

Shallow waters formed during spring thaw when melt water
overflows stream basins or is trapped in vegetated tundra
depressions. These temporary wetlands form in shallow
basins, polygons, or broad meadows.

Class II: Shallow-Carex

Shallow ponds with a gently sloping shore zone surrounded
by and usually containing emergent Carex aquatilis in a
central open water zone. These ponds, formed in low
centers of polygonal ground, often produce a mosaic pattern
of ridges and ponds.

Class III: Shallow-AYcfopfr/'/a

Ponds or pools in beaded streams containing Arctophila
fulva in the central zone and shoreward stands of A. fulva or
Carex aquatilis. Shores are more abrupt than those of Class
II ponds. Maximum water depths range from 20 to 50 cm.

Class IV: Deep- Arctophil 'a

Wetlands of either large pond or lake size that lack
emergents in the central zone and contain stands of
Arctophila fulva near the shore.

Class V: Deep-open

Large, deep lakes that have abrupt shores, sublittoral
shelves, and a deep central zone.

Class VI: Basin-complex

Large, partially drained basins that may contain nearly
continuous water in spring due to flooding of the bottom by
melt water. A composite of several other classes.

Class VII: Beaded stream

Small, often intermittent, streams consisting of a series of
channels formed in ice-wedges and linked to pools that
develop at ice-wedge intersections.

Class VIII: Coastal wetlands

Aquatic habitats that occupy low areas within a zone directly
influenced by sea water.

Class IX: Upland tundra

Ponds characterized by depressions in upland tussock
vegetation {Eriophorum vaginatum) that fill with melt water
and evolve from ephemeral pools to permanent ponds.
When mature, these ponds are typically less than 10 by 3 m
with a maximum depth of 1 m.

Class X: Ice- wedge pools

Small (10 m wide) pools formed in ice-wedge troughs in both
high and low center polygon areas. Older, enlarged pools
resemble Class II ponds and are more diverse than smaller,
deep, acidic pools.

156

Tundra

Uncompahgre frittilary butterfly.

Pingo and associated tundra ponds.

insectivorous birds regularly breed above timberline.
Eight carnivorous mammal species and seven to
eight raptor species spend at least part of the year in
this alpine environment (Hoffmann 1974; Price 1981).
Carnivorous mammals and predatory hawks and owls
seem to constitute a relatively greater proportion of
the overall species array compared with the propor-
tion in temperate or tropical biomes.

In alpine environments the large mammals, such
as grizzly bears, bighorn sheep (Ovis canadensis),
elk, Dall's sheep (Ov is dalli), and mountain goat
(Oreamnos americanus), generally make altitudinal
movements, descending to more wooded habitats
in winter. Dall's sheep and mountain goats may on
occasion use tundra ranges in winter.

Some bird species also make seasonal altitudinal
movements. In the western contiguous states, white-
tailed ptarmigan (Lagopus leucurus) use a broad
array of alpine habitats, moving in winter to lower
elevation shrub communities. Farther north, where
there are three sympatric ptarmigan species, the
white-tailed occupies a narrower niche.

Altitudinal movements may even be on a daily
basis (Hoffmann 1974). Many raptors nest at lower
elevations, but take advantage of upwelling wind
currents to move up to open alpine terrain for forag-
ing during the day.

SPECIAL HABITAT FEATURES

In areas of continuous or near-continuous
permafrost, suitable habitat for denning or burrowing
species may be very limited. The depth of the active
layer (seasonally thawed soil zone) can be an impor-
tant variable influencing wildlife habitat values. Den-
ning or burrowing species depend upon well-drained
or exposed substrates, such as pingos, where the

permafrost zone has receded or is ice-free. Pingos
are conical-shaped mounds pushed up out of the flat
tundra by hydrostatic pressure. These features reach
65 m (213 ft) in height and, because of drainage
and aspect, may provide relatively high substrate
temperatures as well as lookouts (Hoffmann 1974).
Like pingos, coastal bluffs, cut banks, gravel bars,
talus slopes, and sand dunes can be important habitat
features. Their number and extent may actually be
the limiting factor for certain populations. Sandy
banks of only 1 m (1 yd) high may be used by arctic
foxes (Alopex lagopus) for denning (Bee and Hall
1956).

Because of the generally flat, treeless character
of the arctic tundra, river and coastal bluffs are often
the principal sites for nesting raptors outside of the
mountainous zone. Specialized systems have been
developed for classifying these raptor habitats (Rit-
chie 1979) which also include escarpments, rock
outcrops, and talus slopes as one moves into the arc-
tic-alpine environment. Any habitat surveys or other
intensive activities with aircraft should be conducted
so as to have minimal disturbance on nesting raptors,
peregrine falcons (Falco peregrinus) in particular
(Fyfe and Olendorff 1976). Because such sites are fre-
quently also preferred sites for extraction of con-
struction materials (gravel or sand) or sometimes
even as camp or construction sites, it is important to
inventory these features and assess their value as
habitat.

The presence of deep snowdrifts is an important
habitat feature for denning wolverines {Gulo gulo),
polar bears (Ursus maritimus), and brown bears (U.
arctos). The drift itself may form the den or it may
form an additional insulative cover over shallow rock
or earth dens. On the other hand, Reynolds et al.
1976 suggested that brown bears in the arctic do
not den until about the top 10 cm (4 in.) of soil is
frozen because otherwise the coarse-textured soils
would collapse on the excavation.

Tundra

157

Some habitat features may be transient. For ex-
ample, certain arctic river deltas may provide the
only open water for early arriving waterfowl and
shorebirds at a time when tundra lakes are still fro-
zen. These critical areas of open water are easily
discernible from aerial reconnaissance or from Land-
sat imagery (NPR-A Task Force 1978; U.S. Fish and
Wildlife Service 1982; Lent 1985). Aufeis or naled,
the sheets of ice remaining in river and braided
streambeds late into summer, are another special
habitat feature that provides cool, insect relief areas
for caribou, moose (Alces alces), and other large
mammals.

Thus, the special importance of riparian habitats
and riparian-associated habitat features within arctic
tundra landscapes cannot be stressed too strongly.
Not only do they provide microrelief and thawed
soils for burrowers and bluffs and banks for nesting
sites, they also provide narrow zones of brushy habi-
tats that are critical to the requirements of other
species (NPR-A Task Force 1979; Pamplin 1979; Moi-
toret et al. 1985).

INVENTORY AND MONITORING SYSTEMS

Range Surveys

Most of the early efforts in the 1920s and 1930s
to map, describe, and inventory arctic and subarctic
tundra ranges were conducted because of their eco-
nomic importance for domestic reindeer in the
U.S.S.R. and in Alaska (Palmer 1926). With the col-
lapse of the reindeer industry in Alaska and the dis-
ruptions resulting from World War II in the Soviet
Union, there was a hiatus of scientific effort. In the
late 1950s, systematic attention was first turned to
the study of caribou and their habitats in Alaska and
Canada. This was made possible by improvements
in the performance and availability of small aircraft,
providing biologists with a mobility and a perspec-
tive not previously possible. This work tended to
focus on taiga ranges used by wintering caribou.

In the 1950s, scientific studies of Scandinavian
domestic reindeer got underway and Soviet work
was intensified. Banfield (1954) and Tener (1965)
also began their classic studies in the Canadian Arc-
tic. In Alaska, two major reviews of literature and
field work relating to caribou range were done by
Courtright (1959) and Pegau (1968). Pegau put
emphasis on specific comparisons of vegetation in-
ventory techniques suitable for tundra ranges.

Among the pioneer workers in Scandinavia,
Skuncke (1969) devised a detailed system for aerial
surveys of reindeer pastures to estimate pasture car-
rying capacities and trends. Much of the system is
specialized for the Scandinavian situation, however,
and also requires very experienced observers.

Beaded streams furnish good nesting habitat for white-
fronted geese and ducks.

Special inventory techniques are required for
tundra areas where lichens are a significant compo-
nent or likely to be important forage plants. Pegau
( 1970) used a 919-cnT (1-ft") frame and visually esti-
mated lichen cover based on random throws. Using
this simple technique, he was able to assess the dam-
age to lichens caused by holding or herding reindeer
on lichen ranges during dry periods. Pegau (1968)
also measured growth rates for lichens in Alaska.
Although the randomness of random throws has
been legitimately criticized, the technique has
proven useful (Greig-Smith 1983).

Changes over time in arctic tundra vegetation
have been successfully measured in other studies as
well. In a study of the effects of grazing by confined
musk-oxen on tussock and mat cushion tundra, Mc-
Kendrick ( 1981 ) used the walking point method
of Owensby (1973) to measure basal cover and spe-
cies composition. With these techniques, he was able
to demonstrate significant reduction in lichen and
shrub-standing crops caused by musk-oxen. Thus the
measured difference with regard to musk-oxen influ-
ence on shrubs was significantly different from that
reported from numerous reindeer studies. Intense
grazing pressure by reindeer tends to promote inva-
sion by shrub species. Racine (1979) compared soil
condition, permafrost depth, and vegetation on plots
before and after tundra fires on the Seward Penin-
sula. He used simple square meter (square yard)
plots, recording stem density and percentage cover
to measure significant differences before and after
the fires.

Thilenius (1979) presented the best recent dis-
cussion of the problems, concepts, and inexact state-
of-the-art for assessing and monitoring alpine range
communities. He noted that classifying range condi-
tion is a great problem in alpine range management.
Range condition classes often seem to be confused
with site climax classes. For example, cushion plant
communities dominated by Geum rossii have been

158

Tundra

treated by range managers as poor condition com-
munities when in fact such forb-dominated commu-
nities, common on xeric sites, may represent
excellent range condition. He recommended the
system of Lewis (1970) who, he stated, properly rec-
ognized the unique microclimate and soil features
of alpine ranges. Thilenius (1979) also emphasized
that alpine range monitoring must be based on an-
nual sampling because of the high degree of year-to-
year variability.

As with any class of vegetation or habitat, de-
scription and mapping of soils types is an important
part of habitat analysis. The characterization, inven-
tory, and mapping of tundra soils has become partic-
ularly important in the last two decades because
arctic development has required an increased com-
mitment to land planning and engineering skills
(Everett 1982).

In arctic soils, the presence of permafrost, its
depth and water content, as well as the nature of the
overlying vegetation are important considerations.

When permafrost is present, it is frequently associ-
ated with areas of relatively little relief. This, plus
the generally fine texture of soils and spongy moss
cover, limits water movement and results in satu-
rated soils (Figure 4). Gley (wet) soils are indicated
by cottonsedge (Eriophorum sp.) tussock and dwarf
heath communities. Bog soils are characterized as
cottongrass sedges (Carex aquatilis) and Dupontia
Plant communities dominated by lichens, mosses,
and dwarf heaths with scattered herbs prevail over
shallow dry or rocky soils where the permafrost
is also dry and generally 1 m (1 yd) or more deep.
Similar vegetation also occurs on dry raised patches
of peaty material (Linell and Tedrow 1981).

Remote Sensing Applications

Because of the vast areas involved in arctic tun-
dra management, the lack of access, except by ex-
pensive aircraft, and the short field season, there is a
strong rationale for the use of remote sensing sys-
tems for mapping and inventory of vegetation and

ACTIVE
LAYER

WELL-DRAINED SITE

POORLY DRAINED SITE

BOG

Figure 4. Idealized cross-section of tundra landscape from higher well-drained ground on left to peaty
ground on right (after Linell and Tedrow 1981).

Tundra

159

wildlife habitats. Several efforts have been directed
in recent years toward the use of satellite sensors for
that purpose, particularly Landsat. These trial efforts
are described by Anderson and Belon (1973), Ander-
son et al. ( 1974), LaPerriere ( 1978), George et al.
(1977), and Nodler et al. (1978). Anderson and Be-
lon (1973), Lent and LaPerriere (1974), and Hall
et al. ( 1980) described the use of Landsat imagery to
delineate and monitor change in areas that have
been subject to recent wildfires. This then provides a
useful tool to monitor changes in wildlife habitats
due to the wildfires and subsequent recovery of tun-
dra vegetation. Lent (1980) also used Landsat im-
agery to monitor gross changes in extent of snow
cover. More recently, Acevedo et al. (1982) used
Landsat for environmental mapping and related base-
line studies on a portion of the coastal tundra of
the Arctic National Wildlife Refuge. To a certain ex-
tent, wildlife biologists have used wildlife habitat
inventory and mapping efforts, and use classification
schemes that take advantage of the broad land cover
types discriminated from Landsat imagery (U.S. Fish
and Wildlife Service 1982, 1984). As noted earlier,
Moitoret et al. ( 1985) reported on some of the po-
tential problems with such approaches.

Satellite imagery from Landsat and weather satel-
lites has also been used to assess the snow and ice
conditions encountered by nesting geese in the Ca-
nadian Arctic. By assessing the year-to-year timing of
snow and ice melt on a broad scale, this nesting
habitat information has been used to successfully
predict productivity of species such as brant {Branta
bernicla), snow geese {Chen caerulescens), arctic-
breeding Canada geese (B. canadensis), and greater
white-fronted geese (Anser albifrons) (Reeves et
al. 1976).

A major limitation to the use of satellite imagery
over much of the arctic is the high frequency of
overcast days. The large number of overcast days is
partially compensated for by the polar convergence
of Landsat orbits, resulting in bursts of imagery on
successive days at high latitudes. Because the phen-
ology changes so rapidly and the windows of good
weather are frequently so short, it is difficult to ob-
tain comparable imagery in successive years to docu-
ment long-term trends.

The use of satellite imagery for mapping and
monitoring alpine habitats is at present much more
difficult because of the problems resulting from top-
ographic complexity, including the effects of shad-
ows on surface albedo. However, Craighead et al.
(1982) classified and mapped alpine and subalpine
vegetation in western Montana, using Landsat im-
agery; a computer-assisted, multi-spectral analysis
system; and multiple-level, ground-truthing to inven-
tory grizzly bear habitat.

Simple remote sensing techniques involving
small aircraft are also useful. For example, good suc-
cess in mapping tundra-vegetation types has been
reported for false-color, infrared photography and
simple hand-held, 35 mm cameras in fixed- wing air-
craft (Holt 1980). Hesjedal and Larsson (1975) re-
ported color-infrared film to be useful for the
detection of eroded and overgrazed areas in Fenno-
scandian alpine regions, but recommended use of
stereo-panchromatic photography as most cost-effec-
tive for generalized alpine surveys.

Use of stereoscopic pairs of medium- to large-
scale imagery is helpful for raptor nesting habitat
surveys. Some habitat features such as pingos are eas-
ily mapped from small-scale aerial photographs (Sims
1983).

Sims (1983) used large-scale, color-infrared, 70-
mm aerial photography to map lichen types which
revealed reindeer winter range types and carrying
capacities on the Tuktoyaktuk Peninsula, Canada.
Microdensitometric readings on the original film
were diagnostic for areas with over 5% lichen cover.
About 81% of sample sites were correctly classified.

Inventory and Monitoring for Impact
Assessment

The manager of wildlife and wildlife habitat in
tundra systems is increasingly involved in issues and
decisions involving human uses of these lands. In the
arctic, exploration, development, transportation,
and production of oil, gas, and mineral ores are
prime examples of these land-use problems. In alpine
environments, issues are frequently related to inten-
sive recreational use, livestock use, or mineral devel-
opment. Thus this chapter places considerable
emphasis on inventory and monitoring techniques
that have had demonstrable applications in such
management situations.

Surface disturbance and alteration of permafrost
regime have clearly had undesirable aesthetic effects
in arctic areas, as well as caused localized erosion
problems. The long-term ecological consequences of
such changes have been less clear-cut.

Nevertheless, for purposes of planning and in-
ventory before proposed developments in the arctic,
mapping of permafrost temperature, depth, moisture
content, and protective vegetative cover have re-
ceived paramount attention (Linell and Tedrow
1981). Similarly, monitoring efforts have often fo-
cused on changes in permafrost status and the over-
lying insulating vegetation (Brown et al. 1978). The
fragility or sensitivity to damaged tundra areas is
directly proportional to the ice content of the
permafrost layer and to the mean annual ground
temperature (Ives 1970), as well as to the insulative
qualities of the surface cover. The ability to predict

160

Tundra

such sensitivity is now good. A monitoring program
may track the success of stipulations relating to sur-
face disturbance or recovery programs necessitated
by unavoidable disturbance (Hok 1969, 1971; van
Cleve 1977; Brown and Berg 1980).

During the major monitoring effort associated
with construction of the Trans-Alaska Pipeline Sys-
tem, terrain along the corridor was classified into 1 2
wildlife habitat types. Of these, four (subalpine, al-
pine tundra, tussock tundra, and wet-meadow tun-
dra) are clearly associated with or are subcategories
of tundra. Three others (wetlands, riparian willow,
and unvegetated floodplains) frequently occur as
islands or narrow strips within tundra habitats. A
panel of biologists gave independent ratings for the
relative quality of these wildlife habitats, weighing
the importance of 24 wildlife groups (mostly taxa at
the family level) equally. Based on this relatively
crude evaluation process, wet-meadow tundra and
wetlands received the highest quality ratings. Ripar-
ian willow and tussock tundra were rated high to
moderate. The alpine and subalpine types were rated
low-moderate. A comparison of preconstruction and
post-construction imagery (aerial photography) re-
vealed that 12,709 ha (31,403 a.) of wildlife habitat
were altered or destroyed during this major project.
Nearly half of these lands were in high quality wet-
lands and wet-meadow tundra habitats. However, the
greatest deviation from impacts predicted in the
environmental impact statement (EIS) was in the
land altered by construction material sites. The EIS
predicted that the disturbance for material (primarily
gravel) sites would be about 2,400 ha (6,000 a.)
compared with the actual figure of nearly 4,800 ha
( 1 2,000 a. ), as determined in this monitoring study
(Pamplin 1979).

Because of the rapid succession of seasonal
changes in tundra areas and the importance of these
events to phenology, animal abundance, and sensitiv-
ity, it is important that the biologist prepare prelimi-
nary charts of these events. Freezing over of lakes,
rivers, and coastal lagoons; the timing of critical ice
depths; establishment of permanent winter snow
cover; and rate of spring ablation are all events criti-
cal to planning inventories and environmental assess-
ments. These events also constitute essential
information for any environmental impact analysis
(U.S. Fish and Wildlife Service 1982; Lent 1985).

In what is certainly the most elaborate mapping
project undertaken for a tundra area and perhaps
one of the most elaborate done anywhere, Walker
et al. (1980) produced a geobotanical atlas of the
Prudhoe Bay region in Alaska. Among the many maps
are ones showing relative sensitivity of areas to oil
spills and sensitivity to damage by vehicles. Also
shown are lichen cover, breeding bird density, and
the thickness of the active layer (layer subject to
seasonal freezing and thawing). One of the more use-

ful maps, an oil spill sensitivity map and the proce-
dures for its production, is described in detail in
Walker et al. 1978. They were able to demonstrate
great differences in the sensitivity of various plant
communities to either diesel or crude oils by meas-
uring live cover on 1-m" ( 10.9-ft") quadrats before
and 1 year after experimental "spills." These map
types are useful to those who must assess and inven-
tory wildlife habitats and who are involved in plan-
ning for tundra areas in relation to potential impacts
on wildlife.

The production of an atlas like that of Walker et
al. (1980) is beyond the budget of most land man-
agers. Furthermore, many of the methodologies used
in producing the maps are not described in detail
and the reader must refer to the references cited in
the atlas.

Sensitivities of alpine tundra areas to human
impacts are also well-documented by Willard and
Marr (1971 ). Using sample species cover measures,
they monitored the disturbance and recovery of
alpine vegetation. In general they reported ex-
tremely slow recovery rates; in some instances no
recovery had occurred after 4 years of protection,
based on total cover and species composition meas-
ures. For Kobresia sp. meadows it was estimated that
damage to the turf caused such severe erosion and
loss of the thin humus-enriched horizons that
hundreds of years might be necessary to restore the
original climax system.

Bell and Bliss (1973) used 10x50 cm (4x20 in.)
clipping plots and line intercept sampling of plant
cover to measure recovery rates of alpine plant com-
munities after human trampling. Dry tundra and
scree slopes were most fragile and slowest to re-
cover. Recovery rates were significantly more rapid
in meadows and snowbed communities. They also
measured rates of downslope movement as indica-
tors of disturbance.

Snow Cover and Tundra Habitats

Any survey or inventory of winter habitats im-
portant to wildlife must include the collection and
analysis of information on snow cover if it is to pro-
vide the manager with a proper assessment of carry-
ing capacity, relative importance of various subareas,
and temporal patterns of use. In their simplest form,
inventories of snow cover may be limited only to
recording the presence or absence and the timing of
accumulation and ablation. Obviously, data on depth
or thickness greatly increase the value of a survey.
Data on density, hardness, ice, crust layers, and pat-
terns of drifting or distribution in relation to topog-
raphy are also desirable.

Pruitt ( 1959) was the first in North America to
employ extensive aerial surveys and snow measure-

Tundra

161

merits to relate snow-cover characteristics to caribou
movements and distribution. The study area for his
classic investigation of migratory caribou ecology lay
in the taiga, with only a small portion in the tundra
zone. Nevertheless, the concepts and techniques
he pioneered are applicable to both zones.

Pruitt made snow measurements using a set of
standard snow instruments developed and adopted
by the National Research Council of Canada (Klein
et al. 1950). With these instruments, several meas-
ures of snow-cover morphology may be made, of
which thickness, density, and hardness are usually
considered most pertinent for ecological investiga-
tions. Richens and Madden (1973) reported on modi-
fications to these instruments to make them more
suitable for wildlife biologists.

Other workers (LaPerriere and Lent 1977; Lent
and Knutsen 1971; Skogland 1978; Brooks and Collins
1985) carried out snow surveys using the Ramsonde
penetrometer (Figure 5). The penetrometer was
developed in the European Alps for avalanche stud-
ies and is most suitable for use with dense, wind-
blown snow with complex thin crusts and ice layers,
such as is typical of alpine and some tundra environ-
ments. It measures the force required to penetrate
a given layer from above.

HAMMER

GUIDE ROD

SHAFT

60° CONE

V

Figure 5. Schematic representation of Ramsonde
penetrometer.

Pruitt's studies and those of later investigators
working with barren-ground caribou (LaPerriere and
Lent 1977) and mountain reindeer (Skogland 1978)
have confirmed the preference of winter ranges with
relatively low depths, densities, and hardnesses.
Musk-oxen have shown even stronger selection for
such winter ranges (Lent and Knutsen 1971). For
musk-oxen, it has been shown repeatedly that an un-
derstanding and knowledge of the prevalence of
crusts, ice layers, and groundfast ice on winter
ranges is critical to proper habitat evaluation (Vibe
1967; Lent and Knutsen 1971; Thomas et al. 1981).

Miller et al. (1982) combined surveys of cari-
bou winter trails and feeding sites with surveys of
snow cover and ice conditions with subsequent veg-
etation sampling. Vegetation was recorded in sum-
mer as percentage frequency of occurrence in
quadrats at the feeding sites and at points away from
the feeding sites. This investigation showed clearly
that several range types with desirable forage species
were being underutilized because of difficulty of
access. Winter feeding was directed primarily at
those poor forage species that were exposed and
directly available. Similarly, Gaare and Skogland
(1975) concluded that only 25% of the potential
winter forage on Norwegian alpine reindeer ranges
was actually available for consumption. Thus, sum-
mer-only inventories of winter ranges can greatly
exaggerate the winter range carrying capacities.
Brooks and Collins (1985) reported on snow survey
methods to assess forage availability.

Although the presence of a deep, dense snow
cover may restrict or prevent winter use of certain
tundra ranges by ungulates, it may also restrict use
by other species, such as wolves (Canis lupus), wol-
verine, and ptarmigan. The same thick snow cover
may provide protection from the abrasive, desiccat-
ing actions of winter winds and windblown snow.
This "subnivean" environment is relatively warm,
moist, and stable and supports small mammals and
invertebrates. Special techniques for studying this
environment were described by Schmidt (1984).

The best lichen growths on reindeer ranges are
associated with moderately deep snow cover be-
cause of the protection and the spring moisture pro-
vided. Snow accumulations in streambeds and other
depressions provide specialized, critical habitats for
numerous species by protecting tundra shrubs from
winter damage. The vegetative associations of these
snowbeds in turn support some specialized wildlife
species (for example, the singing vole [Microtus
miurus]) and provide high quality late summer for-
age for certain herbivores after other forage plants
have gone to senescence. In a detailed study of an
alpine area, Thorn ( 1982) mapped and inventoried
the distribution of the pocket gopher {Thamomys
talpoides) and the distribution of related plant com-

162

Tundra

munities. He found that the plant communities were
most highly correlated with the mean duration and
mean winter depth of the snowpack.

OTHER SPECIAL CONSIDERATIONS FOR
WILDLIFE MANAGEMENT IN THE ARCTIC

The low primary production rates for vascular
plants in tundra communities cannot be emphasized
too strongly. Annual net production estimates be-
tween 50 and 500 g/m (2.1 and 21.0 oz/yd ) are
typical. Lichens and mosses may add 200 g/m (8.4
oz/yd ) to these figures. Furthermore, in both arctic
and alpine environments, the ratio of belowground
to aboveground biomass for vascular plants is ex-
tremely high in comparison with temperate zone
communities; values as high as 30:1 have been re-
ported (Wielgolaski 1975).

Not only are animal populations constrained,
directly or indirectly, by this low primary produc-
tion but many wildlife populations in the arctic are
particularly difficult to inventory or monitor because
their numbers display extreme fluctuations under
natural conditions. Population peaks can markedly
alter vegetation and other habitat features in some
instances, leading to declines and completion of
a population cycle.

Willow ptarmingan in dry- tundra area.

The brown lemming {Lemmns sibiricus) is the
classic example of a cyclic species with extreme
fluctuations in numbers. In identical trapping efforts
for over a decade from less than 1/ha (2.5/a.) to highs
in 2 years when it reached about 200/ha ( 500/a. ) in
the wet tundra of coastal Alaska, density of this spe-
cies has varied. Even in tussock tundra the cyclic
variations can be pronounced. Pruitt (1966) de-
scribed variation of about 700% in biomass of small
mammals trapped on the same plots over 3 years
of study.

Certain predatory birds that are closely tied to
small mammal numbers, such as pomarine jaegers
(Stercorarius pomarinus) and snowy owls, may be
absent over large areas during years of low numbers
of small mammals. The implications of such extreme
fluctuations for the design of monitoring systems
are obvious: baseline conditions are difficult if not
impossible to establish, and relating population
change to casual management factors is also an often
impossible task.

Clumping phenomena also complicate inventory
and monitoring efforts. As noted earlier, even so-
called resident species may be highly mobile in arc-
tic tundra regions. Arctic foxes may travel hundreds
of miles, deserting previous denning areas (Bannikov
1970). Ptarmigan may coalesce by the thousands in
certain favorable river valleys in winter but disperse
widely over hundreds of square miles of upland tun-
dra in summer (Irving et al. 1967; NPR-A Task Force
1979). Caribou may go several years without making
significant use of certain winter ranges. Then, for
reasons yet unknown, they may converge in the
thousands on those same wintering areas for several
successive years.

Not only does such clumping, erratic distribu-
tion patterns and population fluctuations make it
difficult to measure natural baselines by which man's
influences can be determined, they also make it diffi-
cult to apply concepts of sustained yield for either
population or habitat management (Beddington and
May 1977). Sustained stability is rare and generally
unachievable as a management goal in arctic
environments.

Dunbar (1973) stated the problem in somewhat
different terms by pointing out that arctic systems
show great instability on small geographic scales, but
overall stability on a large geographic scale. Thus,
large expanses are necessary for recovery from
either natural or man-made perturbations. Dunbar
(1973:180) sums this up succinctly: "Arctic systems
must have space. They must also have time." The
latter statement is a reference to the slow recovery
rates of these systems following stresses. Dunbar
considered arctic lakes, because of their slow turn-
over rates and because they are often isolated (that
is, lacking the possibilities for rapid restoration from

Tundra

163

a broader system), as the most vulnerable parts of
arctic systems. Migratory shorebirds and waterfowl
are a principal class of consumers dependent in large
part on these lakes and ponds and their relatively
productive invertebrate populations and emergent
vegetation. They are thus vulnerable because of their
ties to the aquatic system. This localized vulnerabil-
ity is usually compensable because of the broad spa-
tial distribution of the populations. Critical
vulnerabilities occur whenever major portions of
populations exist in concentrations (as with the
molting brant in the Teshekpuk Lake area of Alaska)
or as isolated clumps with little interchange (as with
snow goose colonies). Snow goose nesting colonies
represent situations analogous to islands or to iso-
lated lakes. Because of the strong traditions tying
each breeding group with a specific location, they
function as isolated units not readily restored by
inputs from the broader system.

productivity, high degree of annual variability, short
growing season, and habitat patchiness. Migratory
species are relatively large components of the faunas.
The short season for conventional field work and
the rapid phenological changes during that season
present special challenges.

A major limitation to the design of ecologically
based monitoring systems in arctic and alpine habi-
tats is the lack of understanding of successional pro-
cesses and stages.

The distribution, duration, depth, and physical
attributes of snow cover are of particular importance
in tundra habitats and contiguous boreal and alpine
areas. Special techniques are available to describe
and measure these elements of wildlife habitats. Dis-
tribution and physical attributes of permafrost and
frost phenomena are also of special concern in tun-
dra areas.

CONCLUSION

Tundra habitats are dominated by low herba-
ceous growth forms; shrubs are important compo-
nents in restricted areas, and lichens and mosses are
relatively important components in comparison with
other major habitat types. These arctic and alpine
habitats present many unique ecological attributes
for the wildlife manager to consider in inventory,
monitoring, and assessment work. These include low

The immense geographic scale of arctic tundra
areas and the unusually great costs in undertaking
field work there leads to special emphasis on the use
of remote sensing techniques, especially those em-
ploying satellite-borne sensors. Such techniques are
generally less useful in alpine areas, but small-scale
aerial photography has had several applications. Al-
pine areas present special problems because of the
frequent existence of small habitat units and isolated
populations.

Arctic fox.

164

Tundra

LITERATURE CITED

ACEVEDO, D., W WALKER, L. GAYDOS, and J. WRAY.
1982. Vegetation and land cover, Arctic National
Wildlife Refuge, Coastal Plain, Alaska. U.S. Geological
Survey Map 1-1443- Misc. Investigations Ser. 1 sheet.

ANDERSON, J.G., C.H. RACINE, and H.R. MELCHIOR.
1974. Preliminary vegetation map of the Espenberg
Peninsula, Alaska, based on an Earth Resources Tech-
nology Satellite image, Pages 290-310 in Melchior,
H.R. ed. Final Report. Chukchi-Imuruk Survey. Coop.
Park Studies Unit, Univ. of Alaska, Fairbanks. 5 1 7pp.

ANDERSON, J.H. and A.E. BELON. 1973- A new vegetation
map of the western Seward Peninsula, Alaska, based
on ERTS-1 imagery. No. E73- 10305. NTIS. Springfield,
VA. 20pp.

BANFIELD, A.W.F. 1954. Preliminary investigation of the
barren-ground caribou. In two parts. Wildl. Manage.
Bull. Ser. 1. 10A and 10B. Can. Dep. of Northern
Affairs and Nat. Resour. 79pp. and 1 1 2pp.

BANNIKOV, A.G. 1970. Arctic fox in the U.S.S.R. Pages

121-129 in Fuller,W.A. and P.G. Kevan eds. Productiv-
ity and Conservation of Northern Circumpolar Lands.
IUCN Publ. New Ser. 16. Morges, Switzerland.

BEDDINGTON, JR. and RM. MAY. 1977. Harvesting natu-
ral populations in a randomly fluctuating environ-
ment. Science 197:463-465.

BEE, J.W. and E.R. HALL. 1956. Mammals of northern
Alaska on the Arctic Slope. Misc. Publ. 8. Mus. Nat.
Hist., Univ. Kansas, Lawrence. 309pp.

BELL, KL. and L.C. BLISS. 1973. Alpine disturbance studies:
Olympic National Park. Biol. Cons. 5:25-32.

BERGMAN, R.D., R.L. HOWARD, KF. ABRAHAM, and M.W.
WELLER. 1977. Water birds and their wetland re-
sources in relation to oil development at Storkerson
Point, Alaska. Resource Publ. 129. U.S. Dep. Inter.,
Fish and Wildl. Serv. Washington, DC. 38pp.

BILLINGS, W.D. 1973- Arctic and alpine vegetations: simi-
larities, differences and susceptibility to disturbance.
Bio Science 23:697-704.

BLISS, L.C. 1981. Tundra biome: past and present. Pages 3-
24 in Bliss, L.C., O.W. Heal, and J.J. Moore eds. Tun-
dra Ecosystems: A Comparative Analysis. Cambridge
Univ. Press, Cambridge.

BROOKS, J. and W. COLLINS. 1985. Snow cover and inter-
pretation of vegetation habitat inventions. Pages 203-
210 in LaBau, V.J. and C.L. Kerr eds. Inventorying
Forest and Other Vegetation of the High Latitude and
High Altitude Regions. Proc. Int. Symp. Am. For. Reg.
Conf. Soc. Am. For. Washington, DC.

BROWN, J. and R.L. BERG eds. 1980. Environmental engi-
neering and ecological baseline investigations along
the Yukon River — Prudhoe Bay Haul Road. CRREL
Report 80-19. Hanover, NH. 187pp.

BROWN, R.W., R.S. JOHNSTON, and K VAN CLEVE. 1978.
Rehabilitation problems in alpine and arctic regions.
Pages 23-44 in Reclamation of Drastically Disturbed
Land. Soil Sci. Soc. Am. Madison, WI.

COURTRIGHT, A.M. 1959. Range management and the
genus Rangifer: A review of selected literature. Un-
publ. M.S. Thesis, Univ. Alaska. Fairbanks. 172pp.

CRAIGHEAD, J.J., J.S. SUMNER, and G.B. SCAGGS. 1982. A
definitive system for analysis of grizzly bear habitat
and other wilderness resources. Wildlife- Wilderness
Inst. Monogr. 1, University of Montana, Missoula.
279pp.

DERKSEN, D.V., W.D. ELDRIDGE, and M.W. WELLER.

1982. Habitat ecology of Pacific black brant and other
geese moulting near Teshekpuk Lake, Alaska. Wild-
fowl 33:39-57.

DRISCOLL, R.R., D.L. MERKEL, D.L. RADLOFF, D.E. SNY-
DER, and J.S. HAGIHARA. 1983. An ecological land
classification framework for the United States. U.S.
Dep. Agric, For. Serv. Misc. Publ. 1439.

DUNBAR, M. 1973. Stability and fragility in arctic ecosys-
tems. Arctic. 26:179-185.

EVERETT, KR. 1982. Some recent trends in the physical
and chemical characterization and mapping of tundra
soils, Arctic Slope of Alaska. Soil Sci. 133:264-280.

FYFE, R. and R OLENDORFF. 1976. Minimizing the danger
of nesting studies to raptors and other sensitive spe-
cies. Can. Wildl. Serv. Occ. Publ. 23- Ottawa, Canada.
17pp.

GAARE, E. and T. SKOGLAND. 1975. Wild reindeer food
habits and range use at Hardangervidda. Pages 196-215
in Wielgolaski, F.G. ed. Fennoscandian Tundra Ecosys-
tems. Part 2. Springer Verlag, Berlin.

GABRIEL, H.W. and S.S. TALBOT. 1984. Glossary of land-
scape and vegetation ecology for Alaska. BLM-Alaska
Tech. Rep. 10. U.S. Dep. Inter., Bur. of Land Manage.
Anchorage, AK 137pp.

GALL, L.J. 1984. Population structure and recommenda-
tions for conservation of the narrowly endemic alpine
butterfly (Boloria acronemia [Lepidoptira: Nymphali-
dae]). Biol. Conserv. 28:111-138.

GEORGE, T.H., W.J. STRINGER, J.E. PRESTON, W.R. FI-
BIEL, and PC. SCORUP. 1977. Reindeer range inven-
tory in western Alaska from computer-aided digital
classification of LANDSAT data. 28th Alaska Science
Conf. Anchorage, AK

GREIG-SMITH, P. 1983. Quantitative plant ecology. 3rd.
ed. Univ. Calif., Berkeley. 359pp.

HALL, D.K, J.P. ORMSBY, L. JOHNSON, and J. BROWN.

1980. Landsat digital analysis of the initial recovery of
burned tundra at Kokolik River, Alaska. Remote Sen-
sing of Environment 10:263-272.

HESJEDAL O. 1974. Terrestrial vertebrates. Pages 475-570
in Ives, J.D. and R.G. Barry eds. Arctic and Alpine
Environments. Methuen, London.

and J. LARSSON. 1975. Remote sensing of vegeta-
tion in conservation of tundra landscapes. Pages 237-
242 in Wielgolaski, F.E. ed. Fennoscandian Tundra
Ecosystems; Part 2, Animals and Systems Analysis.

HOFFMANN, R.S. 1974. Terrestrial vertebrates. Pages 475-
570 in Ives, J.D. and R.G Barry, eds. Arctic and Al-
pine Environments. Methven, London.

1984. Small mammals in winter. The effects of

altitude, latitude and geographic history. Pages 9-23
in Merritt, J.E. ed. Winter Ecology of Small Mammals.
Carnegie Mus. Spec. Publ. 10.

HOK, J. R. 1969- A reconnaissance of tractor trails and
related phenomena on the North Slope of Alaska. Un-
publ. U.S. Dep. Inter., Bur. of Land Manage. Anchor-
age, AK.

HOLT S. 1971. Some effects of vehicle operation on Alaska
arctic tundra. M.S. Thesis, Univ. of Alaska, Fairbanks.
85pp.

1980. Vegetation patterns and effects of grazing on

caribou ranges in the Sondre Stromfford area, West
Greenland. Pages 57-63 in Reimers, E., E. Gaare, S.
Skjenneberg eds. Proc. 2nd Int. Reindeer/Caribou
Symp. Roros, Norway.

Tundra

165

IRVING, L., G.C. WEST, L.J. PEYTON, and S. PANEAK

1967. Migration of willow ptarmigan in arctic Alaska.
Arctic 20:77-85.

IVES, J. D. 1970. Arctic tundra: how fragile? A geomorphol-
ogist's point of view. Pages 39-42 in Hare, E.K ed.
The Tundra Environment. Royal Soc. Canada. Univ.
Toronto Press, Toronto.

JENKINS, J.F. and D.H. ASHTON. 1970. Productivity stud-
ies on Macquarie Island vegetation. Pages 851-863
in Holdgate, M.W. ed. Antarctic Ecol., Vol. 2. Aca-
demic Press, London.

KESSEL, B. and G.B. SCHALLER. I960. Birds of the upper
Sheenjek Valley, Northeastern Alaska. Biol. Pap., Univ.
Alaska, Fairbanks. 4:1-59-

KING, J.G. 1970. The swans and geese of Alaska's arctic
slope. Wildfowl. 21:11-17.

KING, R. 1979. Results of serial surveys of migratory birds
on NPR-A in 1977 and 1978. Pages 187-226 in Lent,
P.C. ed. Studies of Selected Wildlife and Fish and their
Use of Habitats on and Adjacent to the National Pe-
troleum Reserve — Alaska — 1977-1978. Field Study
Rep. 3- U.S. Dep. Inter., Fish and Wildl. Serv. Anchor-
age, AK

KLEIN, G.J., DC. Pearce, and L.W. Gold. 1950. Method of
measuring the significant characteristics of a snow
cover. Nat. Res. Coun. Can. Tech. Mem. 18. 56pp.

LaPERRIERE, A.J. 1978. Use of Landsat imagery for wild-
life habitat mapping in northeastern and east-central
Alaska. Final Report. NASA Contract NAS520915.
NTIS. Springfield, VA. 36pp.

and P.C. LENT. 1977. Caribou feeding sites in rela-
tion to snow characteristics in northeastern Alaska.
Arctic. 30:101-108.

LENT, P.C. 1980. Synoptic snowmelt patterns in arctic
Alaska in relation to caribou habitat use. Pages 71-76
in Reimers, E., E. Gaare, S. Skjenneberg eds. Proc.
Second Int. Reindeer/Caribou Symp. Roros, Norway.
Direktoratet for Vilt of Verskvannsfisk, Trondheim.

1985. Cold region vegetation information needs

from the perspective of wildlife and fisheries. Pages
20-27 in LaBau, V.J. and CL. Kerr eds. Inventorying
Forest and Other Vegetation of the High Latitude and
High Altitude Regions. Proc. Int. Symp. Am. For. Reg.
Conf, Soc. Am. For. Washington, DC.

and D. KNUTSEN. 1 971. Musk-ox and snow cover

on Nunivak Island, Alaska. Pages 50-62 in Haugen, A.
ed. Proc. Snow and Ice in Relation to Wildlife and
Recreation Symp. Iowa State Univ., Ames.

and A.J. LaPERRIERE. 1974. Applications of ERTS

imagery to the study of caribou movements and win-
ter habitat. Final Rep. NASA Contract NAS521833.
NTIS. Springfield, VA. 50pp.

LEWIS, ME. 1970. Alpine rangelands of the Uinta Moun-
tains, Ashley and Wasatch National Forests. U.S. Dep.
Agric, For. Serv., Reg. 4. Ogden, UT 75pp.

LINELL, K.A. and J.C.F. TEDROW. 1981. Soil and permafrost
surveys in the Arctic. Clarendon Press, Oxford.
279pp.

MCKENDRICK, J.D. 1981. Responses of arctic tundra to
intensive musk-ox grazing. Agroborealis 13:49-55.

MILLER, F.L., E.J. EDMONDS, and A. GUNN. 1982. Forag-
ing behaviour of Peary caribou in response to spring-
time snow and ice conditions. Canadian Wildl. Serv.
Occ. Pap. 48. 41pp.

MOITORET, C.S., PA. MILLER, R.M. OATES, and MA.
MASTELLER. 1985. Terrestrial bird populations and
habitat use on coastal plain tundra of the Arctic Na-

tional Wildlife Refuge. Pages 362-446 in 1984 Update
Report, Arctic National Wildlife Refuge Coastal Plain
Assessment. U.S. Dep. Inter., Fish and Wildl. Serv.
Anchorage, AK

NODLER, FA., A.J. LaPERRIERE, and DR. KLEIN. 1978.
Vegetation type mapping in northwestern Alaska in
relation to caribou and reindeer range potentials.
Alaska Coop. Wildlife Research Unit Spec. Rep. 2.
Univ. of Alaska, Fairbanks. 33pp.

NPR-A (National Petroleum Reserve — Alaska) Task Force.
1978. Fish and Wildlife resources. Values and Re-
source Analysis. Section 6, Vol. 3, Study Rep. 2. U.S.
Dep. Inter. 224pp.

1979. Ecological Profile. National Petroleum Re-
serve in Alaska. Study Rep. 4. U.S. Dep. Inter.

OWENSBY, C.E. 1973. Modified step-point system for
botanical composition and basal cover estimates. J.
Range Manage. 26:302-303.

PALMER, L.J. 1926. Progress of reindeer grazing investiga-
tions in Alaska. U.S. Dep. Agric. Bull. 1423- 37pp.

PAMPLIN, W.L 1979. Construction-related impacts of the
Trans-Alaska Pipeline System on terrestrial wildlife
habitats. Special Report 24, Joint State Federal Fish
and Wildlife Advisory Team. U.S. Dep. Inter., Fish and
Wildl. Serv. Anchorage, AK 1 32pp.

PEGAU, RE. 1968. Reindeer range appraisal in Alaska.
Arctic 21:255-259.

1970. Effect of reindeer trampling and grazing on

lichens. J. Range Manage. 23:95-97.

PEWE, T. 1983- Alpine permafrost in the contiguous

United States: A review. Arctic and Alpine Research

15:145-156.
PITELKA, F.A. 1979. An avi-faunal review for the Barrow

region and North Slope of Arctic Alaska. Arctic and

Alpine Research 6:161-184.
PRICE, L.W. 1981. Mountains and Man. Univ. of Calif. Press,

Berkeley. 506pp.
PRUITT, W.O., Jr. 1959. Snow as a factor in the winter

ecology of barren-ground caribou. Arctic. 12:159-179-

1 966. Ecology of terrestrial mammals. Pages 519-

564 in Environment of the Cape Thompson Region,
Northwest Alaska. U.S. Atomic Energy Commission.
U.S. Govt. Print. Off.

1978. Boreal Ecology. Edward Arnold Publ. Ltd.

London, iv + 73pp.

RACINE, C.H. 1979. The 1977 tundra fires in the Seward
Peninsula, Alaska: effects and initial revegetation.
BLM-Alaska Technical Report 4. U.S. Dep. Inter., Bur.
Land Manage. Anchorage, AK 51pp.

REEVES, H.M., F.G. COACH, AND RE. MUNRO. 1976.
Monitoring arctic habitat and goose production by
satellite imagery. J. Wildl. Manage. 40:532-541.

REYNOLDS, H.V., J. CURATOLO, and R. QUIMBY. 1976.
Denning ecology of grizzly bears in northeastern
Alaska. Pages 403-409 in Pelton, M.J., J.W., Lentfer,
and E. Folk eds. Bears - their biology and Manage-
ment. IUCN New Series 40. Morges, Switzerland.

RICHENS, V.B. and C.G. MADDEN. 1973. An improved
snow study kit. J. Wildl. Manage. 37:109-113.

RITCHIE, R. 1979. A survey of cliff-nesting raptors and
their habitats. Pages 313-336 in Lent, P.C. ed. Studies
of Selected Wildlife and Fish Populations and their
Use of Habitats on and Adjacent to NPR-A. 1977-
1978. U.S. Dep. Inter., Fish and Wildl. Serv. Anchor-
age, AK

SCHMIDT, WD. 1984. Materials and methods of subnivean
sampling. Pages 25-32 in Merritt, J.E. ed. Winter Ecol-

166

Tundra

ogy of Small Mammals. Carnegie Mus. Nat. Hist. Spec.
Publ. 10.

SIMS, R.A. 1983- Ground-truth and large scale 70 mm aerial
photographs in the study of reindeer winter range-
land, Tuktoyaktuk Peninsula area, NWT. PhD. disserta-
tion. Univ. of British Columbia, Vancouver. 1 78pp.

SKOGLAND, T. 1978. Characteristics of the snow cover
and its relationship to wild mountain reindeer (Ran-
gifer tarandus tarandiis L. ) feeding strategies. Arctic
and Alpine Research 10:569-580.

SKUNCKE, F. 1969. Reindeer ecology and management in
Sweden. Biol. Paper. Univ. Alaska, Fairbanks. 82pp.

SMITH, R.H. AND R.P. ALLEN. 1948. An aerial waterfowl
reconnaissance of the far north. U.S. Dep. Inter., Fish
and Wildl. Serv. Spec. Rep. 60:5-12.

TENER, J.S. 1965. Musk-oxen in Canada. Can. Wildl. Serv.
Monogr. Ser. 2. Dept. of Northern Affairs and Nat.
Resources. Ottawa. 166pp.

THILENIUS, J.F. 1979. Range management in the alpine
zone. Pages 43-64 in Johnson, D.A. ed. Special Man-
agement Needs of Alpine Ecosystems. Range Science
Serv. Society for Range Management, Denver, CO.
5:43-64.

THOMAS, DC, F.L. MILLER, R.H. RUSSELL, and G.R. PAR-
KER. 1981. The Bailey Point region and other musk-
ox refugia in the Canadian Arctic: A short review.
Arctic. 34:34-36.

THORN, C.E. 1982. Gopher disturbance: Its variability by
Braun-Blanquet vegetation units in the Niwot Ridge
alpine tundra zone, Colorado Front Range, U.S.A.
Arctic and Alpine Research 14:45-51.

U.S. FISH AND WILDLIFE SERVICE. 1982 Initial report
Baseline study of the fish, wildlife and their habitats.
Arctic National Wildlife Refuge coastal plain resource
assessment. U.S. Fish and Wildl. Serv., Anchorage,
AK. 507pp.

. 1983-1984. Update report. Baseline study of the

fish, wildlife and their habitats. Arctic National Wild-
life Refuge coastal plain resource assessment. U.S.
Dep. Inter., Fish and Wildl. Serv. Anchorage, Alaska.

613pp.

VAN CLEVE, K 1977. Recovery of disturbed tundra and
taiga surfaces in Alaska. Pages 422-455 in Cairns,
J., KL. Dickson, and E. Herricks eds. Proc. of an Int.
Symp. on the Recovery of Damaged Ecosystems. Vir-
ginia Polytechnic Inst, and Univ. Blacksburg, VA.

VIBE, C. 1967. Arctic animals in relation to climatic fluc-
tuations. Meddelelser om Gronland 170(5):l-227.

VIERECK, LA. and C.T. DYRNESS. 1980. A preliminary

classification system for vegetation of Alaska. U.S. Dep.
of Agric, For. Serv. General Tech. Rep., PNW-106.

, C.T. DYRNESS, and A.R. BATTEN. 1982. Revision

of preliminary classification for vegetation of Alaska.
Inst. North. Forestry, Univ. Alaska, Fairbanks. 72pp.

WALKER, DA., K.R. EVERETT, P.J. WEBBER, and J.

BROWN. 1980. Geo-botanical atlas of the Prudhoe
Bay Region, Alaska. CRREL Report 80-14. Cold Re-
gions Research and Engineering Laboratory. Hanover,
NH. 69pp.

, P.J. WEBBER, KR. EVERETT, and J. BROWN. 1978.

Effects of crude and diesel oil spills on plant commu-
nities at Prudhoe Bay, Alaska, and the derivation of oil
spill sensitivity maps. Arctic 31:242-258.

WIELGOLASKI, F.E. 1975. Productivity of tundra ecosys-
tems. Pages 1-12 in Productivity of World Ecosys-
tems. Proc. of a Symposium. National Academy of
Sciences. Washington, DC.

WILLARD, BE. and W. MARR. 1971. Recovery of alpine
tundra under protection after damage by human
activities in the Rocky Mountains of Colorado. Biol.
Cons. 3:181-190.

WILLIAMSON, F.S.L., M.C. THOMPSON, and J.Q. HINES.
1966. Avi-faunal investigations. Pages 437-480 in Wili-
movsky, N.J. and J.N. Wolfe eds. Environment of the
Cape Thompson Region, Alaska. U.S. Atomic Energy
Comm. Washington, DC.

ZWINGER, A.H. and BE. WILLARD. 1972. Land above the
trees: a guide to American alpine tundra. Harper
and Row, New York, NY. 489pp.

Tundra

167

RIPARIAN
HABITAT

Robert D. Ohmart and Bertin W. Anderson

Center for Environmental Studies
Arizona State University
Phoenix, AZ 85281

Editor's Note: This chapter is the first of two on
wetland habitats. These habitats are extremely im-
portant, not only because of their high inherent
wildlife value, but also because of their effects on
the adjacent upland and aquatic areas and their
associated biota This chapter covers wetland areas
associated with running water, while the following
chapter on marshes covers areas of standing water.
These are not universally accepted definitions but
provide a convenient breakdown for this book. The
length of this chapter reflects the high importance
of and current interest in riparian areas. However,
much still needs to be learned about these areas,
and a critical need exists for better management of
them. Inventorying and monitoring riparian habi-
tats will be a central part of such efforts.

INTRODUCTION

One of the most important assignments in the
career of a wildlife biologist is to monitor or inven-
tory riparian ecosystems. Only a decade ago few
people, including wildlife biologists, had any appreci-
ation or knowledge of these very limited and highly
valuable wildlife habitats. Even today, books on wild-
life habitats or plant communities seldom separate
riparian ecosystems from adjacent upland plant com-
munities. Before undertaking an assignment, we
recommend obtaining copies of four riparian sym-
posiums (Johnson and Jones [1977]; Johnson and
McCormick [1978]; Warner and Hendrix [1984]; and
Johnson et al. [1985]) to use as references.

Riparian as an adjective is defined as "relating to
or living or located on the bank of a natural water-
course (as a river) or sometimes of a lake or a tide-
water" (Webster's New Collegiate Dictionary
1979:991 ). To many, riparian is synonymous with
wetland, but wetland is often defined as consisting
primarily of emergent or marsh communities, which
are not discussed in this chapter. (See Chapters 10,
1 1 , and 1 2 for treatment of marshes, streams, and
lakes, respectively.) For this chapter, which ad-
dresses only terrestrial riparian ecosystems, we will
adhere to the following definition: "A riparian associ-
ation of any kind [excluding marshes] is one which
is in or adjacent to drainageways and/or their flood-
plains and which is further characterized by species
and/or life-forms different than that of the immedi-
ately surrounding non-riparian climax" (Lowe
1964:62). This definition includes plant communities
along permanently flowing or intermittent drainages.
Some of these drainages may not flow for many years
or even in our lifetime, but they are riparian commu-
nities if the plant species along these drainages are
different from those of the adjacent upland.

Obligate or riparian-dependent species such as
cottonwoods (Populus sp. ) and willows (Salix sp. )

Riparian Habitats

169

are frequently referred to in the literature as phrea-
tophytes, referring to vegetation species having their
roots in perennial ground water or in the capillary
fringe above a water table. Most of these species
transpire large quantities of water, and water man-
agers believe that if the streamside vegetation is
removed, this water will be saved or remain in the
aquifer.

In past water management practices, thousands
of acres of riparian vegetation have been removed to
prevent this wicking of water into the atmosphere.
For example, between 1967 and 1971, about 21,600
ha (54,000 a.) of riparian vegetation was removed
along the Pecos River in New Mexico to save water.
Although about 18,800 ha (47,000 a.) of the origi-
nally cleared area has remained cleared since 1971
(U.S. Department of Interior, Bureau of Reclamation
1979), preliminary results indicate that the amount
of water saved is probably insignificant. Riparian
ecosystems do not stand alone; they are fed by wa-
tersheds, which when destroyed, also destroy the
riparian ecosystem.

Nutrients, water, and detrital materials are trans-
ported into the riparian system from its watershed.

A healthy watershed generally indicates a healthy
riparian system. Degraded watersheds produce high
surface runoff carrying valuable soil into the stream,
which reduces productivity of both the aquatic and
terrestrial portions of the system. As any competent
hydrologist knows, to reduce the volume and sever-
ity of floods, one must start with the point source —
the watersheds; other efforts are "treating symptoms
and not the disease."

The importance of western riparian habitats to
humans has long been recognized, as indicated by
the early settlement patterns of native North Ameri-
cans and Europeans. During drought periods, settle-
ments that were not near permanent water sources
were forced to relocate; many died in the process.
Riparian habitats provided and still provide water,
rich fertile soils for agriculture, lush forage for do-
mestic livestock, recreation, and home sites. Their
importance is amplified in the arid western states,
but is also obvious in the East where riparian areas
are termed bottomland hardwood habitats.

Riparian habitats in the West are very limited
when compared to the amount of acreage they con-
stitute versus upland habitats. Riparian ecosystems

Degraded riparian area.

170

Riparian Habitats

when compared with upland habitats may total up to
0.5% of the landscape or < 0.1%. Because of the
small and finite nature of riparian habitats, their vital-
ness to human survival in arid environments, their
recreational values, and their high fishery and wild-
life values, they should receive critical concern in all
land-planning and management efforts. For example,
Johnson (1978) reported that 64 wildlife species
presently listed as endangered, and 47 more species
being considered for listing, are dependent on ripar-
ian habitats. In the past, riparian habitats have been
treated as sewage transport systems and refuse land-
fill sites and have been subjected to numerous other
types of habitat degradation.

The treatment of riparian habitats in the past
and their current condition is alarming, which ampli-
fies the need to pay special management concern to
these ecosystems. For example, estimates are that
70-90% of the natural riparian ecosystems in the U.S.
have been lost to human activities (U.S. Council on
Environmental Quality 1978; Warner 1979a). Re-
gional losses in these ecosystems have been esti-
mated to exceed 98% in the Sacramento Valley in
California (Smith 1977) and 95% in Arizona (Warner
1979b). Johnson and Carothers (1981) estimated

that in the Rocky Mountains/Great Plains region, 90-
95% of the cottonwood-willow riparian ecosystems
of the plains and lower foothills have been lost. Pos-
sibly as much as 80% of the remaining riparian eco-
systems in the U.S. (both privately and publicly
owned) are in unsatisfactory condition and are domi-
nated by human activities (Almand and Krohn 1978;
Warner 1979b). In the West, these ecosystems sup-
port a disproportionately greater number of wildlife
species than their upland counterparts.

Public support for wildlife and its habitat needs
must be interpreted with caution. Kellert (1980),
reporting on a nationwide opinion poll that focused
on land-use allocations for wildlife, showed that
many Americans were willing to make economic
sacrifices in commodity resource production for en-
dangered nongame species and certain featured big
game species. Sixty percent agreed that livestock
grazing on public lands should be limited if it de-
stroyed vegetation used by wildlife, even if it re-
sulted in higher meat prices; 34% disagreed. Over
75% believed that logging should be done in a man-
ner to enhance wildlife even if lumber and paper
prices rose. Conversely, almost 50% believed that
natural resources must be developed, even if it

Same riparian area, 1 0-years later.

Riparian Habitats

171

meant less wilderness and lowered wildlife popula-
tions. These results indicate public awareness for the
needs of wildlife and management changes that need
to be enacted on public lands to maximize natural
resource values. The public at large generally sup-
ports the concept of multiple-use management on
public lands, but interest groups strongly disagree on
specific issues.

Sports enthusiasts in eastern Oregon favorably
responded to improved grazing management prac-
tices (Megank and Gibbs 1979). Almost 70% of
anglers surveyed stated that their recreation experi-
ences would be reduced by management practices
that further degraded riparian systems. Their reac-
tion to fences was positive in that it represented
better livestock control. Hunters thought that man-
agement that improved forage production for live-
stock would also help deer and elk.

Many values of riparian habitats to our society,
seldom considered by terrestrial wildlife biologists,
should be factored in as sound arguments toward
managing these systems in their natural state. The ri-
parian vegetation adjacent to streams or even large
rivers is extremely important as an energy source to
the aquatic organisms (see Streams, Chapter 1 1 ). In
small headwater streams, 99% of the energy for het-
erotrophic organisms comes from the vegetation
along the stream, whereas only 1% comes from pho-
tosynthesizing autotrophs (Cummins 1974). In large
river systems, such as the Missouri River, as much
as 54% of the organic matter consumed by fish is of
terrestrial origin (Benner in Kennedy 1977). A fac-
tor that has not, to our knowledge, been seriously
considered by fishery biologists is the quality of the
organic input or species composition of the stream-
side vegetation. Trees may be desirable as shade
to prevent large fluctuations in water temperature,
but some tree species will ultimately prove more
important as energy input sources than others, e.g.,
cottonwood and willow leaves and other tree parts
are probably of more value to aquatic detritivores
than leaves and other parts of the exotic salt cedar
(Tamarix chinensis). The former would certainly
impose less change on total dissolved solids and
water chemistry than the latter. Biologists must man-
age for healthy streamside vegetation and, as knowl-
edge progresses, some efforts should be directed
toward encouraging establishment of tree species
that have higher nutrient input value to the aquatic
fauna. This approach toward managing terrestrial
vertebrates is already underway along large river
systems.

Streamside vegetation is very important in deter-
mining the structure and function of stream ecosys-
tems (Knight and Bottorff 1984). Mahoney and
Erman (1984) found that riparian vegetation is an

important source of food to stream organisms, pro-
vides shade over small-order streams, and serves
to stabilize banks in preventing excessive sedimenta-
tion and intercepting pollutants. Asmussen et al.
(1977) reported that vegetation buffer strips were
very effective in reducing pollution from agricultural
chemicals. Karr and Schlosser (1977, 1978) reported
that proper management of streamside vegetation
and the channel may substantially improve water
quality in agricultural watersheds. Corbett and Lynch
( 1985) stressed the importance of streamside zones
in water-quality management for municipal water
supplies. Haupt (1959) presented guidelines for
buffer strip widths in road-building projects, and
Benoit (1978) similarly presented guidelines for tim-
ber harvest operations in Oregon. Treating water to
bring it to potable standards is expensive; stopping
pollution at its source and managing for productive
riparian vegetation will significantly reduce these
costs. Finally, riparian vegetation can be important in
flood-control efforts (Chaimson 1984) by reducing
water velocity and its erosive energy during flood
stage (Li and Shen 1973). The vegetation may also
reduce streambank damage from ice, log debris, and
animal trampling (Platts 1979; Swanson et al. 1982);
armor levees; and prevent channel changes during
high flows.

Importance of riparian systems to wildlife has
not been quantified or demonstrated to any convinc-
ing extent until the past 1 5 years. The efforts of
Carothers et al. (1974) to quantify avian densities in
cottonwood habitats along the Verde River in central
Arizona and those of Ohmart and Anderson ( 1974)
along the lower Colorado River were beginnings.
A riparian symposium (Johnson and Jones 1977;
Hehnke and Stone 1978; Thomas et al. 1979) along
with others were fruitful in focusing attention on
these ecosystems. These studies and subsequent ones
indicated that some of the highest densities of breed-
ing birds in North America were found in riparian
habitats, and more than 60% of the vertebrates in
the arid Southwest were obligate to this ecosystem
(see Ohmart and Anderson 1982 for a review). An-
other 10-20% of the vertebrates were facultative
users (present for a portion of the annual cycle but
not fully dependent on riparian habitats) of stream-
side vegetation. Mosconi and Hutto (1982) reported
that in western Montana, 59% of the species of land
birds use riparian habitats for breeding, and 36%
breed only in riparian habitats. Similarly, of the 363
species of land vertebrates in the Great Basin of
southeast Oregon, 299 either directly depend on
riparian habitats or utilize them more than any other
habitat types (Thomas et al. 1979). Therefore, if
these ecosystems were totally lost or continued to
be reduced to vestiges of their original state, con-
ceivably 60-80% of our native wildlife species could
be lost in the western U.S. The Colorado River is a
classic example of this, in that at least four species of

172

Riparian Habitats

birds have recently been extirpated and unless some
dramatic management changes are made, another
six species could be lost in the next 20 years
(Hunter et al. unpubl. data).

RIPARIAN HABITAT CLASSIFICATION

Small riparian ecosystems occurring in wa-
tersheds at higher elevations eventually connect to
form major drainage systems. These major drainages,
such as the Green River in Wyoming, connect into
larger systems, such as the Colorado River. To aid in
understanding these western riparian ecosystems at
different elevations, we recommend using the Life
Zone concept developed by Merriam ( 1 890 ), which
is widely used in the West for its utilitarian value.

The Brown et al. (1979) concept for naming
riparian ecosystems in Merriam's Life Zones is also
widely accepted in the West because its hierarchical
system for North American biomes has an evolution-
ary and genetically based approach. It is also
digitized, which makes it computer-compatible. Im-
portantly, it is concordant with the Life Zone
concept.

Succession

In highly modified or managed rivers, such as
the lower Colorado, Pecos, and Rio Grande, there is
little evidence of classical succession. If an area is
burned or cleared, it tends to return as trees or
shrubs and remains in that state. Succession is poorly
studied and understood, and it may be that riparian
floodplains are so rich in nutrients and water that
classical plant succession does not occur. Brady et al.
(1985) presented the developmental continuum of
a riparian gallery forest ranging from a nursery bar to
a mature forest. Bock and Bock ( 1985) presented
data on patterns of reproduction in a species of syca-
more (Platanns urightii) in southeastern Arizona.

Communities

Naming individual plant communities varies
depending on the biologist mapping and sampling
the communities, but is generally based on one or
more dominant species (either by density or stat-
ure). For example, a cottonwood-willow community
may only have 1 cottonwood for every 10 willows in
a stand, but because of the size and presence of cot-
tonwoods, it may be called a cottonwood-willow
community. In another example, a community may
be called honey mesquite (Prosopis glandulosa)
although it contains mostly shrubs. What a commu-
nity is called is unimportant as long as the commu-
nity names and accompanying written descriptions
can be interpreted by others. Iaurenzi et al. (1983)

presented a habitat classification system for mixed
broadleaf riparian forests in the Upper Sonoran Life
Zone.

Structural Types

Some riparian ecosystems may be extremely
dynamic through time whereas others are relatively
stable. Recognizing and classifying structural stages
(young through mature stages) allows a quick assess-
ment of the riparian ecosystem's health. If no young
plant communities develop to replace the old, deca-
dent communities of similar species composition,
then animal species dependent on those communi-
ties may be lost to the fauna. Also, young plant com-
munities may support a fauna much different from
that found in a mature plant community. These
structural differences (in plant community age
classes and foliage layers) are very important in man-
aging for maximum riparian productivity and verte-
brate species richness.

The concept of structural classification of ripar-
ian communities is an important one and not difficult
to understand if an area is envisioned as going from
bare soil to supporting a mature cottonwood-willow
forest; all the structural types would then be present
over this continuum. Figure 1 shows the structural
types in lower elevational systems; Type VI is the
young or beginning community, and as it grows it
passes to a V, IV, etc., until it becomes a Type I
which is a mature cottonwood-willow community. In
Type VI, most of the foliage volume is in the grass
and shrub layers; as the community matures, a good
overstory (Type II) shades out much of the under-
story. Type I contains overstory, midstory, and un-
derstory as some trees die opening the overstory for
ingression of shorter trees, shrubs, and annuals. By
dividing a continuous process into stages or types,
both plant communities and structural types can be
assessed together in demonstrating user-oriented
impacts. However, structural complexity and mean
canopy height are generally reduced where riparian
systems are under heavy water management, live-
stock grazing, pollution, and recreational activities.

In the following example of structural types, a cot-
tonwood-willow community was used to demon-
strate Types VI (young) through I (mature). At
higher elevations this community could consist of
sycamores, narrow-leaf cottonwoods (Popnlns an-
gustifolia), quaking aspen (P. tremuloides), or other
tree species that have similar vertical and foliage
density attributes as cottonwood-willow. Not all tree
species form communities that have the same
growth pattern as those discussed above, but their
community development shows some of the same
structural stages. For example, salt cedar communi-
ties generally reach maturity as structural Type II,

Riparian Habitats

173

1 J v

/(^ [A

I $ I *«JL

J^j^

j^^yinkj

W^\

\ / ^

is-A

*\r\

j\~ >

~fir

JqL

L •»

.vKj ^~

JU,/,-^

VI

$$*%

Figure 1. Vegetation structural types found in lower elevations.

honey mesquite as Type III, and arrowweed (Tes-
saria sericea) and other shrub communities are al-
ways Type VI. Managers may find this classification
system less confusing than trying to establish a more
complex and, hence, more complicated approach
to recognizing structural types. Most importantly, the
structural types are real to wildlife and its use of the
vegetation.

In habitat discussions today of the Colorado
River, managers have wanted to know the commu-
nity and structural type under consideration. They
can easily visualize the community and structural
types, and we have provided a document that gives
species richness, densities, and other wildlife use
values seasonally for 3 to 5 years along with mean
values. These data are available for only a few ripar-
ian ecosystems, and our approach in this chapter
is to help managers initiate development in their re-
source management areas. The system can always
be expanded and improved toward better managing
these invaluable wildlife habitats.

IMPORTANT HABITAT ATTRIBUTES TO
WILDLIFE

Discerning habitat attributes that are most im-
portant to wildlife is not an easy task, and reliable

identification of those habitat components only
comes from long-term, in-depth studies. Seasonal and
annual variances of wildlife numbers are generally
very high, and until a number of seasons and years
have been studied, the habitat components that are
truly important to wildlife may not be discernible or
fully understood.

Four habitat components, not discussed in depth
in this section but important to wildlife, are plant
community size (number of acres), continuity of
riparian habitat along the streambed, edge or eco-
tone, and water. The streambed has been widely
discussed in island biogeography theory and in terms
of habitat fracturing. The more extensive stands of a
riparian forest are reduced, the less wildlife value
they have for some species. As the tract gets smaller,
its wildlife values diminish to the point that the for-
est contributes little, if any, to overall wildlife values.
It probably reaches its zenith of importance in ripar-
ian habitats at lower elevations where expansive
alluvial floodplains support extensive stands of con-
tinuous wildlife habitat. T.E. Martin (Arizona State
University, unpubl. ms) identified area size as the pri-
mary factor accounting for variation in number of
breeding pairs of birds in five of seven of his eco-
logical groups in high-elevation, riparian habitats in
central Arizona. As knowledge expands and under-
standing of the needs of wildlife increases, we ob-

174

Riparian Habitats

serve that continuous areas of some particular
riparian community are more than twice the value of
that same community in 0.8- to 4-ha (2- to 10-a.)
blocks. For example, many bird species have been
lost along the Colorado River and others will be lost
in the near future as island size stands of cotton-
wood-willow are reduced. Stands of 28 ha (70 a.)
only begin to fill the needs of some bird species,
allowing the presence of a few breeding pairs at
best. For these reasons, habitat island size is impor-
tant regardless of whether it is along small streams at
high elevations or along rivers in the desert.

An example of continuous riparian habitat.

Continuity of riparian vegetation along the
floodplain is extremely important to some species
such as small mammals, reptiles, and amphibians.
Small, discontinuous blocks may not fulfill the needs
of many of these species, causing reductions in num-
bers and, possibly, extirpation of some populations
attempting to use highly fractured riparian habitats
as movement corridors. Amphibians and reptiles that
need riparian habitats are extremely vulnerable to
this fracturing and, as pointed out by Brode and
Bury (1984), what were once continuous populations
of some species in California are now isolated relict
populations. Dispersal routes for pioneering individu-
als and gene exchange have essentially been halted
because once continuous riparian habitats have been
disrupted.

The edge or ecotone component may be of less
concern because the very nature of riparian ecosys-
tems or ribbons of habitat running through uplands
is edge at its maximum. Some believe that riparian
ecosystems are all edge, regardless of alluvial flood-
plain width, because the river constantly meanders
across the first terrace, scouring old plant communi-
ties and depositing new soils that are revegetated

with early stage plant communities. Continuing this
productive, valuable wildlife habitat as edge for wild-
life should be the most important factor in good
management decisions.

Water may or may not be present aboveground
in riparian ecosystems. When present, it is an impor-
tant component for large mammals and, to a lesser
degree, small animals. Most smaller vertebrates gain
adequate moisture in their diet but will drink or
bathe when the opportunity arises. The primary im-
portance of water to terrestrial wildlife, whether
above or below ground, is that it supplies terrestrial
vegetation with the quality and quantity needed
for health and growth.

Birds

As stated earlier, habitat features most important
to wildlife are very difficult to extract from data
sets and virtually impossible to extract and confirm
from 1- to 2-year studies. What may appear to be an
important habitat component one year may not even
be a significant component again in a 5- or 10-year
study. Bird habitat components were derived from
one large river system studied for at least 7 years
with intensive monthly sampling over that time span.
These data have been tested on other large desert
riparian systems, including a revegetation site in
which the plant community contained all of the
most important habitat components except snags for
nest cavities. These were added by erecting nest
boxes. The revegetation site is now about 6 years
old, the cottonwood trees are more than 15 m (50
ft) high, and the area is replete with the avian spe-
cies that it was designed to attract. Unfortunately,
the area is too small (28 ha [70 a.]) to attract and
house many pairs of species that require large areas
for breeding territories — again demonstrating the
importance of island size effect.

Keep in mind that these habitat components
were derived for birds in desert riparian systems. Al-
though similar data in a long-term study are being
collected in quaking aspen stands in Colorado (Win-
ternitz 1973, 1976; Winternitz and Cahn 1983), to
our knowledge these kinds of data do not exist for
birds at higher elevations. Not all of these compo-
nents may be important at higher elevations and
if they are, their order of importance may also differ.
These habitat components should be used with cau-
tion at higher elevations, and where supporting data
are available we have included the references. Again,
other and better references may be available and
should be used.

Frequently, biologists need bird species lists to
illustrate the importance of riparian habitats, to refer-
ence particular species, or for other reasons. A list
of desert riparian species in each desert and their

Riparian Habitats

175

dependency on riparian revegetation species in each
desert is presented by Ohmart and Anderson (1982).
Knopf ( 1985) provided a list of birds observed in a
2-year study along the Front Range in Colorado.

We have found the following habitat compo-
nents, in the order given, most important for entire
avian communities in desert riparian areas:

1. Tree species

a. Fremont cottonwoods (Populus fremontii)
and Goodding willows (Salix gooddingii)

b. Honey mesquite

2. Shrubs

a. Quail bush (Atriplex lentiformes)

b. Iodine bush or inkweed (Suaeda torreyana)

3. Mistletoe (Phoradendron californicum)

4. Foliage density

5. Foliage height diversity (FHD)

6. Snags

7. Patchiness (PI)

Tree Species.

Cottonwoods and Willows. This tree
component was widespread along most permanent
streams and rivers at lower elevations in the western
U.S. before European settlers. Its widespread
distribution and antiquity in western North America
(Axelrod 1958) undoubtedly provided many
opportunities for an evolving insectivorous avifauna.

A lush canopy provided shade, cover, and a myriad
of insects. The rough, ever-sloughing bark attracts
wood-boring larvae plus a number of other
arthropods, which provide forage for bark-gleaning
and trunk-scaling birds. The soft wood is easily
excavated by woodpeckers and, when abandoned,
secondary cavity- nesting species such as Lucy's
warblers (Vermivora luciae), brown-crested
flycatchers (Myiarchus tyrannulus), and elf owls
(Micrathene whitneyi) have an array of vacant
cavities for their nesting activities. Even in winter
after leaf drop, the rough, splitting bark provides
foraging opportunities for numerous bird species. In
early spring the flowers, laden with pollen, are
swarmed by arthropods, and wintering and migrant
birds consume the pollen and insects attracted to
the flowers. As gallery forests of these trees age,
some die providing snags and light penetration to
the forest floor. Shrubs, other trees, and annuals
invade to provide patchiness that attracts other bird
species.

These large (18-24 m [60-80 ft] tall and 2-3 m
[6-10 ft] DBH), branching trees with their attendant
insect fauna, bird life, and proximity to a stream with
its aquatic life, also attract a number of wintering
and breeding raptors. These birds, at the top of the
food chain, find a rich and varied prey base upon
which to feed.

Studies of this community type in California
(Gaines 1977), Arizona (Carothers et al. 1974; An-
derson and Ohmart 1984; Rice et al. 1984), New
Mexico (Hubbard 1971; Hink and Ohmart 1984),
and in Texas (Wauer 1977; Engel-Wilson and
Ohmart 1978) attest to its wildlife value. As this
community type becomes extirpated from the West
Coast eastward, at least 10 bird species will be lost
as well.

sA m

■■■■ 'f*^m'--

m4 8|§p&; ' v "■'■'*

Cottonwoodwillow association.

Woodpeckers can easily excavate cavities in soft-wood
trees, such as the cottonwoods.

176

Riparian Habitats

Deciduous trees at higher elevations may be as
important to birds as cottonwoods and willows are
in desert riparian systems. Bull and Skovlin (1982),
working in Oregon, reported that bird diversity and
species composition changed with the amount of
deciduous vegetation as it ranged from high
(> 40% ), to moderate (15-30% ), to low (< 1% ).
Birds using deciduous vegetation in the area sampled
were the only group highly dependent on this habi-
tat component. A long-term study in quaking aspen
stands in Colorado has documented the value of this
deciduous tree to birds (Winternitz 1976).

A 2-year study by Knopf (1985), at elevations
ranging between 1,200 and 2,750 m (3,973 and
9,022 ft), indicated that breeding birds in riparian
habitats were more simplistically structured to habi-
tat components at higher elevations than at lower
elevations. He reported dramatic changes in species
richness at intermediate elevations, which could
indicate less structuring of avian communities at
these elevations or that his study period was too
short to fully document what was occurring at these
elevations. At elevations above 1,909 m (6,263 ft),
the importance of riparian habitat declined; the up-
lands contained a more diverse avifauna than the
riparian areas. Further testing of these correlations
are needed, but before wildlife biologists de-empha-
size intermediate elevation habitats, we strongly
recommend that the manager fully determine
whether the wildlife values are indeed low and the
potential for maximum riparian productivity on a
site has been achieved.

Possibly, the entire western riparian avifauna has
evolved more closely with deciduous tree species at
various elevations than other habitat components.
These trees provide the essentials for life, e.g., food,
cover, and space for these insectivorous birds that
can also obtain needed water from the riparian sys-
tem. Box elder (Acer negundo), walnut (juglans
major), sycamore, narrow-leaf cottonwood, and oth-
ers may be ecologically equivalent to cottonwood-
willow at lower elevations. The importance of these
trees to wildlife cannot be overstressed if they are
true surrogates; riparian habitats should be managed
to ensure healthy communities with young replace-
ment structural types.

Honey Mesquite. This tree is deciduous in
desert riparian habitats. It grows slowly and is found
primarily on the highest terrace (second terrace)
away from the river where channel cutting by the
river seldom occurs. Again, the flowers attract wild-
life, both by the pollen and the insects feeding on
the flowers. The fruits or beans produced by this
tree are rich in carbohydrates and are consumed by
a wide range of wildlife species.

IS

Honey mesquite.

In low areas where floodwaters are trapped,
colloidal materials are deposited and form heavier
soils that support a mixture of shrub species, primar-
ily quail bush, four-winged saltbush (Atriplex canes-
cens), wolfberry (Lycium sp), and inkweed. These
shrubs provide important values to a number of
wildlife species and enhance the avian productivity
of the honey mesquite community.

Honey mesquite communities are also enhanced
by numerous annual plants that have high seed pro-
duction after a wet summer or winter. Seeds pro-
duced by these annuals are utilized by a large
granivorous guild that includes white-crowned spar-
rows (Zonotrichia leucophyrs), Gambel's quail (Cal-
lipepla gambelii), and Brewer's sparrows (Spizella
breweri).

Riparian Habitats

177

Shrubs.

Quail Bush. In optimum growing conditions,
quail bush reaches heights of3to4m(10to 14ft)
and a single plant could cover 13m ( 1 40 ft ).
Although an evergreen, this shrub drops leaves and
stems that over the years form a thick layer of
composting material. This dense shrub shades the
soil to help hold moisture that in turn expedites
decomposition of the litter accumulation. Ground-
dwelling birds, such as quail, thrashers (Toxostoma
sp.), and towhees (Pipilo sp.), find protection and
productive, insect foraging areas in the decomposing
litter under this shrub. Small insectivorous birds
forage extensively in and among the dense stems and
leaves.

When this shrub is mixed with a community of
trees, such as honey mesquite or salt cedar, it greatly
enhances the wildlife values of the trees. Also, it
forms a monoculture where quail and other ground-
dwelling bird species can attain high densities. Its
seeds are important to these birds as well because
the plant is a prolific seed-producing species and the
seeds fall to the ground throughout the year.

Iodine Bush or Inkweed. This 0.5- to 1-m (2-
to 3-ft) tall shrub is used by numerous ground-dwell-
ing birds, but is most important to sage sparrows
(Amphispiza belli}. Wintering sage sparrows ac-
tively select habitats containing this plant or revege-
tated areas containing inkweed (Meents et al. 1982).
Seeds and plant parts of inkweed were found in sage
sparrow gizzards, but why they actually select habi-
tats with a preponderance of inkweed is unknown.

Mistletoe. Mistletoe is a parasite, widespread on
trees and shrubs in the pea family (Leguminosae or
Fabaceae). Although it can be found on many
different host species, it appears to do best on honey
mesquite trees. The dense clumps provide shelter
for perching birds, nesting cover for breeding
species, and berries for a number of frugivorous
(fruit-eating) birds, especially Phainopepla
{Phainopepla nitens; Anderson and Ohmart 1978).
Although mistletoe may eventually kill some
mesquite trees, its value to wildlife appears to offset
its negative effects to mesquite. In healthy riparian
systems, life and death are integral parts of a produc-
tive community. As mesquite trees die, they produce
hardwood snags. The snags attract wood boring
insects and can become potential nest sites for the
ladder-backed woodpecker {Picoides scalaris).
These snags also provide perches for hunting raptors.

Foliage Density. Until recently, the importance of
foliage density (surface area of leaves and stems/
area ) or foliage height diversity could not be readily
separated. In recent strip-clearing studies, foliage

density was reduced by 20-40% , whereas foliage
height diversity did not change. However, bird life in
the area was drastically reduced (Anderson and
Ohmart unpubl. ms). This seems to indicate that
foliage density is more important in the higher
vertical layers than foliage height diversity, as it
provides better forage substrate for insect-gleaning
birds, concealment from predators, and sites for
nesting. Canopy and mid-canopy layers of vegetation
are low quality to wildlife unless foliage density is
moderate or high. It is generally believed that by
increasing foliage density in each layer of vegetation,
the carrying capacity of insectivorous birds is
increased and new niches are added.

Foliage Height Diversity (FHD). Foliage height
diversity measures how evenly foliage is distributed
among the vertical layers. A plant community with
little or no foliage density in the tree canopy layer
would have a lower FHD value than a similar plant
community with dense foliage at the understory,
midstory, and canopy levels. Thus, foliage height
diversity increases when foliage density values are
nearly equal among all vertical layers of vegetation.

:. ■

>C '■',

Salt cedar community.

Figure 2 shows two communities with differing
height diversities; the one on the left shows low-,
mid-, and upper-canopy volumes while the
community on the right shows some midstory
volume and a little overstory.

This habitat component is unquestionably valu-
able to wildlife, but less so than foliage density.
Other species of trees can have height density values
similar to cottonwood and willow habitat, yet not
have similar species richness values and bird densi-
ties. Rice et al. (1984) examined the importance
of FHD, patchiness (PI), and individual tree species
to avian communities. Response of individual bird

178

Riparian Habitats

ft?

42?

Figure 2. Two plant communities showing different foliage height diversities.

species was significantly higher with greater fre-
quency of individual tree species than with other
vegetation variables depicting structure. Although
height densities correlated well with other tested
habitat components, they were strengthened by add-
ing individual tree species. Generally, it is assumed
that as one adds layers of vegetation, these additional
layers provide niches for additional species of birds.

Snags. Snags are extremely important to a number
of species in riparian ecosystems. Many species use
snags for nesting, feeding, roosting, as hunting and
loafing perches, for hibernacula while overwintering,
and as a moist refuge in dry summer months. Snags
can be, and have been reported as, limiting to the
presence and abundance of cavity-nesting birds
(Haapanen 1965; Balda 1975; Conner et al. 1975;
Evans and Conner 1979; Scott 1979; Mannan et al.
1980; Dickson et al. 1983). A wildlife snag
symposium (Davis et al. 1983) provided a
compilation of papers on this subject and should be
referred to for more specific information.

Many of the above studies and others examining
the importance of snags to birds have been con-
ducted in upland habitats, but snags in riparian habi-
tats are as limiting, if not more so, than those in
upland areas and have the same effect on many pri-
mary and secondary cavity-nesting species. T.E. Mar-
tin (Arizona State University, unpubl. ms) examined
the importance of snags in 1 3 riparian habitats
where the riparian species were primarily big tooth
maple (Acer grandidentatum), quaking aspen, and
New Mexico locust (Robinia neomextcana), with
an understory of golden pea (Thermopsis pineto-
rum) and raspberry (Rubus strigosus). Snags were
primarily aspen and were significantly (P < 0.001)
more abundant along streams on north-facing slopes
than on south-facing slopes. Densities of snag-using
species were greatest in snag-rich habitats, and more
snag-using species were present.

Brush et al. (1983) found that this habitat com-
ponent, especially softwood snags, is generally a

limiting factor in heavily managed desert riparian
systems. Fires, floods, and even removal of snags to
protect water skiers has virtually eliminated this
important component in the lower Colorado River
riparian ecosystem.

Primary and secondary cavity-nesting species are
vulnerable unless management is aware of their
needs and places high value on ensuring large, qual-
ity snags for wildlife. Unfortunately, a prerequisite
to snags is living trees; this handicap has been partly
overcome and artificial snags are being excavated
by some woodpeckers (Grubb et al. 1983; Peterson
and Grubb 1983). Presence or absence of snags
should be noted as riparian habitats and assessed for
wildlife.

Snags provide essential habitat for about 85 bird
species (Scott et al. 1977) that use natural cavities,
excavate their own, or use holes excavated by other
species. Many of these species are obligate riparian
forms. In quaking aspen stands in Colorado, almost
40% of the breeding bird species using this type
used nest holes, and trees containing these cavities
were usually well over 100 years old (Winternitz
and Cahn 1983).

Patchiness. Horizontal foliage diversity or
intraplant community patchiness on the horizontal
scale has long been recognized as valuable to
wildlife. Figure 3 attempts to show patchiness
in a plant community; the open areas between
interlocking trees and some midcanopy vegetation
produce a patchy effect in this forested habitat.
Patchiness, like foliage height diversity, is thought to
create additional niches for birds. Some species are
found primarily in continuous riparian forests,
whereas others are attracted to openings in the
canopy where lesser trees or shrubs provide patches
in an otherwise continuous canopy. These patches
offer habitat space for new species that would
otherwise be absent, because as patchiness increases
so does species richness. Anderson et al. (1983)
recently developed a method to quantify patchiness.

Riparian Habitats

179

1^ VP \W

Jk, k-^Ok^ JWii'a*

Figure 3. Patchiness in a plant community
(diagrammatic).

Mammals

Large Mammals. The total importance of riparian
habitats to large mammals is not fully understood,
but undoubtedly these ecosystems provide three
essential resources to this group: food, water, and
cover. Dealy et al. (1981) pointed out that riparian
habitats in general, and quaking aspens in particular,
provide both thermal cover and forage for ungulates.

In the Modoc National Forest in northern Cali-
fornia, Salwasser and Shimamoto (1984) examined
animal use of wetlands that have developed since the
1920s behind water storage dams. They examined
wetland use for three classes of large mammals: feral
horses, domestic cattle, and pronghorn antelope
(Antilocapra amertcana). Although these wetlands
made up 10% of the available habitats, they were
used heavily by all classes from midsummer through
fall. Pronghorn antelope use peaked at 80% in Au-
gust and remained over 40% into October. Feral
horse use peaked at 78% in July and remained over
40% into early August. Cattle use peaked at 80%
in August and exceeded 50% since late June. Annual
use varied by the three classes, but all used the de-
veloped wetlands far more extensively than the
wetlands that were available among habitats in mid-
summer to fall.

Studies examining habitat use by elk in mixed-
forest types have shown that of 1 5 associated habitat
types, riparian zones were most heavily used (Peder-
son et al. 1979). Riparian corridors also served as
travel routes between areas.

Although elk will graze areas used by cattle,
domestic livestock grazing appears to diminish use of
the area by elk (Blood 1966; Nelson and Burnell
1975; Burbridge and Neff 1976; Skovlin et al. 1976).
Diminished elk use of areas where cattle have grazed
is probably related to season of use and grazing in-
tensity by cattle before or during the time of elk use.
Fall and winter use by elk declined significantly fol-
lowing cattle use in late spring and summer on pine-
bunchgrass uplands near riparian zones in the Blue
Mountains of Oregon (Skovlin 1984).

How much vegetation is removed in riparian
habitats by big game is not well studied, but Skovlin
(1984) indicated that in northeastern Oregon, deer
and elk accounted for about one-third of the total
browsing on riparian trees and shrubs. Cattle appar-
ently removed the remaining two-thirds. He also
reported that studies conducted near the above sites,
where deferred or seasonal livestock grazing has
been practiced, showed similar vegetation removal
rates for big game and livestock.

The extent to which elk (Cervus elaphiis) use
and depend on riparian habitats, other than as travel
routes, in the nonsummer months is not well known.
However, in summer in western Montana, 80% of
elk use in July was within 0.4 km (0.25 mi.) of per-
manent water (Marcum 1975). Skovlin (1984) re-
ported heavy preference for summer range within
0.8 km (0.5 mi.) from water. Whether the elk are at-
tracted to riparian habitats for the green lush forage,
water, or both is unclear; perhaps only lactating
cows depend on free water sources at this time
(Marcum 1975; Black et al. 1976; Thomas et al.
1976).

Soil compaction from grazing by wild ungulates
is normally not a problem in riparian zones in fall
and winter, but some compaction damage of satu-
rated or wet soils in the upland range in early spring
may occur (Skovlin 1984). He also reported that
shrub use by elk in early spring and summer, before
grazing by domestic livestock, can significantly affect
shrub survival during the ensuing growing season.
Cattle browsing on the same shrubs grazed by elk
could add to the shrub mortality problem.

Medium-Sized Mammals. Numerous medium-sized
species of mammals are either obligate or facultative

180

Riparian Habitats

users of riparian systems at all elevations. For
example, in Big Bend National Park in west Texas,
1 5 medium-sized species were reported using
riparian ecosystems and one, the beaver {Castor
canadensis), was a true obligate (Boeer and
Schmidly 1977). Williams and Kilburn (1984)
reported that of the 502 recent native species and
subspecies of land mammals in California, about 25%
(133 taxa) were limited to or largely depended on
riparian and other types of wetland communities.
They further stated that "No other general type of
mammalian habitat in California approaches riparian
and other wetland communities in importance to

mammals "Of the 1 5 species they listed, 5 were

in the medium-sized category. Similarly, in Mexico,
seven medium-sized mammals, threatened with
extinction, are confined to riparian ecosystems
(Ceballos 1985).

Lack of any systematic and long-term studies of
medium-sized mammals makes it difficult for man-
agers to establish specific habitat management
criteria that would ensure these species are not ex-
tirpated by water management, timber harvest, or
grazing practices. As in most past management deci-
sions, the best approach is to review natural history
studies of these species and avoid eliminating habitat
components that have been identified as important.

Many of these species are aquatic or semi-
aquatic and feed on plant and animal matter in or
along the stream. Consequently, instream flows and
water quality represent important habitat factors.
Streamside vegetation also provides food and cover
for these species, so severe vegetation removal by
domestic livestock can be important. Therefore, to
provide optimum conditions for these species, one
must be cognizant of the specific needs for each
species found in the area.

In California, five medium-sized mammal species
considered obligate to riparian habitats have been
recommended for special consideration (Williams
and Kilburn 1984). One, the snowshoe hare (Lepus
americanus), occurs in dense thickets of alders (Al-
nus sp. ), willows, and other shrubs in the Sierra
Nevada range (Orr 1940). Dense thickets of young
conifers also serve as cover for this hare (Williams
1985). This species may not depend on riparian
habitat in other portions of its range, which may not
be unusual for some high-elevation forms.

The mountain beaver (Aplodontia rufa), also
found along moist, forested habitats along the Pacific
Slope, occurs in small colonies. It favors moist
slopes, supporting lush growths of forbs, and often
excavates its burrows next to a stream. Its diet in-
cludes a variety of forbs and the buds, twigs, and
bark of willow and dogwood (Cornus sp.). It also
cuts forbs to pile as hay (Grinnell and Storer 1924).

A number of medium-sized mammals use ripar-
ian ecosystems extensively although they occur in
the uplands. Densities are lower in upland habitats
than along riparian systems, as noted by studies of
the gray fox (Urocyon cinereoargenteus) along
Putah Creek near Davis, California (Hallbery and
Trapp 1984 ). Two males and females with radio col-
lars spent 76% of their nocturnal and 92% of their
diurnal time in riparian habitats and the remainder in
agricultural habitats. An example of the latter is ex-
emplified in ringtails {Bassariscus astutus ), where
this species' densities range from 2.5 to 5 times
greater in riparian ecosystems as opposed to upland
habitats (Belluomini and Trapp 1984). Numerous
other facultative riparian species obviously fit into
these categories, stressing the importance of these
habitats to medium-sized mammals.

The fisher (Martes pennanti) is poorly studied
especially in the western U.S., but in general, is
thought to depend on riparian habitats for travel and
escape routes (de Vos 1951, 1952; Kelly 1977; Buck
1982; Mullis 1985). Of the studies conducted thus
far, only Kelly (1977) has reported that the long axis
of most home ranges in New Hampshire have tended
to parallel drainages. Regardless, these riparian habi-
tats are believed to be very important to this species.

In the West, fisher populations have been
thought to be declining because of activities such as
timber harvesting, road building, and overtrapping.
In California the trapping season was closed in 1946,
and in recent years these populations appear to have
increased (Mullis 1985). Indications are that the
marten (M. americana) may similarly need riparian
habitats.

Beavers, once a dominant aquatic mammal in
riparian systems, have virtually been eliminated in
western streams through trapping, shooting, instream
flow reductions, and other factors. The beaver needs
streams 1 m (3 ft) or more deep, and timber for
food, lodge, and dam construction. Its propensity to
construct dams has frequently made it an undesira-
ble riverine species. Although primarily preferring to
live in lodges, the beaver will excavate dens in the
riverbank (Nelson and Hooper 1976). Its damming
activities frequently have flooded agricultural lands,
forest lands, and damaged irrigation systems.

Parker et al. ( 1985) recently suggested that
beavers might have played an important role in re-
sisting minor perturbations in lower-order streams.
Their model and arguments are not too difficult to
visualize when the impact this species can have on
streams and riparian trees is considered. Lang and
Weider (1984) suggested that beavers have altered
the structure of forests in West Virginia. Platts et
al. (1985) supported the beaver model for large
storm events. However, the Parker et al. (1985)

Riparian Habitats

181

model needs further testing before it can be used as
a possible management tool in resisting mild
perturbations.

Scott (1984) justified the importance of beaver,
mink (Mustela vison), and muskrat (Ondatra zibeth-
icus) by examining the dollar return of these species
in California. In some instances, the fur-return value
of these species may represent up to 71% of an indi-
vidual's annual income. In economically depressed
communities this could be an important source of
income.

Small Mammals. Soil texture, structure, and
moisture seem to be important in habitat selection
of many small mammals that burrow. Other rodents
respond to riparian habitats in varying manners:
sciurids respond more to tree species and height of
trees; heteromyids, more to soil and open habitats;
castorids, to water and forage availability; cricetids,
to vegetation density and structure; and zapodidae,
possibly to grass height and stem densities in moist
soils.

Although identifying habitat components for all
small mammals in riparian communities is far from
complete, there is good evidence to demonstrate the
importance of these habitats to small mammals.
Cross (1985) found that riparian habitats in south-
western Oregon, composed of mixed conifer and
deciduous broadleaf trees, invariably had greater spe-
cies richness and total small mammal biomass than
upland sites. Studies (Stamp and Ohmart 1979) in
riparian habitats in the Sonoran Desert of Arizona
show similar results.

Herptofauna

To our knowledge, there are no in-depth, long-
term data from a riparian ecosystem at any elevation
dealing with habitat factors important to an entire
herptofauna. This is a serious omission in riparian
studies in that reptiles and amphibians are probably
as important, and possibly more so, than birds and
mammals in energy flow and nutrient cycling. There
are a number of autecological studies on specific
species, however, which we used to delineate some
of the most important habitat features for amphibians
and reptiles. Until in-depth, long-term community
ecology studies are conducted, we will be making
many assumptions.

In the Hubbard Brook Experimental Forest in
New Hampshire, Burton and Likens (1975) esti-
mated salamander densities of 2,950/ha ( 1,180/a.),
which exceeded density estimates of birds and mam-
mals. Salamander biomass exceeded that of birds by
216 times and approximated that of small mammals.
Similar densities of Siskiyou Mountain salamanders
(Plethodon stormi) have been estimated at densities

of 2,700/ha (1,080/a.; Nussbaum 1974) in optimum
habitats in Oregon; and Murphy and Hall (1981)
reported that in Oregon, the Pacific giant salamander
(Dicamptodon eusatus) may make up as much as
99% of the total predator biomass in some streams.

Reptile densities are equally high in the West.
The western pond turtle (Clemmys marmorata) has
been reported at densities of 425/ha ( 170/a.; Bury
1979) and Sonoran mud turtles (Kinosternon sono-
riense) in Arizona at 825/ha (330/a.; Hulse 1974).
Southern ringnecked snakes (Diadophis punctatus)
have been reported at densities of 1,000-1,500/ha
(400-600/a.; Sullivan 1981).

In general, amphibians as a species group are
more dependent on riparian ecosystems than are
reptiles. Much of this revolves around their evolu-
tionary life-style; they are aquatic or semi-aquatic
and lay open eggs (nonshelled) in water or very
moist areas. For aquatic species, many of the same
needs of fishes would apply to these forms, e.g.,
shaded stream for cooler water and higher oxygen
levels, productive stream bottoms, and escape cover.
For more terrestrial forms, such as some salaman-
ders, the presence of rotting logs and dense ground
cover (litter or vegetation) is essential.

Because most species in this group are insecti-
vorous and carnivorous, and most spend their active
periods in shallow water, on the ground, or in trees,
any action that reduces densities of trees, shrubs,
or other vegetation and reduces high insect biomass
has to be considered negative. On the Colorado
River, actions that help birds were also considered
to help this group, especially reptiles. This is proba-
bly not totally true, but revegetation sites that are
primarily designed for birds and small mammals also
support high populations of reptiles — higher than
those found in most natural communities. Because
many biologists believe that birds are nothing more
than glorified reptiles, these results should not be
surprising.

- - - — *»-»v.

182

Riparian Habitats

For some amphibians, the maintenance of back-
waters with dense-to-moderate emergent stands of
vegetation is essential. There should be stable levels
of instream flow with good bank development stabi-
lized with vegetation. Large, decaying logs in the
floodplain and in the adjacent uplands are necessary
habitat components for numerous amphibians, such
as salamanders.

Many herpetile species either overwinter in the
soil of the floodplain or in decaying wood under
logs. Others overwinter in hibernacula in downed,
decaying logs; snags; and even live trees with natural
cavities or heart rot. These habitat features are im-
portant to amphibians for their overwinter survival.
Loss of one or more of these critical habitat features
can break the annual life cycle of a species, resulting
in its local extirpation.

Another important feature of riparian habitats to
amphibians and reptiles, frequently overlooked by
managers, is the use of these systems as corridors for
dispersal and genetic continuity between popula-
tions. Brode and Bury (1984) stressed the impor-
tance of continuous riparian ecosystems to help
maintain genetic heterogeneity, and noted that habi-
tat disruption has resulted in isolated populations
of many species in California riparian habitats.

Although all the habitat components needed for
species survival may be present along a stream, a
population or species can be extirpated for a num-
ber of reasons: ( 1 ) their density is so low that an
adequate number of matings do not occur to sustain
population levels; (2) genetic drift; (3) genetic het-
erozygosity is not adequate to sustain a healthy pop-
ulation; and (4) some density-independent mortality
factor eliminates the population, e.g., severe and
prolonged freezing or drought. For this group, habi-
tat continuity and patch size should be considered
when making management decisions.

Habitat-Wildlife Correlations

A few years ago, MacArthur and MacArthur
( 1961 ) reported a strong correlation between east-
ern forest birds and FHD. Over the years, ecologists
and managers have looked for one or more habitat
features that highly correlated with wildlife. Bird
species diversity (BSD) is identical in basic concept
to FHD, but is based on the number of bird species
(species richness) in the plant community and how
evenly distributed the density of each species is
among the total number of species. An avian commu-
nity containing 1 5 species, where 2 or 3 make up
90% of the total birds, would have low BSD values
compared with an avian community where the total
bird numbers were more evenly distributed among
the 15 species.

The BSD and FHD values and their relationships
have not been well understood. For a brief period
it was thought that the key to good management was
to manage for high BSD values. The weakness of
this approach becomes fairly obvious in the follow-
ing example. In most habitat situations the density of
rare or federally and state-listed species is low. Con-
sequently, in the earlier example of 1 5 bird species
in a community, the 2 or 3 species in very low den-
sity could be lost and hardly change the BSD value.
Therefore, one can see the fallacy of trying to man-
age habitats or ecosystems based on BSD values.
Keep in mind that BSD and species richness (num-
ber of species present ) are very different terms and
should not be used interchangeably.

Since high BSD-FHD relationships were reported
by MacArthur and MacArthur ( 1961 ) in eastern for-
ests, many studies have supported or refuted this
relationship throughout North America. Many scien-
tists collecting data have not looked at other habitat
components and have been satisfied with relatively
good correlations between BSD and FHD. A few
other studies looked at important vegetation fea-
tures, but these research results suffer from one or
more of the following: lack of tight experimental
design; limited data collection, such as only during
the breeding season; and short-term studies, con-
ducted for only 1 or 2 years. These are central crite-
ria in judging the value of field data as applied to
management situations. Much money, time, and ef-
fort have been and can be lost when management
decisions are based on poorly conducted studies.

Many federal and state agencies have tried to re-
duce ecosystems to one or a few variables for man-
agement purposes. Our long-term research, plus that
of others, argues persuasively that ecosystems, plant
communities, and even small habitats are too com-
plex to reduce to one or a few numbers for manage-
ment purposes. We know of no shortcut approaches
to good wildlife and habitat management and seri-
ously doubt any will be found. Unless managing
agencies discover this and proceed with the task at
hand, which is the collection of in-depth, long-term
changes in plant and animal communities, we as
biologists will always be playing "catch-up biology"
or patching data together along with our "gut feel-
ings" to make important management decisions and
recommendations.

DATA COLLECTION PRIORITIES

Biologists are so frequently enthusiastic about
wildlife that when presented with the opportunity to
study an area, they immediately begin thinking of
ways to census various classes of vertebrates, study
life histories of animals, or collect fauna. Although
important, highly desirable, and needed, these are
probably the last steps that should be undertaken.

Riparian Habitats

183

We have developed a list of priorities that we be-
lieve will help the land manager develop a data base
on the riparian ecosystem and also document pro-
cesses occurring or about to occur in the ecosystem.
Since the limiting factor in any study is money,
which is people, vehicles, equipment, or some other
factors requiring funding, our list assumes that funds
will be limited.

Developing Vegetation Maps

Develop a fine-grained map of the riparian vege-
tation ecosystem(s) under management considera-
tion. By fine-grained, we mean a map depicting and
naming the plant communities at a resolution of
possibly 0.2 ha (0.5 a.) or less, if necessary, to show
discrete plant communities. Common sense must
prevail because overzealous type mapping could
reduce a community to one tree, then to a limb, and
to a leaf. If two tree species occur together along a
stream or river, the community might be delineated
as a cottonwood-willow association. Or if a tree wil-
low grows in one area and a shrub willow grows
in another, you would have Goodding willow (Salix
gooddingii) as a separate community from coyote
willow (S. exigua) and so forth. Because riparian
systems are linear, and frequently vary in width from
a few yards to many miles, maps must be scaled
upward to include small communities that may sup-
port endangered wildlife species or very highly spe-
cialized species.

If numerous riparian ecosystems are under man-
agement consideration and funding only allows half
of them to be adequately mapped or all to be super-
ficially mapped, then we recommend prioritizing
the streams and intensively mapping only half of
them. Otherwise, a poor job on all of them provides
little, if any, good information from which to make
crucial management decisions and will probably
be meaningless to future managers attempting to in-
terpret and use the maps.

Good vegetation type maps should be accom-
panied by the criteria used in naming each major
community. For example, if you elect to name cot-
tonwood-willow associations based on cottonwoods
being abundant to scattered in the community, this
information should accompany the maps. Or you
may elect to name the community as such if about
50% of the trees are cottonwoods and the other half
willows. The important point is that good quantita-
tive criteria should be established, adhered to in
mapping, and always accompany the maps as legends
or as an appendix. These maps will be invaluable to
future managers as they attempt to assess and under-
stand habitat changes. The probability is high that
you will be transferred in a few years and the person
replacing you must be able to interpret the informa-
tion that you collected while working on riparian
systems.

^>Kr,-.%v-

mmmm

WWkMm™

A large scale airphoto (1:2,400) provides a basis for map -
ping (B) brush, (H) herbaceous vegetation, and (X) bare
ground, and (T) trees where present.

Classifying Plant Community and Structural
Types

Quantify vegetation structure and tree species
composition in a subset of available plant com-
munities along the stream under management con-
sideration. These data are important in naming
communities and assessing structural configurations
of vegetation. As riparian ecosystems are perturbed
they tend to support less foliage volume at higher
layers. For example, fires or floods tend to remove
trees, reducing canopy foliage volume. Incoming
shrubs or trees increase foliage volume at the herba-
ceous and shrub layers. Also, as man perturbs these
habitats through reduction of instream flow or in-
creased silt concentrations, the perturbations be-
come manifested through vegetation structure as
reduced foliage volumes in higher layers of vegeta-
tion or the tree canopy.

In our work with community structural types,
we have used structural Type I as the mature, most
structurally diverse community, and Type VI as the
early stage of returning vegetation. A healthy riparian
ecosystem will contain mature plant communities
and structural types, intermediate types, and Type VI
communities that will eventually replace old, deca-
dent stands. In heavily grazed ecosystems, such as

184

Riparian Habitats

those along the Salt and Verde rivers in central Ari-
zona, replacement cottonwood communities have
been virtually eliminated and only mature, decadent
stands remain. No replacement communities are
forthcoming and remedial efforts, such as planting of
cuttings, have been attempted to provide replace-
ment cottonwood stands. Although riparian ecosys-
tems can be revegetated more economically through
natural flooding, federal agencies are supplementing
regrowth through fencing and revegetation efforts to
ensure riparian habitats are not lost forever.

Censusing Vertebrates

Establish census lines for birds, snap-trap small
mammals (make voucher specimens), and install can-
trap grids for reptiles and amphibians. Bird transects
must be censused no less than nine times per plant
community and structural type per season. Number
of census transects for birds should be in about
equal proportion to the aerial extent of the commu-
nity and structural types. Small-mammal trapping and
can traps should also follow the same protocol.

Original type maps can be modified to show
new plant communities and structural types. Your
data on classification and criteria depicting structural
types should be archived for subsequent use in the
area.

Data should be collected for at least 2 years; 3
years is better. Seasonal and annual variation in ani-
mal numbers should be determined before realistic
densities and habitat affinities can be made.

Determining Vertebrate Species Richness
and Relative Abundance (Breeding)

Determine vertebrate species richness and spe-
cies composition for each plant community and
structural type during the breeding season, and year-
round when possible. Relative densities determined
by a gradation of rare-to-common for each bird spe-
cies is adequate. Small mammals can be assessed
through snap trapping for three consecutive nights
and expressing each species' relative values as X
number of each species caught/number of trap nights
(cumulative total). If small mammal trapping is un-
dertaken, voucher specimens must be prepared and
housed in a specimen collection where they are
properly cared for. If certain community types are
important to large mammals for feeding, thermal
cover, or breeding grounds, these should be noted.

Attempt to determine relative densities of am-
phibians and reptiles through trap-and-release with
can traps and observations. Values can be based on
can-trap days, species and numbers observed per
hour, or some other relative index.

We believe if managers concentrated on priori-
ties 1 through 3, they would be in a strong position
to defend management decisions relating to domes-
tic livestock grazing, presence or absence of endan-
gered species, wildlife values of community and
structural types, or wildlife values of areas behind
proposed dams. As it is, managers often do not have
enough data to demonstrate the high wildlife value
of riparian ecosystems that are the most productive
wildlife habitats in western North America.

EFFECTS OF LAND MANAGEMENT
ACTIVITIES ON RIPARIAN SYSTEMS

Under the multiple-use concept, managers must
be aware of impacts or potential impacts of ongoing
activities on the riparian resource. Activities of recre-
ationists, once considered harmless, now could be
shown to have some profound effects on some verte-
brate groups and plant species in the riparian zone.
As with other activities, such as domestic livestock
grazing and impoundments, we know many impacts
can be mitigated to avoid or reduce some of the
damage to riparian systems. We hope some of the
activities covered in this section will help you avoid
some of the major management problems or provide
solutions to existing problems.

Determining Vertebrate Species Richness
and Relative Abundance (Winter)

Determine vertebrate species richness and spe-
cies composition for each plant community and
structural type in winter. These values will vary from
winter to winter depending on the severity of the
winter and the previous breeding success of the
overwintering species. Rodents may hibernate and
large mammals may move to lower elevations.

Domestic Livestock Grazing

Livestock grazing is a historic use on public
lands. Overgrazing has been a problem in many areas
of the West, however, and overgrazing in upland
areas has caused the removal of virtually all nontoxic
forage in adjacent floodplains. Reid and Pickford
(1946) reported that cattle congregate and utilize
riparian forage much more intensively than the vege-
tation of adjacent ranges. Biologists must be aware

Riparian Habitats

185

that the riparian ecosystem can potentially be much
more productive for wildlife under better grazing
strategies. Although other land-management activities
have imposed losses or serious reductions in produc-
tivity of wildlife habitat, livestock grazing may be
the major factor negatively affecting wildlife in the
1 1 western states (Oregon- Washington Interagency
Wildlife Council 1978). We agree with Skovlin et al.
( 1977) that because of the highly limited area of
riparian habitats in planning large pasture grazing ap-
proaches, riparian systems are sacrificed on areas of
domestic livestock grazing on most state and federal
land today.

The value of these riparian systems to domestic
livestock grazing (essentially the permittee) can only
be appreciated through forage production values.
Further, the difficulty in bringing about proper stock-
ing levels is deeply embedded in historical and polit-
ical realms. There are 1.62 million ha (4 million a.)
of mountain meadows at elevations between 1 ,890
and 150 m (6,200 and 6,400 ft) in the 11 western
states (U.S. Department of Agriculture, Forest Service
1972). These moist meadows support more beef
per acre than any other range type (Skovlin 1984).
In northeast Oregon, these meadows are so produc-
tive that 1 ha (2.5 a.) is equal in forage production
to 10 to 15 ha (25 to 38 a.) of forested rangeland
(Reid and Pickford 1946). Although only 1 to 2% of
the summer range is meadow in the Pacific North-
west, it potentially produces 20% of the forage (Reid
and Pickford 1946; Roath and Krueger 1982). Be-
cause of livestock concentrations, topographical con-
straints on livestock, and distribution of water, the
forage in the riparian zone accounted for 81% of
that removed by livestock in the Blue Mountain graz-
ing allotment (Roath and Krueger 1982).

Cattle are probably attracted to the riparian
ecosystem for the same reasons as other large mam-
mals (Ames 1977; Severson and Boldt 1978). Many
riparian plant species remain green and succulent
longer than upland vegetation, and sedges (Carex
sp.) contain higher sustained protein and energy
content than important upland plant species (Mc-
Lean et al. 1963; Skovlin 1967; Paulsen 1969).

Two alternatives that may hold promise are
changing season of forage use and changing kinds
and classes of livestock. Siekert et al. (1985) re-
ported that spring grazing showed no significant
changes in channel morphology, whereas summer
and fall grazing did. The level of grazing (30 cow-calf
pairs on 48 ha [120 a.] for 10 days) and duration
are probably unrealistic for most permittees. Further,
in many riparian ecosystems, most if not all, tree
seedlings would be eliminated under this approach.
Marlow and Pogacnik (1985) reported that grazing
of a riparian system in Montana after the stream-
banks had dried (< 10% soil moisture) protected
the stream channel from damage. They recom-
mended fencing riparian habitat, rest-rotation, light
grazing (20% forage removal), and grazing after
streambanks have dried to 10% moisture. Fencing
has been ruled out by Skovlin ( 1984) in that it is too
expensive regardless of ownership.

Stocking moderately with steers or trying differ-
ent breeds might reduce damage to riparian ecosys-
tems caused by cow-calf operations. Lactating cows
with calves appear to concentrate in areas with
green forage and water, whereas steers might range
more into the uplands and not concentrate in ripar-
ian habitats. Neither of the above approaches will
work unless stocking rates are reduced as well. Most
examined riparian systems and their watersheds

Platts (1984) listed eight major approaches
managers should consider in managing riparian-
stream habitats under multiple-use systems. These
range from eliminating grazing permanently or until
after recovery to rehabilitating through revegetation
and artificial stream structures. Most of these ap-
proaches are unacceptable to permittees today but
may seem more reasonable with time or as public
concern forces political action.

Cowand-calf herd in riparian zone.

186

Riparian Habitats

L

were unquestionably overgrazed, and use of any
management approach without AUM reductions ap-
pears to be a waste of time and money. Streams in
overgrazed pastures are characterized by being
wider and shallower; they contain more fine sedi-
ment and have more unstable banks, less bank un-
dercut, and higher summer water temperatures
(Marcuson 1977; Van Velson 1979; Platts 1979,
1984).

Large storm events and the response of riparian
habitats have been of interest to biologists because
of the effects of those events on riparian ecosystems
(Gregory and Madew 1982; Lyons and Beschta
1983). An interesting data set has been reported for
three streams under study in Nevada and Utah where
historically the watersheds of these streams have
been heavily grazed (Platts and Nelson 1983; Platts et
al. 1983). Peak flows (2-14 times normal) occurred
in Gance Creek in 1983 and 1984. Big Creek and
Chimney Creek were not grazed, but Big Creek emp-
ties into Bear Creek. In 1983, flows of 3,630 cfs ex-
ceeded all recorded flows for the past 40-year
records (Platts et al. 1985).

Chimney Creek had received heavy grazing until
1981 and showed little bank development. It was
rested during 1982 and 1983, and the banks were
developing some overhanging vegetation (Platts et
al. 1985). Large decomposing aspen logs in the
stream and on the banks were evidence of past for-
ests that once lined the banks of Chimney Creek.
Aspen forest return has probably been prevented by
blowdown, beaver, and heavy grazing of sprouts by
livestock (Platts et al. 1985). The severe floods in
1983 and 1984 straightened the meandering channel
and widened the streambank, reducing the develop-
ing bank overhanging vegetation. The large aspen
logs that helped hold the stream previously were de-
composed or flushed from the stream by the high
discharges.

The riparian vegetation along Gance Creek was
dominated by large trees from the floodplain to the
streambanks. Flood damage was previously vertical
cutting and some lateral movement of the channel
(Platts et al. 1985). Damage was minimal and the
authors believed that had beavers still controlled the
stream as they had done in the 1950s and 1960s,
the flood damage would have been lessened.

Changes that occurred on Big Creek were most
interesting because a portion of the study stream had
been rested for about 10 years and was comparable
to two other reaches (one above and one below the
rested site) that had received normal grazing pres-
sures. The rested section had dramatically recovered
and showed good floodplain vegetation and stream-
bank development. During the flood years, stream
width in the grazed portions increased by 40% with
extensive lateral stream movement and redeposition

of bedload sediments. The rested section with im-
proved streambanks was able to contain the high
flows and showed only a slight increase in channel
width (Platts et al. 1985). Floodplain vegetation and
soils were altered dramatically in the grazed sections
following the storm events, whereas the rested sec-
tion showed little evidence of vegetation change
or newly eroded sediments. Results from this study
amplify the admonitions of Heede ( 1985) that man-
agers should understand the interrelationships be-
tween vegetation and hydrologic processes in
riparian ecosystems before attempting any type of
management change that alters these natural systems.

The economic values of healthy riparian ecosys-
tems and their attendant wildlife are difficult to es-
tablish (Everest 1977), but approaches have been
made based solely on fisheries. Olson and Armour
(1979) suggested that a hypothetical reach of 14,484
km (9,000 mi.) of depleted fishable streams on U.S.
Bureau of Land Management land be set aside exclu-
sively for recreation. Based totally on increased fish-
ery visitor days due to restored habitat, they
estimated a first-year benefit-cost ratio of 1:1.66. Or,
for every $1.00 spent to fence the riparian corridor,
there would be $1.66 generated by fishermen. Other
values such as backpacking, camping, bird watching,
erosion control, and improved water quality were
not included in this economic return value.

The above approach needs further study and
confirmation because, if true, the cost of fencing
could be replaced by the economic return, making it
a valid alternative. If fencing is not economical
(Skovlin 1984), then it is entirely possible that only
light or no livestock grazing of riparian systems is
the ultimate answer if maximum wildlife productiv-
ity is the management goal. However, Armour
(1977) quoted Harmay as stating:

Vegetation in meadows and drainages is
closely utilized (by domestic livestock) under
any stocking rate or system of grazing Reduc-
ing the livestock or adjusting grazing season
usually will not solve the problem

Findings by Severson and Boldt (1978) support
Harmay in that in the northern Great Plains the ri-
parian habitats were excessively used, regardless of
stocking rates. In eastern Oregon, Gillen ( 1981 )
reported that at a moderate rate, the meadows only
produced 3 to 16 times more forage, again support-
ing the contention that regardless of stocking rate
the riparian habitats will receive the greatest grazing
pressures. Under continuous grazing (60-100% utili-
zation) in Nevada and Utah, there was no evidence
of riparian improvement and under present stocking
rates with continuous grazing, the riparian-stream
ecosystem continually deteriorated (Platts 1984).

Riparian Habitats

187

If not limited by factors such as high salinity and
no flooding, riparian habitat can recover following
heavy grazing (Davis 1977; Glinski 1977; Crouch
1978). Eliminating grazing for 10-12 years may be
necessary at higher elevations (2,650 m [4,800 ft])
in willow communities where grazing pressure was
heavy and chronic (Knopf and Cannon 1981). Knopf
and Cannon ( 1981 ) further pointed out that it is
more difficult to improve a damaged riparian ecosys-
tem by eliminating grazing than to maintain good
conditions in one that is being grazed.

A number of studies have shown a dramatic
increase in wildlife values where riparian systems
were abused by domestic livestock grazing; the areas
were fenced and monitored a number of years after
domestic livestock removal. Numbers of small mam-
mals, songbirds, and raptors increased by 350% (Wi-
negar 1977; Duff 1979; Van Velson 1979) in an area
fenced for 8 years after grazing. Game animals such
as ring-necked pheasants (Phasianus colchicus),
deer, and waterfowl increased as well (Van Velson
1979). On the South Platte River in Colorado,
Crouch (1982) found more ducks, upland game, and
twice as many terrestrial birds in areas fenced for 7
years compared with adjacent grazed habitats. Signifi-
cant differences in bird species richness and foraging
guilds have been reported between heavily grazed
2.5 cow-calf units/ha ( 1/a.) and lightly grazed (0.3
cow-calf units/ha (0.75/a.) riparian habitats. Total
density was not significantly different indicating in-
creases in some species that were already present
and the extirpation of some species such as flycatch-
ers, ground-foraging thrushes, and foliage-gleaning
insectivores.

Small mammals are also adversely affected by
domestic livestock grazing in riparian communities.
Small mammal densities before and after grazing with
a stocking rate of 5 to 6.25 a./AUM (2.0 to 2.5 ha/
AUM) declined from 320 to 33/a. (800 to 83/ha) in a
Douglas hawthorn {Crataegus douglasii >dominated
community, from 180 to 24/a. (450 to 60/ha) in a
riparian meadow, and from 52 to 17/a. (129 to 42/
ha) in a black cottonwood (Populus trichocarpa)-
mixed conifer community. Ten months after grazing
ceased, no significant difference was found between
the small mammal densities in the grazed versus
ungrazed plots.

Some grazing investigators have reported in-
creased rodent species richness under moderate or
heavy grazing pressures (Moulton 1978). We do not
doubt some of these results but point out that small
mammal species that are added or increase in num-
bers are usually habitat generalists whose habitat
requirements are broad. Habitat specialists, such as
many microtine rodents, are usually reduced or elim-
inated when grazing pressures are high. Under these
grazing conditions, species in the genus Peromyscus

and Perognathus may increase or be added because
the former are generalists and the latter require
more open habitat.

Moulton (1978) reported that moderate grazing
(2.3 a./AUM [0.9 ha/AUM]) for 6 months in a cotton-
wood riparian system reduced prairie vole (Microtus
ochrogaster) numbers and increased deer mouse
(Peromyscus maniculatus) numbers. Eight species
of small mammals were trapped in the grazed plot
and four in the ungrazed control. The control site
had been ungrazed for 11+ years, and the vegeta-
tion had moved toward a uniform, dense grass struc-
ture, unsuitable for a number of rodent species. Light
or moderate grazing would have altered plant struc-
ture and species composition, making the habitat
suitable to other rodent species.

Where grazing can be controlled in riparian
habitats and seasonally light-to-moderate forage re-
moval is practiced, the impact can be small to ripar-
ian vegetation and wildlife. But, as pointed out, by
incorporating riparian areas into large pastures, these
productive wildlife habitats become sacrifice areas
where most, if not all, of the annual plant production
is removed. As suggested by May and Davis (1982),
riparian habitats should be separated and managed as
distinct units.

Preliminary data on riparian populations of the
wandering garter snake (Thamnophis elegans va-
grans) on fenced (1972 and 1975) and unfenced
plots on the Rio de los Vacos near Santa Fe, New
Mexico, provided some interesting results (Szaro et
al. 1985). The fenced plots (only cattle excluded)
supported a stand ( 18% ) of trees and shrubs com-
posed of thin-leaf alder (Alnus tenuifolia), irrorata
willow (5. irrorata), Scouler willow (S. scouleriana),
coyote willow, and Mexican cliff-rose (Cowania
mexicana). In contrast, 0.1% of the grazed plots
supported a mixture of thin-leaf alder and irrorata
willow. Herbaceous ground cover (71% versus
88% ) and down and dead debris (0.4% versus 5% )
was significantly (P < 0.05) different in grazed ver-
sus ungrazed plots, respectively. Streamside shrubs in
the fenced plots filtered out debris during floods to
form debris piles up to 4 m ( 12 ft) in diameter and
2 m (6 ft) high. These decomposing piles supported
numerous worms and slugs that made up 62% and
18% of snakes' diet. Snakes were five times more
abundant in the ungrazed versus the grazed plots,
even though they were more difficult to find in the
vegetation and debris. Other species of reptiles are
undoubtedly affected by domestic livestock grazing
as foliage for insects and cover for the reptiles are
reduced or eliminated.

Impacts to wildlife by heavy domestic livestock
grazing vary from moderate to extreme depending
on whether grazing is seasonal or yearlong. Seasonal

188

Riparian Habitats

grazing may allow limited tree and shrub regenera-
tion that provides some habitat and forage for wild-
life, whereas heavy, yearlong grazing eventually leads
to removal of most, if not all, of the palatable ripar-
ian vegetation. In the latter instance, forage and ther-
mal cover for large mammals are slowly eliminated
along with food and habitat for medium-sized and
small mammals, birds, and herps. In seasonal heavy
grazing, some forage and thermal cover may be left
for large mammals, but food and cover for medium-
sized and small mammals are generally eliminated.
Mid-canopy and understory birds may be affected to
the point of exclusion. Yearlong, heavy grazing on
the Verde River in central Arizona has resulted in re-
maining stands of cottonwood-willow communities
of structure Type I or mature communities tending
toward decadence (Higgins and Ohmart 1981). Al-
though floods produce good seedling development,
these seedlings are consumed before they reach
0.6 m ( 2 ft ) in height. Unless corrective measures
are taken, in a few years the old decadent communi-
ties will expire and there will be no young replace-
ment communities. Crouch (1978) reported a 50%
reduction of cottonwoods in a grazed stream in Col-
orado over an 18-year period.

Birds, for example (Ohmart and Anderson 1982),
are associated with four layers of vegetation on the
lower Colorado River:

19 species are associated with the 7.6 m (25 ft)
or taller layer,

10 species with the 4.6 to 7.6 m (15 to 20 ft)

layer,

13 species with the 1.5 to 4.6 m (5 to 15 ft)
layer, and

11 species with the 0.15 to 1.5 m (0.5 to 5 ft)

layer.

The overstory or canopy group (19 species) are
specialists and were absent or poorly represented
when this layer was absent or foliage density highly
reduced. The 23 species in the two middle layers
and some of the 1 1 species in the understory group
were generalists. Some of these species will be pres-
ent even when their foliage layer is absent or poorly
represented. Heavy grazing not only affects the her-
baceous and shrub layers, but over time affects the
upper canopy layers as riparian tree regeneration
is stopped or curtailed.

Domestic livestock grazing in riparian habitats
may be used as a management tool to enhance areas
for wildlife. This approach has potential, but one
would have to know more about the wildlife and its
habitat needs than presently known or would have
to experiment with different levels of forage removal
to bring about the desired wildlife results. Moulton
( 1978) suggested that grazing in a riparian area in

eastern Colorado may increase rodent species num-
bers by creating microhabitat diversity. He found
eight species of small mammals in a grazed plot (2.3
a./AUM [0.9 ha/AUM]) and four species in the un-
grazed plot. The grazed area had been stocked from
July through December with spring use deferred
for 17 years. This type of seasonal grazing, after max-
imum spring plant growth, allowed livestock selec-
tion of preferred species creating a patchy
microhabitat. The ungrazed plot ( 1 1 or more years
of rest) promoted a uniform vegetation structure
without any horizontal patchiness. Prairie vole num-
bers were higher in the protected site but not ex-
cluded in the grazed site.

Prairie vole.

Another possible use of cattle grazing as a man-
agement tool has been reported by Krueger and
Anderson (1985) in dense shrub-willow communi-
ties at high elevations on the North Platte drainage
in Wyoming. Apparently, fish populations are not
harmed by this activity, at least on the Little Des-
chutes River in Oregon (Lorz 1974), but this was
not examined in the study by Krueger and Anderson
(1985). Tunnels created by cattle grazing through
the dense shrub-willow altered plant community
structure, creating a more diverse set of ecological
conditions for birds. Many species respond positively
to this treatment except for the green-tailed towee
(Pipilo chlorums), which inhabits dense stands of
vegetation.

These management tools need to be examined
carefully to ensure that wildlife species targeted
for enhancement receive the benefits. Domestic live-
stock grazing is so entrenched on public land that
we doubt that use of domestic livestock grazing will
become an important management tool in many
ecological situations. First, we must reverse the gen-
eral downward trend in riparian vegetation condi-
tions before cattle are needed as a management tool

Riparian Habitats

189

on a broad basis. Fish habitat appears to be more
sensitive to livestock grazing impacts than terrestrial
habitats.

Mining

Mining can have profound effects on riparian
ecosystems ranging from total sterility of the riparian
system to intermittent effects following heavy pre-
cipitation. The effects vary in areas in the West and
depend on the mineral(s) being mined. In the arid
Southwest, the material is frequently sand and gravel;
removal of these products has caused extensive
channel cutting, reduced water quality, and even
flooding and loss of homes. Ruptured holding ponds
of leached materials from copper mines has resulted
in virtually permanent losses of the flora and fauna in
some streams.

Little advice other than extreme caution and a
full knowledge of the mining operation, proximity of
potential toxins to the stream, safeguards to avoid
pollutants, and good common sense must reign in
this situation. Also, be aware of potential and real
secondary impacts.

An excellent approach is to inventory the re-
source intensively before any mining disturbance, in-
cluding seasonal water quality samples (have a
control stream if at all possible), and then monitor
every 3-5 years thereafter. The control will provide
normal variance data and a comparison should litiga-
tion ever occur.

Recreational Activities

Recreational activity and its effects on wildlife
can range from relatively minor to so severe that
virtually all the vegetation is destroyed locally. In
many instances, the agency developing the recrea-
tional opportunities in riparian habitats builds roads
through the habitats allowing vehicles total access to
the recreational area. Consequently, users drive off
the roads, camp at random, and many assume an
attitude of "destroy anything you want, we won't
ever return." Wood gathering for firewood consumes
down and dead trees, limbs, snags, and many times
standing live trees. Many forms of wildlife leave the
area and others, such as lizards, snakes, frogs, and
salamanders, are destroyed by children and pets.

The impacts of recreational use are poorly docu-
mented, but simply by viewing some recreational
areas one is left with the impression that only the
heartiest and persistent wildlife are left. Aitchison
(1977) studied bird densities and species composi-
tion in a seasonal-use campground in Oak Creek
Canyon in Arizona for 3 years. His control and recre-
ational site primarily supported ponderosa pine
(Pinus ponderosa), cottonwoods, and Arizona wal-

nut at a 1,646-m (5,400-ft) elevation. The camp-
ground was opened from about Memorial Day to
Labor Day each year which spans the bird breeding
season. In the first year of study (first year the devel-
oped campground opened), there was a 40% de-
crease in bird density on opening day. Agency
personnel destroyed 20% of the nests of the Steller's
jay (Cyanocitta stelleri) by removing and slashing
trees. Aitchison (1977:178) reported: "Campers de-
stroyed 30 percent more of the Steller's jay nests
and 20 percent of the robin (Turdus migratorius)
nests by removing branches for firewood, making
room for tents, and other reasons." Many species
abandoned their nests but foraged in the camp-
ground area. Ultimately, bird species remaining in
the campground were larger, different, and fewer
than on the control site.

In highly stressed riparian systems, trailer park
development can be positive to wildlife, especially
birds, as was exemplified in one of our studies. Along
the Colorado River, the cottonwood-willow associa-
tion is rapidly disappearing and a wise developer
planted native trees in the park for better tree sur-
vival and growth, and to attract birds for the enjoy-
ment of the residents. This small oasis supports a few
pairs of birds that were once common along the
lower river (Grinnell 1914), but are now rapidly ap-
proaching extirpation (Anderson and Ohmart 1977).
This type of action should be encouraged in devel-
opments where trees will be protected, but man-
agers must plan for these developments.

Riparian systems are very attractive to recrea-
tionists in that the systems contain water, interesting
plants and animals, shade, and numerous other en-
joyable features in the otherwise arid and semiarid
environments. Hoover et al. (1985) reported in a
visitor information study that environmental attri-
butes receiving highest user ratings were primarily
ecological features present in healthy riparian eco-
systems. Managers should educate the public of the
fragile nature and unique values of these systems.
This may seem impossible, but there are some guide-
lines available such as the Recreational Carrying
Capacity in the California Desert (U.S. Department of
the Interior, Bureau of Land Management 1978) and
the California Desert Area Conservation Plan (U.S.
Department of the Interior, Bureau of Land Manage-
ment 1980).

Martin (1984) provided an excellent approach
in using recreation planning to restore and protect
riparian systems. He recommended ways to control
visitor use subtly and directly in an intensively used
riparian system replete with wildlife values, water
recreation, and large metropolitan areas nearby on
the American River in California. Approaches and
successes in the California State Park System (Barry
1984) and potential problems and questions in the
proposed wildlife enhancement and recreational

190

Riparian Habitats

development at Oristimba Creek in California (Mor-
ris 1984) may also be helpful in better planning. If
new campgrounds are absolutely necessary, Aitchi-
son's (1977) suggestions may apply: locate new
campgrounds in nonsensitive areas, periodically
close to the campground to allow revegetation and
reduce stress on wildlife; open the campground be-
fore or after the height of the breeding season; con-
trol visitors and agency habitat destruction; and
educate the public through signs showing good
camping procedures. We might add to disallow col-
lection of wood for any reason.

Impoundment Construction

Impoundments are constructed for a number of
reasons and some are multiple-use structures. Gener-
ally, each has regulations that outline its purpose
and function such as flood control, hydroelectric
power, and water storage for agriculture or munici-
palities. During construction, roads are built, recrea-
tional facilities may be installed, and numerous other
secondary impacts to wildlife may occur, along with
the eventual inundation of the vegetation in the stor-
age reservoirs. In some instances, the secondary
impacts of reservoirs to wildlife can equal or exceed
primary impacts.

Reservoirs. Depending on how rapidly water
surface levels fluctuate behind reservoirs, there may
be a potential for productive wildlife habitats to
develop. Rapidly and wildly fluctuating water levels
are not conducive to the development of emergent
plant communities such as cattails (Typha sp.) and
bulrushes (Scirpus sp. ). These types of reservoirs
develop a "bathtub ring" for a shoreline where only
annual plants grow and perennials are drowned.
Reservoirs with slow fluctuating water levels
generally develop good emergent plant communities
that support animals such as muskrat, beaver, rails,
and gallinules. Waterfowl tend to more heavily use
reservoirs that have an abundance of emergent
vegetation that provides cover and greater foraging
opportunities. Reservoirs with rapidly fluctuating
shorelines tend to attract only a few diving ducks.

Downstream. The wildlife value of areas below
reservoirs tends to degrade slowly over the years.
Generally, instream flows are lower than natural
flows. Natural floods that provide new soil deposition
and enrichment are stopped, and riparian plant
health and vigor slowly decline. If controlled
releases from the dam or floods do occur, they are
generally greater than would have occurred without
the dam, and the health of the riparian system may
have degraded to a point that is no longer resilient
to a heavy flood.

Recreational activities can affect riparian areas.

Riparian Habitats

191

Another problem tends to occur below dams
where instream flows are highly regulated and cessa-
tion of natural floods prevent leaching and soil reju-
venation. Total Dissolved Solids (TDS) tend to
increase in areas where the water table is near the
soil surface. If the stream carries high TDS loads, the
process is relatively rapid and the TDS are wicked
to the soil surface where they accumulate as the soil
moisture evaporates. Sodium or salt levels eventually
accumulate to a point that most native species, ex-
cept halophytes (salt-tolerant plants), cannot germi-
nate or survive. Natural floods generally leached
and removed these deposits, and new soils were de-
posited in low areas.

In some situations below dams the riverbed is
essentially dewatered until another major perennial
tributary enters. The Salt and Gila rivers in central
Arizona and the Rio Grande from about El Paso to
Presidio, Texas, are classic examples. Water return-
ing to these dry reaches is in one of two forms: agri-
cultural waste waters high in chemicals and salts or
impoundment releases because reservoirs are near or
at capacity. Agriculture waste waters generally poi-
son productive riparian vegetation and create condi-
tions that favor growth of less desirable trees or
shrubs. These waters seldom flow in the original
channel and eventually the channel is obliterated.

During high rainfall years, releases from up-
stream dams must occur, and the dense, low-growing
trees and shrubs that cover the floodplain and block
the channel form a living dam that spreads the re-
leased water laterally to inundate everything in the
floodplain. Floodwaters drain slowly (generally tak-
ing months) and frequently relict, productive native
plant communities drown.

Logging and Roads

We stated earlier that riparian ecosystems can
be and are affected by any major perturbation,
whether it be natural (fire, storms) or man-made
(logging, roadbuilding, or grazing). Therefore, man-
agement must consider all disturbances that could
potentially or actually affect riparian ecosystems.
Productive fisheries can be lost to high stream sedi-
ment loads, to stream channel and streamside vegeta-
tion destruction by floods following abuse of
watersheds, and to perennial streams becoming in-
termittent because of continued abuse of watersheds.

Logging and roadbuilding on the watersheds
and near riparian systems destroy the natural ground
cover and churn and mix the soil to produce trans-
portable sediment. Sediment entering a stream
comes from both natural and man-made activities
and its rate of passage varies depending on slope,
size of area disturbed, severity of disturbance, kind
and type of streamside vegetation to stabilize the

transported materials, instream sediment traps, and
the periodicity and duration of large streams. Dunn
and Leopold (1978) estimated that 5 or more years
are needed for a transportable sediment load to to-
tally pass through a stream system. Mahoney and
Erman ( 1984) reported that the sediment load cur-
rently moving through any stream is the product
of past years' land-use activity and major storms.

Leaving buffer strips near riparian vegetation is
apparently successful. Aubertin and Patric (1974)
studied a 34-ha (14-a.) clearcut in West Virginia and
reported only slight increases in stream turbidity
following timber harvest. They attributed the success
to leaving a 10 to 20-m (33 to 66-ft) forested strip
adjacent to the riparian vegetation. Moring (1975)
demonstrated similar results in his 15 -year study
in Oregon. He showed a 3-8-fold reduction in sus-
pended stream sediments in the clearcut with a
buffer strip versus a clearcut without the buffer strip.

A large study on streams in northern California
examined macroinvertebrate changes relative to
logging with buffered and unbuffered strips (Erman
et al. 1977; Ruby et al. 1977; Newbold et al. 1980).
Where buffer strips were > 30 m ( > 98 ft ) on
logged sites, there were no differences between in-
vertebrate populations in experimentals and con-
trols. Where buffer strips were less, differences of
invertebrate populations were detectable between
experimentals and controls.

Buffer strips also reduce pollutants and other
chemical substances from surface runoff (Young et
al. 1980). Karr and Schlosser (1977) extensively
reviewed literature on the value of near-stream vege-
tation on water quality and stream biology and
should be consulted for more in-depth coverage.

Not only are buffer strips effective in reducing
physical and, ultimately, biological damage to lower
animals, they also protect small mammals from inten-
sive logging operations (Cross 1985). Where buffer
or leave strips varying from 12 to 70 m (39 to 230
ft) wide were retained, those remaining habitats
supported small mammal communities comparable
to undisturbed sites. These studies were conducted
in southwestern Oregon in mixed-coniferous riparian
vegetation. Harris (1984) suggested maintaining
riparian corridors as a means of connecting forest
habitat islands in similar stands of old-growth Doug-
las fir (Pseudotsuga menziesii).

Riparian Ground-water Withdrawals

Ground-water pumping may become the most
serious threat to North American riparian systems.
Water diversions and reduced instream flows can be
devastating to riparian habitats, but pumping of ripar-
ian ground waters for industrial and municipal devel-
opment will totally annihilate most, if not all,

192

Riparian Habitats

riparian plant species. Extensive mesquite bosques
were killed around Casa Grande Ruins or the Casa
Grande National Monument by ground-water pump-
ing in central Arizona (Judd et al. 1971 ). The water
table has receded at 2+ m(7+ ft) per year which,
in turn, resulted in the death of this large mesquite
forest.

Long-term vegetation changes are well docu-
mented by aerial photography for a 32-km (2-mi. )
reach of the Carmel River near the Monterey Penin-
sula in California (Groeneveld and Griepentrog
1985). Time-series documentation for a 24-year pe-
riod (1956-80) conclusively demonstrated a marked
decline in riparian trees such as red willow (S. laevi-
gata), black cottonwood {P. trichocarpa), California
sycamore {Platanus racemosa), and white alder
(Alnus rhombifolia). Along with reduced riparian

plant cover was an invasion of weedy perennials,
annuals, and xerophylic (arid-adapted) shrubs. The
riverbanks have become noticeably eroded due to
increased channel width.

Groeneveld and Griepentrog (1985) cited other
unpublished studies where lowering of the water
table by pumping had caused ecological change. This
is a real and final threat to many riparian ecosystems
in the West. The U.S. Geological Survey is currently
studying the effects of ground-water withdrawal on
native riparian species along the Owens River south
of Bishop, California (Dileanis et al. 1985). Unfortu-
nately, no tree species were involved in the study,
but data for Nevada saltbush (Atriplex torreyi) and
rubber rabbitbush (Cbrysothamnus nauseosus) may
provide insight into how sensitive obligate-riparian
species are to gradual and drastic declines of the
water table.

Riparian Habitats

193

LITERATURE CITED

AITCHISON, S.W. 1977. Some effects of a campground on
breeding birds in Arizona. Pages 175-182 in Johnson,
R.R. and D.A. Jones, tech. coords. Importance, Preser-
vation and Management of Riparian Habitat: A Sym-
posium. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-43. 217pp.

ALMAND, J.D. and W.B. KROHN. 1978. The position of
the Bureau of Land Management on the protection
and management of riparian ecosystems. Pages 359-
361 in Johnson, R.R. andJ.F. McCormick, tech.
coords. Strategies for Protection and Management of
Floodplain Wetlands and Other Riparian Ecosystems:
Proc. of the Symposium. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. WO-12., Washington, DC. 4 10pp.

AMES, C.R. 1977. Wildlife conflicts in riparian manage-
ment: grazing. Pages 49-51 in Johnson, R.R. and D.A.
Jones, tech. coords. Importance, Preservation and
Management of Riparian Habitat: A Symposium. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. RM-43- 217pp.

ANDERSON, B.W. and R.D. OHMART. 1977. Breeding
bird use of a residential development. Pages 196-201
in Annual Report, submitted to U.S. Dep. Inter.,
Bureau of Reclamation, Boulder City, NV.

and , 1978. Phainopepla utilization of honey

mesquite forests in the Colorado River valley. Condor
80:334-338.

and , 1984. Vegetation management study

for the enhancement of wildlife along the lower Colo-
rado River. U.S. Dep. Inter., Bur. of Rec, Boulder
City, NV.

-, and W.C. HUNTER. 1983. Quantifying

variables for classifying desert riparian vegetation.
Pages 32-44 in Moir, W.H. and L. Hendzel, tech.
coords., Proc. of the Workshop on Southwestern Hab-
itat Types. U.S. Dep. Agric, For. Serv., Southwest
Reg., Albuquerque, NM.

ARMOUR, C.L. 1977. Effects of deteriorated range streams
on trout. U.S. Dep. Inter., Bur. Land Manage., Idaho
State Off., Boise. 7pp.

ASMUSSEN, L.E., AW. WHITE, Jr., E.W. HANSON, and J.M.
SHERIDAN. 1977. Reduction of 2,4-d load in surface
runoff down a grassed waterway. J. Environ. Qual.
6:159-162.

AUBERTIN, G.M. and J.H. PATR1C. 1974. Water quality
after clearcutting in a small watershed in West Vir-
ginia. J. Environ. Qual. 3:243-249.

AXELROD, D.I. 1958. Evolution of the Madro-Tertiary
geoflora. Bot. Rev. 24:433-509.

BALDA, R. 1975. The relationship of secondary cavity
nesters to snag densities in western coniferous for-
ests. U.S. Dep. Agric, For. Serv., Southwest Reg., Wildl.
Habitat Tech. Bull. 1, Albuquerque, NM. 37pp.

BARRY, W.J. 1984. Management and protection of riparian
ecosystems in the state park system. Pages 758-766
in Warner, R.E. and KM. Hendrix, eds. California
Riparian Systems: Ecology, Conservation, and Produc-
tive Management. Univ. California Press, Berkeley.
1035pp.

BELLUOMINI, L. and G. TRAPP. 1984. Ringtail distribution
and abundance in the Central Valley of California.
Pages 906-914 in Warner, R.E. and KM. Hendrix, eds.
California Riparian Systems: Ecology, Conservation,
and Productive Management. Univ. California Press,
Berkeley. 1035pp.

BENOIT, C. 1978. Fluvial sediment delivery as percent of
erosion: The relationship between landslope and
effective streamside bufferstrip width. U.S. Dep. Agric,
For. Serv., Portland, OR. (typescript)
BLACK, N., R.J. SCHERZINGER, and J.W. THOMAS. 1976.
Relationship of Rocky Mountain elk and Rocky Moun-
tain mule deer to timber management in the Blue
Mountains of Oregon and Washington. Pages 11-31 in
Hieb, S.R., ed. Proc. Elk-Logging-Roads Symposium,
Univ. Idaho, Moscow. 142pp.
BLOOD, DA. 1966. Range relationships of elk and cattle
in Riding Mountain National Park, Manitoba. Canadian
Wildl. Serv., Wildl. Manage. Bull. Ser. 1,19. 62pp.
BOCK, J.H. and C.E. BOCK. 1985. Patterns of reproduction
in Wright's sycamore. Pages 493-494 in Johnson,
R.R., CD. Ziebell, DR. Patton, P.F. Ffolliott, and R.H.
Hamre, tech. coords. Riparian Ecosystems and Their
Management: Reconciling Conflicting Uses. Proc. First
North Am. Riparian Conf. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-120. 523pp.
BOEER, W.J. and D.J. SCHMIDLY. 1977. Terrestrial mam-
mals of the riparian corridor in Big Bend National
Park. Pages 212-217 in Johnson, R.R and DA. Jones,
eds. Importance, Preservation and Management of
Riparian Habitat: A Symposium. U.S. Dep. Agric, For.
Serv. Gen. Tech. Rep. RM-43- 217pp.
BRADY, W., DR. PATTON, and J. PAXSON. 1985. The

development of southwestern riparian gallery forests.
Pages 39-43 in Johnson, R.R., CD. Ziebell, DR. Pat-
ton, R.F. Ffolliott, and R.H. Hamre, tech. coords. Ri-
parian Ecosystems and Their Management:
Reconciling Conflicting uses. Proc. First North Am.
Riparian Conf. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. RM-120. 523pp.
BRODE, J.M. and R.B. BURY. 1984. The importance of

riparian systems to amphibians and reptiles. Pages 30-
36 in Warner, R.E. and KM. Hendrix, eds. California
Riparian Systems: Ecology, Conservation, and Produc-
tive Management. Univ. California Press, Berkeley.
1035pp.
BROWN, D.E., C.H. LOWE, and CP. PASE. 1979. A digi-
tized classification system for the biotic communities
of North America, with community (series) and asso-
ciation examples for the Southwest. J. Arizona-Nevada
Acad. Sci. 14 (Suppl. 1):1-16.
BRUSH, T., B.W. ANDERSON, and R.D. OHMART. 1983-
Habitat selection related to resource availability
among cavity-nesting birds. Pages 88-89 in Davis,
J.W., GA. Goodwin, and RA. Ockenfels, tech. coords.
Snag Habitat Management: Proc. Symposium. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. RM-99. 226pp.
BUCK, S.G 1982. Habitat utilization by fisher near Big
Bear, California. M.S. Thesis, Humboldt State Univ.,
Areata, CA. 178pp.
BULL, E.L. and J.N. SKOVLEN. 1982. Relationships be-
tween anifauna and streamside vegetation. Trans.
North. Am. Wildl. Nat. Resour. Conf. 47:496-506.
BURBRIDGE, W.R. and D.J. NEFF. 1976. Coconino Na-
tional Forest — Arizona Game and Fish Department co-
operative roads — wildlife study. Pages 44-57 in Hieb,
S.R., ed. Proc Elk-Logging-Roads Symposium, Univ.
Idaho, Moscow.
BURTON, T.M. and G.E. LIKENS. 1975. Salamander popula-
tion and biomass in the Hubbard Brook experimental
forest, New Hampshire. Copeia 1975:541-546.
BURY, R.B. 1979. Population ecology of freshwater turtles
in Harless, M. and H. Morlock, eds. Turtles: Prospec-

194

Riparian Habitats

tives and Research. John Wiley & Sons, Inc., New
York, NY. 695pp.

CAROTHERS, S.W., R.R. JOHNSON, and S.W. AITCHISON.
1974. Population structure and social organization
of southwestern riparian birds Am. Zool. 14:97-108.

CEBALLOS, G. 1985. The importance of riparian habitats
for the conservation of endangered mammals in Mex-
ico. Pages 96-100 in Johnson, RR, CD. Ziebell, DR.
Patton, P.F. Ffolliott, and R.H. Hamre, tech. coords.
Riparian Ecosystems and Their Management: Recon-
ciling Conflicting Uses. Proc. First North Am. Riparian
Conf. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM 120. 523pp.

CHAIMSON, J.F. 1984. Riparian vegetation planting for

flood control. Pages 120-123 in Warner, RE. and K.M.
Hendrix, eds. California Riparian Systems: Ecology,
Conservation, and Productive management. Univ.
California Press, Berkeley. 1035pp.

CONNOR, R.N., R.G. HOOPER, H.S. CRAWFORD, and H.S.
MOSBY. 1975. Woodpecker nesting habitat in cut and
uncut woodlands in Virginia. J. Wildl. Manage. 39:144-
150.

CORBETT, E.S. andJ.A. LYNCH. 1985. Management of
streamside zones on municipal watersheds. Pages
187-190 in Johnson, R.R, CD. Ziebell, DR. Patton,
R.F. Ffolliott, and R.H. Hamre, tech. coords. Riparian
Ecosystems and Their Management: Reconciling Con-
flicting Uses. Proc. First North Am. Riparian Conf.
U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. RM-120.
523pp.

CROSS, S.P. 1985. Responses of small mammals to forest
riparian perturbations. Pages 269-275 in Johnson,
R.R., CD. Ziebell, DR. Patton, R.F. Ffolliott, and R.H.
Hamre, tech. coords. Riparian Ecosystems and Their
Management: Reconciling Conflicting Uses. Proc. First
North Am. Riparian Conf. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-120. 523pp.

CROUCH, G.L. 1978. Effects of protection from livestock
grazing on a bottomland wildlife habitat in northeast-
ern Colorado. Pages 118-125 in Graul, WD. and S.J.
Bissell, tech. coords. Lowland River and Stream Habi-
tat in Colorado: A Symposium. Colorado Chap. Wildl.
Soc and Colorado Audubon Council. 195pp.

. 1982. Wildlife on ungrazed and grazed bottom-
lands on the South Platte River in northeastern Colo-
rado. Pages 186-197 in Wildlife-Livestock
Relationships Symposium. Proc. 10, Univ. Idaho, For.
Wildl. Range Exper. Sta., Moscow, ID.

CUMMINS, K.W. 1974. Structure and function of stream
ecosystems. BioScience 24:631-641.

DAVIS, GA. 1977. Management alternatives for the ripar-
ian habitat in the Southwest. Pages 59-67 in Johnson,
R.R. and DA. Jones, tech. coords. Importance, Preser-
vation and Management of Riparian Habitat: A Sym-
posium. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-43. 217pp.

DAVIS, J.W., GA. GOODWIN, and R.A. OCKENFELS, tech.
coords. 1983- Snag habitat management: Proc. of
the Symposium. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. RM-99. 226pp.

DEALY, J.E., DA. LECKENBY, and DM. CONCANNON.
1981. Plant communities and their importance to
wildlife. Pages 1-66 in Wildlife Habitats in Managed
Rangelands — The Great Basin of Southeast Oregon.
U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. PNW-120.

DE VOS, A. 1951. Recent findings in fisher and marten

ecology and management. Trans. 16th North American

Wildl. Conf. 16:498-507.

1952. Ecology and management of fisher and mar-

ten in Ontario. Ontario Dep. Lands For. Tech. Bull.
90pp.

DICKSON, J.G., R.N. CONNER, and J.H. WILLIAMSON.
1983 Snag retention increases birds in a clearcut. J.
Wildl. Manage. 47:799-809.

DILEANIS, P.D., FA. BRANSON, and S.K. SORENSON.

1985. Methods for determining effects of controlled
dewatering of shallow aquifers on desert phreato-
phytes in Owens Valley, California. Pages 197-200 in
Johnson, R.R, CD. Ziebell, DR. Patton, R.F. Ffolliott,
and R.H. Hamre, tech. coords. Riparian Ecosystems
and Their Management: Reconciling Conflicting Uses.
Proc First North Am. Riparian Conf. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. RM-120. 523pp.

DUFF, DA. 1979. Riparian habitat recovery on Big Creek,
Rich County, Utah. Pages 91-92 in Proceedings of the
Forum — Grazing and Riparian/Stream Ecosystems.
Trout Unlimited, Inc. Vienna, VA. 94pp.

DUNN, T and LB. LEOPOLD. 1978. Water in environmen-
tal planning. W.H. Freeman Co., San Francisco, CA.
818pp.

ENGEL-WILSON, R.W. and R.D. OHMART. 1978. Floral
and attendant faunal changes on the lower Rio
Grande between Fort Quitman and Presidio, Texas.
Pages 139-147 in Johnson, R.R. and J.F. McCormick,
tech. coords. Strategies for Protection and Manage-
ment of Floodplain Wetlands and Other Riparian
Ecosystems. Proc of the Symposium. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. WO- 12. 4 10pp.

ERMAN, D.C., J.D. NEWBOLD, and K.R. RUBY. 1977. Eval-
uation of streamside bufferstrips for protecting
aquatic organisms. Contr. 165, California Water Re-
sour. Center, Univ. California, Davis. 48pp.

EVANS, K.E. and R.N. CONNOR. 1979. Snag management.
Pages 214-225 in DeGraff, RM. and K.E. Evans, com-
pilers. Management of North-Central and Northeast-
ern Forests for Nongame Birds. U.S. Dep. Agric, For.
Serv. Gen. Tech. Rep. NC-51.

EVEREST, F.H. 1977. How to demonstrate the importance
of fishery resources to interdisciplinary planning
teams. Fisheries 4:15-19.

GAINES, D. 1977. The valley riparian forests of California:
Their importance to bird populations. Pages 57-85
in Sands, A., ed. Riparian Forests in California: Their
Ecology and Conservation. Inst. Ecology, Univ. Cali-
fornia, Davis. 122pp.

GILLEN, R. 1981. 1980 progress report to U.S. Forest
Service Pacific Northwest Forest and Range Experi-
ment Station. Under Coop. Aid Program 218, PNW
Suppl. 218. Oregon State Univ., Corvallis. 10pp.

GLINSKI, R.L. 1977. Regeneration and distribution of

sycamore and cottonwood trees along Sonoita Creek,
Santa Cruz County, Arizona. Pages 1 16-123 in John-
son, R.R. and DA. Jones, tech. coords. Importance,
Preservation and Management of Riparian Habitat: A
Symposium. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. RM-43. 217pp.

GREGORY, K.J. and JR. MADEW. 1982. Land use changes,
flood frequency and channel adjustment. Pages 757-
781 in Hey, R.D., J.C Bathurst, and C.R. Thome, eds.
Gravel-bed. John Wiley & Sons, New York, NY.

GRINNELL, J. 1914. An account of the mammals and birds
of the lower Colorado Valley with special reference
to the distributional problems presented. Univ. Cali-
fornia Publ. Zool. 12:51-294.

Riparian Habitats

195

and T.I. STORER. 1924. Animal life in the Yosemite.

Univ. California Press, Berkeley. 952pp.

GROENEVELD, DP. and T.E. GRIEPENTROG. 1985. Inter-
dependence of groundwater, riparian vegetation,
and streambank stability: A case study. Pages 44-48 in
Johnson, R.R., CD. Ziebell, D.R. Patton, P.F. Ffolliott,
and R.H. Hamre, tech. coords. Riparian Ecosystems
and Their Management: Reconciling Conflicting Uses.
Proc. First North Am. Riparian Conf. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. RM-120. 523pp.

GRUBB, T.C, Jr., D.R. PETIT, and D.L. KRUSAC. 1983-
Artificial trees for primary cavity nesters. Pages 151-
154 in Davis, J.W., G.A. Goodwin, and R.A. Ockenfels,
tech. coords. Snag Habitat Management: Proceedings
of the Symposium. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. RM-99. 226pp.

HAAPANEN, A. 1965. Bird fauna of the Finnish forests in
relation to forest succession. I. Ann. Zool. Fenn.
2:153-196.

HALLBERY, D.L. and G.R. TRAPP. 1984. Gray fox temporal
and spatial activity in a riparian/agricultural zone in
California's Central Valley. Pages 920-928 in Warner,
R.E. and KM. Hendrix, eds. California Riparian Sys-
tems: Ecology, Conservation, and Productive Manage-
ment. Univ. California Press, Berkeley. 1035pp.

HARRIS, L.D. 1984. The fragmented forest— island bio-
geography theory and the preservation of biotic di-
versity. Univ. Chicago Press, Chicago, IL. 211pp.

HAUPT, H.F. 1959. A method for controlling sediment
from logging roads. U.S. Dep. Agric, For. Serv. Inter-
mountain For. Range Exp. Sta., Misc. Paper 22, Ogden,
UT. 22pp.

HEEDE, B. 1985. Interactions between streamside vegeta-
tion and stream dynamics. Pages 54-58 in Johnson,
R.R, CD. Ziebell, D.R. Patton, R.F. Ffolliott, and R.H.
Hamre, tech. coords. Riparian Ecosystems and Their
Management: Reconciling Conflicting Uses. Proc. First
North Am. Riparian Conf. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-120. 523pp.

HEHNKE, M. and CP. STONE. 1978. Value of riparian

vegetation to avian populations along the Sacramento
River system. Pages 228-235 in Johnson, R.R. and
J.F. McCormick, tech. coords. Strategies for Protection
and Management of Floodplain Wetlands and Other
Riparian Ecosystems: Proc. of the Symposium. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. WO-12., Wash-
ington, DC. 4 10pp.

HIGGINS, A.E. and R.D. OHMART. 1981. Riparian habitat
analysis: Tonto National Forest. U.S. Dep. Agric, For.
Serv., Tonto Natl. For., Phoenix, AZ.

HINK, V. and R.D. OHMART. 1984. Middle Rio Grande
biological survey final report. U.S. Army Corps of
Engineers, Albuquerque, NM.

HOOVER, ST., DA. KING, and W.J. MATTER. 1985. A

wilderness riparian environment: Visitor satisfaction,
perceptions, reality, and management. Pages 223-
226 in Johnson, R.R., CD. Ziebell, D.R. Patton, P.F.
Ffolliott, and R.H. Hamre, tech. coords. Riparian Eco-
systems and Their Management: Reconciling Conflict-
ing Uses. Proc. First North Am. Riparian Conf. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. RM-120.
523pp.

HUBBARD, J.P. 1971. The summer birds of the Gila Val-
ley, New Mexico. Nemouria 2. 35pp.

HULSE, A.C 1974. An autoecological study of Kinosternon
soroniense LeConte (Chelonia: Kinosternidae ). PhD
dissertation, Dep. Zool., Arizona State University,

Tempe. 105pp.

JOHNSON, R.R. 1978. The lower Colorado River: A west-
ern system. Pages 41-55 in Johnson, R.R. and J.F.
McCormick, tech. coords. Strategies for Protection
and Management of Floodplain Wetlands and Other
Riparian Ecosystems: Proc. of the Symposium. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. WO-12., Wash-
ington, DC. 4 10pp.

and S.W. CAROTHERS. 1981. Southwestern riparian

habitats and recreation: Interrelationships and impacts
in the Rocky Mountain region. Eisenhower Consor-
tium Bulletin. U.S. Dep. Agric, For. Serv. Rocky Moun-
tain For. Range Exp. Sta., Fort Collins, CO.

and D.A.JONES, tech. coords. 1977. Importance,

preservation and management of riparian habitat:
A symposium. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. RM-43. 217pp.

and J.F. MCCORMICK, tech. coords. 1978. Strate-
gies for protection and management of floodplain
wetlands and other riparian ecosystems. Proc. Sym-
posium. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
WO-12. 4 10pp.

, CD. ZIEBELL, D.R. PATTON, P.F. FFOLLIOTT, and

R.H. HAMRE, tech. coords. 1985. Riparian ecosystems
and their management: reconciling conflicting uses.
Proc. First North Am. Riparian Conf. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. RM-120. 523pp.

JUDD, J.B., J.M. LAUGHLIN, H.R. GUENTHER, and R. HAN-
DERGRADE. 1971. The lethal decline of mesquite
on the Casa Grande National Monument. Great Basin
Nat. 31:153-159.

KARR, JR. and I.J. SCHLOSSER. 1977. Impact of near-
stream vegetation and stream morphology on water
quality and stream biota. Env. Res. Lab. EPA-600/3-77-
097, U.S. EPA, Athens, GA. 91pp.

and . 1978. Water resources and the land-
water interface. Science 201:229-234.

KELLERT, S.R. 1980. What do North Americans expect of
wildlife management agencies? Proc. 70th Meeting
Int. Assoc. Fish, and Wildl. agencies, Washington, DC.
9pp.

KELLY, G.M. 1977. Fisher (Martes pennanti) biology in
the White Mountains National Forest and adjacent
areas. Ph.D. dissertation, Univ. Massachusetts, Am-
herst. 178pp.

KENNEDY, CE. 1977. Wildlife conflicts in riparian man-
agement: grazing. Pages 49-51 in Johnson, R.R. and
D.A. Jones, tech. coords. Importance, Preservation and
Management of Riparian Habitat: A Symposium. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. RM-43- 217pp.

KNIGHT, A.W. and R.L. BOTTORFF. 1984. The importance
of riparian vegetation to stream ecosystems. Pages
160-167 in Warner, RE. and KM. Hendrix, eds. Cali-
fornia Riparian Systems: Ecology, Conservation, and
Productive Management. LIniv. California Press, Berke-
ley. 1035pp.

KNOPF, F.L. 1985. Significance of riparian vegetation to
breeding birds across an altitudinal cline. Pages 105-
111 in Johnson, R.R., CD. Ziebell, D.R. Patton, P.F.
Ffolliott, and R.H. Hamre, tech. coords. Riparian Eco-
systems and Their Management: Reconciling Conflict-
ing Uses. Proc. First North Am. Riparian Conf. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. RM-120.
523pp.

and R.W. CANNON. 1981. Structural resilience of a

willow riparian community to change in grazing prac-
tices in Peek, J., ed. Livestock- Wildlife Relationships

196

Riparian Habitats

Symposium. Univ. Idaho, Moscow.

KRUEGER, HO. and S.H. ANDERSON. 1985. The use of
cattle as a management tool for wildlife in shrub-
willow riparian systems. Pages 300-304 in Johnson,
R.R., CD. Ziebell, DR. Patton, P.F. Ffolliott, and R.H.
Hamre, tech. coords. Riparian Ecosystems and Their
Management: Reconciling Conflicting Uses. Proc. First
North Am. Riparian Conf. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-120. 523pp.

LANG, G.E. and R.K. WEIDER. 1984. The role of beaver in
vegetation patterning and development in Sphagnum-
dominated wetlands in West Virginia. Bull. Ecol. Soc.
Am. 65(2):243.

LAURENZI, A.W., R.D. OHMART, and V. HINK. 1983.
Classification of mixed broadleaf riparian forest in
Tonto National Forest. Pages 72-81 in Moir, W.H. and
L. Hendzel, tech. coords. Proc. of the Workshop on
Southwestern Habitat Types. U.S. Dep. Agric, For.
Serv., Southwest Reg., Albuquerque, NM.

LI, R.M. and W.H. SHEN. 1973. Effects of tall vegetation
and flow sediment. J. Hydraul. Div., ASCE, Vol. 9, No.
HY5, Proc. Paper 9748.

LORZ, H.H. 1974. Ecology and management of brown

trout in Little Deschutes River. Oregon Dep. Fish and
Wildl. Fish. Res. Rep. 8. 49pp.

LOWE, C.H., ed. 1964. The vertebrates of Arizona. Univ.
Arizona Press, Tucson. 270pp.

LYONS, JK. and R.L. BESCHTA. 1983. Land use, floods, and
channel changes, Upper Middle Fork Williamette
River, Oregon (1936-1980). Water Resour. Res.
19:436-471.

MACARTHUR, R.H. and J.W. MACARTHUR. 1961. On bird
species diversity. Ecology 42:594-598.

MAHONEY, D.L. and D.C. ERMAN. 1984. The role of
streamside buffer strips in the ecology of aquatic
biota. Pages 168-176 in Warner, R.E. and KM. Hen-
drix, eds. California Riparian Systems: Ecology, Con-
servation, and Productive management. Univ.
California Press, Berkeley. 1035pp.

MANNAN, R.W., E.C. MESLOW, and H.M WIGHT. 1980.
Use of snags by birds in Douglas fir forests, western
Oregon. J. Wildl. Manage. 44:787-797.

MARCUM, C.L. 1975. Summer-fall habitat selection and use
by a western Montana elk herd. Ph.D. dissertation,
Univ. Montana, Missoula. 188pp.

MARCUSON, P.E. 1977. The effects of cattle grazing on
brown trout in Rock Creek, Montana. Project F-20-R-
21-1 la. Montana Game and Fish, Helena. 28pp.

MARLOW, C.B. and T.M. POGACNIK 1985. Time of graz-
ing and cattle-induced damage to streambanks. Pages
279-284 in Johnson, R.R., CD. Ziebell, D.R. Patton,
P.F. Ffolliott, and R.H. Hamre, tech. coords. Riparian
Ecosystems and Their Management: Reconciling Con-
flicting Uses. Proc. First North Am. Riparian Conf.
U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. RM-120.
523pp.

MARTIN, KE. 1984. Recreation planning as a tool to re-
store and protect riparian systems. Pages 748-757 in
Warner, R.E. and KM. Hendrix, eds. California Ripar-
ian Systems: Ecology, Conservation, and Productive
Management. Univ. California, Berkeley. 1035pp.

MAY, B. and B. DAVIS. 1982. Practices for livestock grazing
and aquatic habitat protection on western rangelands.
Pages 271-278 in Wildlife-Livestock Relationships
Symposium: Proc. 10, Univ. Idaho, Forest, Wildl.,
Range Exp. Sta., Moscow.

MCLEAN, A., H.H. NICHOLSON, and A.L. VAN RYSWYK

1963- Growth, productivity and chemical composi-
tion of a subalpine meadow in interior British Colum-
bia. J. Range Manage. 16.235-240.

MEENTS, JK, B.W. ANDERSON, and R.D. OHMART. 1982.
Vegetation relationships and food of sage sparrows
wintering in honey mesquite habitat. Wilson Bull.
94:129-138.

MEGANK, R.A. and KC GIBBS. 1979. A methodology
applied to the analysis of selected grazing manage-
ment strategy and dispersed recreation. Final report
to U.S. Dep. Agric, For. Serv., Pacific Northwest For.
Range Exp. Sta. under Coop. Aid Prog. PNW-40,
Suppl. 57. Oregon State Univ., Corvallis. 67pp.

MERRIAM, C.H. 1890. Results of a biological survey of the
San Francisco Mountains region and desert of the
Little Colorado in Arizona. U.S. Dep. Agric, North Am.
Fauna 3:1-136.

MORING, JR. 1975. The Alsea watershed study: Effects of
logging on the aquatic resources of the headwater
streams of the Alsea River, Oregon. Part II. Fishery
Res. Rep. 9, Oregon Dep. Fish and Wildlife, Corvallis.
39pp.

MORRIS, R.H. 1984. Planning recreation development and
wildlife enhancement in a riparian environment at
Oristimba Creek. Pages 767-772 in Warner, R.E. and
KM. Hendrix, eds. California Riparian Systems: Ecol-
ogy, Conservation, and Productive Management. Univ.
California, Berkeley. 1035pp.

MOSCONI, S.L. and R.L. HUTTO 1982. The effects of
grazing on land birds of a western Montana riparian
habitat. Pages 221-233 in Wildlife-Livestock Relation-
ships Symposium: Proc. 10, Univ. Idaho For. Wildl.
Range Exp. Sta., Moscow.

MOULTON, M. 1978. Small mammal associations in grazed
versus ungrazed Cottonwood riparian woodland in
eastern Colorado: A symposium. Pages 133-140. Colo-
rado Chap. Wildl. Soc. and Colorado Audubon Coun-
cil, Greeley. 195pp.

MULLIS, C. 1985. Habitat utilization by fisher {Martes

pennanti) near Hayfork Bally, California. M.S. Thesis,
Humboldt State Univ., Areata, CA.

MURPHY, ML. and J.D. HALL. 1981. Varied effects of
clearcut logging on predators and their habitat in
small streams of the Cascade Mountains, Oregon. Can.
J. Fish. Aquat. Sci. 38:137-145.

NELSON, JR. and D.G. BURNELL. 1975. Elk-cattle compe-
tition in central Washington. Pages 71-83 in Range
Multiple Use Management. Univ. Idaho, Moscow.

NELSON, L, Jr. and JK. HOOPER. 1976. California fur

bearers and their management. Univ. California, Coop.
Ext. Serv., Leaflet 2721, Berkeley.

NEWBOLD, J.D., DC. ERMAN, and KB. RUBY. 1980. Ef-
fects of logging on macroinvertebrates in streams
with and without bufferstrips. Can. J. Fish. Aquat Sci.
37:1076-1085.

NUSSBAUM, R.A. 1974. The distributional ecology and life
history of the Siskiyou Mountain salamander {Pletho-
don stormi), in relation to the potential impact of the
proposed Applegate Reservoir on this species. Rep.
submitted to U.S. Army Corps of Engineers, Portland,
OR. 52pp.

OHMART, R.D. and B.W. ANDERSON. 1974. Vegetation
management: Annual report. Submitted to U.S. Dep.
Inter., Bureau of Reclamation, Boulder City, NV.

and . 1982. North American desert riparian

ecosystems. Pages 433-479 in Bender, G.L., ed. Refer-
ence Handbook on the Deserts of North America.

Riparian Habitats

197

Greenwood Press, Westport, CT. 594pp.

OLSON, R.W. and C.L. ARMOUR. 1979. Economic consid-
erations for improved livestock management ap-
proaches for fish and wildlife in riparian/stream areas.
Pages 67-71 in Cope, O.B., ed. Forum — Grazing and
Riparian/Stream Ecosystems. Trout Unlimited, Inc.
Vienna, VA. 94pp.

OREGON-WASHINGTON INTERAGENCY WILDLIFE

COUNCIL. 1978. Managing riparian zones for fish and
wildlife in eastern Oregon and eastern Washington.
Unpubl. rep.

ORR, R.T. 1940. The rabbits of California. Occ. Papers.
California Acad. Sci. 19:1-227.

PARKER, M., F. WOOD, Jr., B.H. SMITH, and R.G. ELDER.
1985. Erosional downcutting in lower order riparian
ecosystems: Have historical changes been caused
by removal of beaver? Pages 35-38 in Johnson, R.R.,
CD. Ziebell, DR. Patton, P.F. Ffolliott, and R. H.
Hamre, tech. coords. Riparian Ecosystems and Their
Management: Reconciling Conflicting Uses. Proc. First
North Am. Riparian Conf. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-120. 523pp.

PAULSEN, H.A., Jr. 1969. Forage value on a mountain

grassland aspen range in western Colorado. J. Range
Manage. 22:102-107.

PEDERSON, R.J., AW. ADAMS, and J. SKOVLIN. 1979. Elk
management in Blue Mountain habitats. Oregon Dep.
Fish and Wildl., Portland. 27pp.

PETERSON, AW. and T.C. GRUBB, Jr. 1983- Artificial trees
as a cavity substrate for woodpeckers. J. Wildl. Man-
age. 47:790-798.

PLATTS, W.S. 1979. Livestock grazing and riparian/stream
ecosystems — An overview. Pages 39-45 in Cope, O.B.,
ed. Proc. of the Forum — Grazing and Riparian/Stream
Ecosystems. Trout Unlimited, Inc., Vienna, VA. 94pp.

. 1984. Riparian system/livestock grazing interaction

research in the intermountain West. Pages 424-429
in Warner, R.E. and KM. Hendrix, eds. California
Riparian Systems: Ecology, Conservation, and Produc-
tive Management. Univ. California, Berkeley. 1035pp.

, KA. GEBHARDT, and W.L. JACKSON. 1985. The

effects of large storm events on basin-range riparian
stream habitats. Pages 30-34 in Johnson, R.R., CD.
Ziebell, DR. Patton, P.F. Ffolliott, and R. H. Hamre,
tech. coords. Riparian Ecosystems and Their Manage-
ment: Reconciling Conflicting Uses. Proc. First North
Am. Riparian Conf. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. RM-120. 523pp.

and R.L. NELSON. 1983- Population and generic

differentiation in the Humboldt cutthroat trout of
Gance Creek, Nevada. Pages 15-19 in California-
Nevada Wildl. Trans.

, O. CASEY, and V. CRISPIN. 1983. Riparian-stream

habitat conditions on Tabor Creek, Nevada, under
grazed and ungrazed conditions. Pages 10-14 in Proc.
Ann. Conf. Western Assoc. Fish and Wildl. Agencies.
63:10-14.

REID, EH. and G.D. PICKFORD. 1946. Judging mountain
meadow range condition in eastern Oregon and east-
ern Washington. U.S. Dep. Agric, Circular 748.

RICE, J., B.W. ANDERSON, and R.D. OHMART. 1984. Com-
parison of the importance of different habitat attri-
butes of avian community organization. J. Wildl.
Manage. 48:895-911.

ROATH, L.R. and W.C KRUEGER. 1982. Cattle grazing
influence on a mountain riparian zone. J. Range Man-
age. 35:100-104.

RUBY, KB., DC. ERMAN, and J.D. NEWBOLD. 1977. Bio-
logical assessment of timber management activity
impacts and bufferstrip effectiveness on National For-
est streams of northern California. Earth Resources
Monogr. 1, U.S. Dep. Agric, For. Serv., Reg. 5, San
Francisco, CA. 170pp.

SALWASSER, H. and K SHIMAMOTO. 1984. Pronghorn,
cattle and feral horse use of wetland and upland habi-
tats. Pages 210-213 in Warner, RE. and KM. Hendrix,
eds. California Riparian Systems: Ecology, Conserva-
tion, and Productive Management. Univ. California,
Berkeley. 1035pp.

SCOTT, LB. 1984. Economic values of three furbearers
inhabiting California riparian systems. Pages 731-738
in Warner, RE. and KM. Hendrix, eds. California
Riparian Systems: Ecology, Conservation, and Produc-
tive Management. Univ. California, Berkeley. 1035pp.

SCOTT, V.E. 1979. Bird responses to snag removal in pon-
derosa pine. J. Forestry. 77:26-28.

, KE. EVANS, DR. PATTON, and CP. STONE. 1977.

Cavity-nesting birds of North American forests. U.S.
Dep. Agric, For. Serv. Agric. Handbook 511, Washing-
ton, DC. 112pp.

SEVERSON, KE. and CE. BOLDT. 1978. Cattle, wildlife,
and riparian habitats in the western Dakotas. Pages
94-103 in Management and Use of Northern Plain
Rangeland. Reg. Rangeland Symp., Bismarck, ND.

SIEKERT, RE., Q.D. SKINNER, MA. SMITH, J.L. DODD,
and J.D. RODGERS. 1985. Channel response of an
ephemeral stream in Wyoming to selected grazing
treatments. Pages 276-278 in Johnson, R.R., CD.
Ziebell, DR. Patton, P.F. Ffolliott, and R. H. Hamre,
tech. coords. Riparian Ecosystems and Their Manage-
ment: Reconciling Conflicting Uses. Proc. First North
Am. Riparian Conf. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. RM-120. 523pp.

SKOVLIN, J.M. 1967. Fluctuations in forage quality on

summer range in the Blue Mountains. U.S. Dep. Agric,
For. Serv. PNW-Res. Paper 44.

. 1984. Impacts of grazing on wetlands and riparian

habitat: A review of our knowledge. Pages 1001-
1 1 04 in Committee on Developing Strategies for
Rangeland Management. Natl. Resour. Council/Natl.
Acad. Sci., Westview Press, Boulder, CO.

, R.W. HARRIS, G.S. STRICKLER, and G.A. GARRI-
SON. 1976. Effects of cattle grazing methods on pon-
derosa pine-bunchgrass range in the Pacific
Northwest. U.S. Dep. Agric, Tech. Bull. 1531. Wash-
ington, DC. 40pp.

W.R. MEEHAN, J.C BUCKHOUSE, and M. VAVRA.

1977. A method of study for determining the influ-
ence of grazing on riparian and aquatic habitats in the
Blue Mountains of Oregon. Pages 164-169 in Live-
stock Interactions with Wildlife, Fisheries, and Their
Environments. Proceedings of the Symposium. U.S.
Dep. Agric, For. Serv. Pacific Southwest For. Range
Exp. Sta., Berkeley, CA.

SMITH, F. 1 977. A short review of the status of riparian
forests in California. Pages 1-2 in Sands, A., ed. Ripar-
ian Forests in California: Their Ecology and Conserva-
tion. Inst. Ecol., Publ. 15, Univ. California, Davis.
122pp.

STAMP, N. and R.D. OHMART. 1979. Rodents of desert
shrub and riparian woodland habitats in the Sonoran
Desert. Southwestern Nat. 24:279-289-

SULLIVAN, B.K 1981. Distribution and relative abundance
of snakes along a transect in California. J. Herpert.

198

Riparian Habitats

15:247-248.

SWANSON, F.J., S.V. GREGORY, JR. SEDELL, and AG.

CAMPBELL. 1982. Land-water interactions: The ripar-
ian zone. Pages 267-291 in Analysis of Coniferous
Forest Ecosystems in the Western United States. US/
IBP Synthesis Series 14, Hutchinson Ross Publ. Co.,
Stroudsburg, PA.

SZARO, R.C., S.C. BELFIT, J.K AITKIN, and J.N. RINNE.
1985. Impact of grazing on a riparian garter snake.
Pages 359-363 in Johnson, R.R., CD. Ziebell, DR.
Patton, P.F. Ffolliott, and R. H. Hamre, tech. coords.
Riparian Ecosystems and Their Management: Recon-
ciling Conflicting Uses. Proc. First North Am. Riparian
Conf. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-120. 523pp.

THOMAS, J.W., C. MASER, and J.E. RODIEK 1979. Wildlife
habitat in managed rangelands — the Great Basin of
southeastern Oregon-riparian zone. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. RNW-80. 18pp.

, R.J. MILLER, H. BLACK, J.E. RODIEK and C.

MASER. 1976. Guidelines for maintaining and enhanc-
ing wildlife habitat in forest management in the Blue
Mountains of Oregon and Washington. Trans. North
Am. Wildl. Nat. Resour. Conf. 41:452-476.

U.S. COUNCIL ON ENVIRONMENTAL QUALITY. 1978.
Environmental quality. The ninth report of the Coun-
cil on Environmental Quality. U.S. Govt. Printing
Office, Washington, DC. (Stock No. 041-011-00040-
8). 599pp.

U.S. DEPARTMENT OF AGRICULTURE, FOREST SER
VICE. 1972. Western regional working conference —
results of work group sessions — delegate meetings.
National Program of Research for Forest and Associ-
ated Rangelands. U.S. Dep. Agric, For. Serv. Inter-
mountain For. Range Exper. Sta., Ogden, UT. 39pp.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF
LAND MANAGEMENT. 1978. Recreational carrying
capacity in the California Desert. U.S. Dep. Inter., Bur.
Land Manage., Sacramento, CA. 115pp.

. 1980. The California desert conservation area plan.

U.S. Govt. Printing Office, San Francisco, CA.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF
RECLAMATION. 1979. Pecos River Basin water sal-
vage project: Final environmental statement. U.S. Dep.
Inter., Bur. Rec, Southwest Reg. Off., Amarillo, TX.

VAN VELSON, R. 1979. Effects of livestock grazing upon
rainbow trout in Otter Creek, Nebraska. Pages 53-55
in Cope, O.B., ed. Proc. of the Forum — Grazing and

Riparian/Stream Ecosystems. Trout Unlimited, Inc.,
Vienna, VA. 94pp.

WARNER, RE., recorder. 1979a. Fish and wildlife resource
needs in riparian ecosystems: Proc. of a workshop.
Nat. Water Resour. Analysis Group, Eastern Energy
and Land Use Team, U.S. Dep. Inter., Fish Wildl. Serv.,
Kearneysville, WV. 53pp.

. 1979b. California riparian study program: Back-
ground information and proposed study design. Plan-
ning Branch, California Dep. Fish and Game,
Sacramento. 177pp.

and KM. HENDRIX, eds. 1984. California riparian

systems: Ecology, conservation, and productive man-
agement. Univ. California Press, Berkeley. 1035pp.

WAUER, R.H. 1977. Significance of Rio Grande riparian

systems upon the avifauna. Pages 165-174 in Johnson,
R.R. and D.A. Jones, tech. coords. Importance, Preser-
vation and Management of Riparian Habitat: A Sym-
posium. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-43. 217pp.

WEBSTER'S NEW COLLEGIATE DICTIONARY. 1979.
G.&C. Merriam Co., Springfield, MA. 1532pp.

WILLIAMS, D.F. 1985. Mammalian species of special con-
cern in California. California Dep. Fish and Game,
Nongame Wildl. Invest., Final Rep., Project E-W-4,
IV-14.1. Sacramento. 184pp.

and KS. KILBURN. 1984. Sensitive, threatened, and

endangered mammals of riparian and other wetland
communities in California. Pages 950-957 in Warner,
R.E. and KM. Hendrix, eds. California Riparian Sys-
tems: Ecology, Conservation, and Productive Manage-
ment. Univ. California Press, Berkeley. 1035pp.

WINEGAR, H.H. 1977. Camp Creek channel fencing-
plant, wildlife, soil, and water response. Rangeman's J.
4:10-12.

WINTERNITZ, B.L. 1973. Ecological patterns in a montane
breeding bird community. Ph.D. dissertation, Univ.
Colorado, Boulder.

1976. Temporal change and habitat preference of

some montane breeding birds. Condor 78:383-393.

and H. CAHN. 1983. Nestholes in live and dead

aspen. Pages 102-106 in Davis, J.W., G.A. Goodwin,
and R.A. Ockenfels, tech. coords. Snag Habitat Man-
agement: Proceedings of the Symposium. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. RM-99. 226pp.
YOUNG, R.A., T HUNTRODS, and W. ANDERSON. 1980.
Effectiveness of vegetated bufferstrips in controlling
pollution from feedlot runoff. J. Environ. Qual. 9:483-
487.

Riparian Habitats

199

10

MARSHES

Milton W. Weller

Department of Wildlife & Fisheries Sciences
Texas A&M University
College Station, TX 77843

Editor's Note: Wetlands compose a small percentage
of lands in the ivestern U.S., but their importance
for wildlife far outweighs their acreage. In this pub-
lication, marshes are considered to be wetlands as-
sociated with standing water. Among the most
complex habitats, marshes are therefore most diffi-
cult to inventory or monitor. As the author empha-
sizes, evaluation, impact assessment, and manage-
ment decisions require first-hand, mud-on-your-
boots experience in the marsh.

INTRODUCTION

"Greater familiarity with marshes on the part of
more people could give man a truer and a more
wholesome view of himself in relation to Nature.
. . . Marshes comprise their own form of wilderness.
They have their own life-rich genuineness and reflect
forces that are much older, much more permanent,
and much mightier than man."

-Paul L. Errington, from Of Men and
Marshes

Marshes, the dynamic meeting places of land
and water, are among the most productive and excit-
ing of all wildlife habitats. Not only do they attract a
great diversity of wildlife, but numbers of some spe-
cies can be overwhelming. As a result, the value of
marshes for recreation such as hunting or bird
watching can outweigh those activities on far larger
tracts of other habitats. Moreover, their values in
damping flood surges, in water purification, and in
water recharge make marshes highly important. In-
ventory and monitoring are essential to identify- criti-
cal habitats and resource values. With this knowl-
edge biologists can develop management alternatives
or mitigation plans on these numerically restricted
but highly valued habitats.

Wetlands are characterized by hydric soils and
water-loving plants (Cowardin et al. 1979). Of the
many diverse types of wetlands, marshes are the
most widely distributed and the best-known form.
They are dominated by emergent plants such as cat-
tail (Typha sp. ), bulrushes (Scirpus sp. ), sedges
{Carex sp. ), and water-tolerant grasses. The general
height and structure of such plant communities dis-
tinguish marshes from swamps (wooded and shrub-
dominated wetlands ) or from bogs ( moss and heath-
covered wet organic soils ). Whereas swamps and
bogs are restricted in distribution by climatic re-
gime, marshes may be found anywhere. They are the
characteristic wetland of intermountain lowlands of
the western U.S. where they form a true oasis for
wildlife, both aquatic and terrestrial.

Marshes often are complete entities, found in
shallow basins. The term may also be used for any
emergent hydrophyte community. The emphasis
here is on shallow basins that are marshes in the
geomorphic as well as biological sense.

Water is the driving force in determining wet-
land type and habitat quality. Water permanency and
associated vegetation are key factors in classifying
wetlands. Because water cycles are variable, marshes
are rarely constant. These fluctuations induce "boom
and bust" in wildlife numbers, but are essential to
nutrient recycling. In shallow marshes such as

Marshes

201

ephemeral prairie potholes, dense vegetation pro-
duced during natural drawdowns is the major food
for detritus-feeding macroinvertebrates (those larger
aquatic invertebrates visible to the unaided eye)
which are in turn important food resources for wa-
terbirds and fish (McKnight and Low 1969; Krapu
1974). In deeper marshes or shallow lakes, drying al-
lows decomposition of bottom organic deposits that
provide nutrients for a new surge of vegetation that
germinates under the low water conditions.

The character of marshes — water, sediment, or-
ganic matter, and dense vegetation — makes walking
or boating difficult. The tendency is to circle around
the area rather than go through it. But evaluation,
impact assessment, and management decisions re-
quire first-hand experience in the marsh. It is hoped
that this chapter will provide the inducement to gain
this experience.

Bureau of Biological Survey employees, 1935.

Because of the dynamics of water and season,
observations of conditions or populations in a single
year have relatively little meaning except when
viewed in relation to the following: ( 1 ) past records
of the same marsh, or (2) a predicted potential
based on knowledge of similar wetlands for which
data are available. Long-term monitoring is vital to
understanding the quality, importance, and condition
of a marsh.

Losses of wetlands occur in many ways. Drain-
age for agriculture is the major cause of loss of
marshes (Frayer et al. 1983). Filling of small marshes
for agriculture and other uses occurs occasionally,
but even large marshes have been filled for urban
and industrial development. Because marshes that
have constant water levels become lake-like, perma-
nent flooding may have serious detrimental impacts
on marsh wildlife (Weller 1981b). The impact of to-
tal loss is obvious. Less measurable negative influ-
ences include modified water flow, increased eutro-
phication or turbidity, and introduction of exotic
animals or plants. These impacts are both long-term
and widespread, but they are difficult to assess with-
out ecological and biological baseline data for mak-
ing comparisons.

A number of wildlife species inhabiting western
marshes are classified as endangered, mostly because
of disappearing habitats. Three subspecies of Clapper
Rail — Yuma (Rallus longirostris yumaensis), Califor-
nia (R. I. obsoletus), and Light-footed (R I. levipes) —
are endangered due to loss of river or coastal marsh.
The salt marsh harvest mouse (Reithrodontomys ra-
viventris) now occurs in only a few scattered
marshes in San Francisco Bay. Wetlands are impor-
tant not only for local wildlife populations but also
for migratory birds, a national and international re-
source. Protection, restoration, and management of
marshes are essential to restore these species.

Yuma clapper rail.

202

Marshes

CLASSIFICATION SYSTEMS AND
DESCRIPTORS FOR MARSHES

Interest in wetlands has been dominated by bi-
ologists studying semiaquatic animal and plant life.
Acquisition of marshes was promoted by waterfowl
hunters, fur trappers, and fishery managers. Different
species of wildlife were associated with different
types of marshes, and efforts to preserve species
through acquisition or management required a
method of choosing different wetlands and assessing
how many of each existed. Hence, biologists devel-
oped classification systems to broadly define and in-
ventory wetlands (e.g., Martin et al. 1953).

A classic document by Shaw and Fredine ( 1956)
outlined 20 types of wetlands across the U.S., de-
scribed their use by wildlife, and estimated the ex-
tent of losses since the earliest surveys. A significant
advancement was made in classification precision by
a detailed regional study in the North Dakota prairie
pothole region (Stewart and Kantrud 1971 ). This
system is strongly influenced by data on water per-
manence and vegetation patterns. The system de-
vised by Millar ( 1979) for western Canada also used
data on basin, area, watershed, and other limnologi-
cal variables to develop a precise classification sys-
tem that could be used when wetlands were dry or
devoid of vegetation. Data on wildlife generally have
not been used as a means of defining or clarifying
types of wetlands except by Bergman et al. ( 1977)
for Alaskan tundra wetlands.

With the advent of widespread interest in wet-
land values for food chain support systems, flood and
erosion control, and energy and sewage treatment, a
broader classification system was needed. An inter-
agency task force, headed by the U.S. Fish and Wild-
life Service, devised a system ranging from the gen-
eral to more and more specific. Known as the
"Classification of Wetlands and Deepwater Habitats
of the United States" (Cowardin et al. 1979), it is
being used for the National Wetland Inventory. It is
hierarchical, like a taxonomic key. All wetlands are
first placed into one of five "systems" by means of
ecological terms of broad usage:

Palustrine

vegetated wetlands such as
marshes

Estuarine

brackish water which may include
marshes or swamps

Lacustrine

lake-like

Riverine

stream-related but not including
marshes or oxbow lakes

Marine

deepwater habitats

These systems are then divided into subsystems
except for Palustrine wetlands. The Palustrine Sys-
tem is divided into units called Classes which are
based on dominant vegetative type. A marsh is
termed an Emergent Wetland as opposed to a Moss-
lichen Wetland (bog) or a Forested Wetland
(swamp). Further subdivisions allow more and more
precise descriptors, so that Emergent Wetlands can
be divided into subclasses by the presence of persist-
ent (lasting over winter or more) plants such as cat-
tail or bulrushes, to the nonpersistent (rapidly de-
composing) plants such as arrowhead (Sagittaria
sp.) or dock {Rutnex sp.). The precise plant species
or community is referred to as dominance type,
which allows a rather precise taxonomic description
of the community. However, the habitat significance
of these descriptors is not evident from the
classification.

On a local basis, some of the earlier classifica-
tion systems are still in use and may be extremely
helpful in wildlife evaluations. They should be used
as an addition to rather than replacement for the
interagency system.

For the western U.S., especially the glaciated
pothole region, the Stewart and Kantrud (1971) sys-
tem is useful because of the use of common terms
related to hydroperiods. Classes I through V reflect
pattern of flooding or mean water depth and are self
explanatory: Ephemeral, Temporary, Seasonal,
Semipermanent, and Permanent. Special wetland
classes include VI Alkali ponds and shallow lakes
rich in submergents and VII-Fens (or alkaline bogs).
Subclasses were used to denote five ranges from
fresh to subsaline.

Because wetlands are so dynamic in water level,
a wetland of the same class can range from bare mud
flat to emergent vegetation through the action of
water levels and herbivores (Weller and Spatcher
1965; Weller and Fredrickson 1974). Special termi-
nology reflecting those short-term successional
stages was used by Stewart and Kantrud ( 1971 ) to
assist in detailed descriptions: emergent phase, open-
water phase, drawdown phase, and bare-soil phase —
the last two inducing germination conditions that

Marshes

203

ultimately result in reestablishment of emergent
vegetation.

Another set of descriptive terms denotes the
typical combination of plants found under mean pre-
vailing water conditions as well as topography. Mil-
lar's ( 1979) system uses traditional terms for vegeta-
tion zones of large wetlands, or for classifying a
marsh by the vegetation of the central part of the
marsh:

1. wet meadow

2. shallow marsh

3. emergent deep marsh

4. open water marsh

5. shallow open water

6. open alkali wetland

7. disturbed areas (due to drought or agricultural
activities).

A wet-meadow community is made up mostly of
nonpersistent annuals and is commonly flooded only
in spring. A deep marsh is dominated by those per-
sistent perennials in the center that do well with
regular flooding, but the edge will have species char-
acteristic of wet meadow or shallow marsh.

Most classification and inventory systems deal
with one basin or a part of one basin. Yet values for
wildlife and fish are enhanced by the proximity of
several different wetlands. The importance of conti-
guity has not been documented, but birds and mam-
mals move from area to area. Adult fish that move
from lakes to wetlands to spawn, or fry that mature
in nursery marshes, must have connecting links (For-
ney 1968). Such relations sometimes can be de-
tected from aerial photographs or detailed maps, but
the extra value to be placed on such spatial relation-
ships is uncertain. Data need to be gathered and ana-
lyzed from this perspective because acquisition, pro-
tection, and management of marshes would be
strongly influenced by this information.

One of the most difficult descriptive decisions of
wetlands is the demarcation of the wetland bound-
ary. Marsh edges often grade imperceptibly into the
uplands, and plants at the marsh edge may be toler-
ant of both wet and dry conditions. Moreover, the
hydrologic influence of a wetland may affect upland
vegetation since increased water availability near the
wetland may encourage or maintain terrestrial vege-
tation that might otherwise not exist.

Coastal wetlands may have the additional com-
plicating variables of lunar tidal regime, wind and
storm surges, and salinity. These too are dynamic
processes with tides varying seasonally, and fresh-
water inflow varying with rainfall. At times, estuarine
marshes may be totally fresh due to high rates of in-
flow, whereas marine storm tides may modify and

even reverse the situation periodically. As a result, a
much reduced flora dominated by grasses such as
salt marsh cordgrasses {Spartina sp.) and saltgrass
{Distichlis spicata) characterizes vast areas of
coastal marsh. Mean tide ranges have produced
zones of high ( less commonly flooded ) and low
(wet) marsh based on the relative degree of flooding
and the influence of salt water (Nixon 1982; Zedler
1982).

MAJOR SPECIES GROUPS

The availability of water, diverse plants, and rich
invertebrate life make marshes ideal for many spe-
cies of wildlife and fish. Some species evolved with
more permanent water and are equipped with gills
for total water emersion throughout life (fish and
some salamanders ) or for part of their life cycle
(most salamanders, frogs, and toads). Others are
semiaquatic, being air breathers throughout life as a
secondary adaptation to fluctuating water regimes
found in marshes. Some species have highly modified
feet (beaver, Castor canadensis, and river otters, Lu-
tra canadensis) or flattened tails for swimming
(muskrats, Ondatra zibethicus). Some avian species
have long legs for wading (herons and storks), or
adaptions for swimming and diving (ducks and
grebes ). Still other species retain greater range of
habitat by means of less inhibiting adaptations to
water, but are common around marshes ( mink, Mus-
tela vision, and raccoons, Procyon lotor). All de-
grees of variation are found between these extremes.
Some species frequent marsh areas and exploit their
resources without being intimately associated. Swal-
lows and bats, for example, catch insects there but
do not use marshes for nesting or cover. Some ter-
restrial species may invade during drawdown stages
to seek out stranded invertebrates or fish. Red foxes
(Vulpes vulpes) take advantage of muskrats unpro-
tected by water or lodges.

A review of some common taxonomic groups
will facilitate a view of the more complex communi-
ties of vertebrates common to marshes of the west-
ern U.S. For further detail, see Weller ( 1979a).

Fish

Considering both freshwater (about 95) and es-
tuarine (about 109) species, wetlands serve over
200 species on a national basis (Adamus 1983)- Inte-
rior palustrine wetlands may have several species or
none at all. The most abundant are small minnow-
like fishes that feed on tiny invertebrates, such as
mosquito fish (Gambusia sp. ). Other species include
herbivores or omnivores like crappie (Pomoxis sp.)
and bluegill (Lepomis macrochirus), herbivore-detri-
tivores like carp (Cyprinus carpio); and carnivores
like walleye (Stizostedium vitreum), pike (Esox In-
dus), and largemouth bass (Macropterus sal-

204

Marshes

moides ). The exact complement varies with size and
permanence of the water. Association with source
streams may contribute to pioneering. Oxygen short-
age is perhaps the most important limiting influence
because marshes may be oxygen depleted during the
summer.

Amphibians

Dominant amphibians range from the most
aquatic sirens {Siren sp. ) that have gills throughout
life, to frogs that breathe air as adults. Some frogs
stay in or near water always (bullfrogs, Rana cates-
beiana ). Others stay near water but feed in wet
meadows and grassy areas (northern leopard frogs,
Rana pipiens ). A few are highly terrestrial species
that associate with moist forest most of the year
(tree frogs, Hyla sp. ). The most terrestrial amphibi-
ans are toads (Bufo sp. ) that range in habitat from
moist uplands to near-desert. Amphibians are impor-
tant as food for various fish, reptile, avian, and mam-
malian predators.

Reptiles

Although the species vary locally, turtles are the
most widespread and often abundant reptilian preda-
tor or omnivore in marshes. Some like the musk tur-
tles (Sternotherns sp. ), mud turtles (Kinostemon
sp), cooters and sliders (Pseudemys sp. ) loaf on logs
or vegetation in the sun and are conspicuous. A few
are true marsh turtles, found nowhere else: Bind-
ing's turtle {Emydoidea blandingii) and chicken tur-
tles {Deirochelys reticularia). Others like the com-
mon ( Chelydra sp. ) and alligator snapping turtles
(Macroclemys temminckii) may be inconspicuous
until they move into uplands during egg-laying or
are exposed during droughts. Snakes, especially
water snakes of the genus Nerodius, are widespread.
The massasauga (Sistrurus catenatus) occurs in
either swamps or wet prairies. Some species like rat
snakes (Elapha sp. ) and garter snakes ( Thamnophis
sp. ) go into any habitat, including marshes, for birds
and their eggs.

Flycatchers
Warblers

Geese •*

iVxi Dabblers
. ^ Sparrows ^
W ^

Blackbirds
(Red Winged) (Yellow-Headed)

Redheads, Ruddy Ducks
Swans

-+■ Mergansers

*WWl

Sedge) (Marsh)

■^w Wrens

Gulls, Terns

Rails
Egrets - Herons
Bitterns
Short-Eared Owls

Gallinules
Snail Kites

Moorhens, Coots
Grebes

MUD FLAT

Harriers
-Cranes
Shorebirds SniPes

Loons, Pelicans

Shorebirds
Waders

WET
MEADOW

SHALLOW
MARSH

UNVEGETATED
SHALLOWS/
FLATS

Figure 1. Waterbirds of wetlands — with emphasis on habitat selection for breeding and feeding.

Marshes

205

Birds

Over 1 40 species of birds are associated with
wetlands intimately or casually (Adamus 1983). Like
fish, but with still greater flexibility, birds may use
marshes only part of the year to fulfill certain life
functions. These include summer breeding and feed-
ing sites (see Figure 1 ), molting sites (ducks), feed-
ing sites away from the nest (pelicans and waders),
migration stops (shore birds, ducks, and geese), food
exploration zones (most waders moving north after
breeding), and wintering (all migratory species). Al-
though breeding sites tend to be valued more highly,
each area plays a vital role in the life cycle of mobile
species. All are essential.

Mammals

Compared with other classes of vertebrates, few
mammals are semiaquatic and very few are adapted

to marshes. Lack of species numbers is compensated
for by numbers of individuals. Several small rodents
characteristic of damp areas swim easily and do not
avoid water or flooded areas: marsh rice rat (Oryzo-
mys palustris) in Gulf Coast marshes (Esher et al.
1978), and California vole (Microtus californicus) in
tidal marshes (Fisler 1961). Both are omnivorous.
Beavers may create their own marsh by damming of
streams, but also use large marshes and build lodges
of willows and cattail.

Communities of Vertebrates

Species associations and numbers are influenced
by ( 1 ) plant structure, ( 2 ) size and diversity of the
unit, and (3) water depth and permanence. Hence, a
wet meadow dominated by nonpersistent, temporary
plants and water cover is likely to be used for breed-
ing by smaller species requiring only low cover
(sparrows, mice, rats, shore birds, rails) or by tran-
sients using the area for food (muskrats, shorebirds,

HABITAT LAYERS

REPRESENTATIVE GUILDS

F B F B F B

Midstory

Terrestrial surface

Bottom of water column

F - Feeding

B- Breeding (Nesting]

1 2 3

Mallard Redhead Muskrat

Pintail

Canvas-
back

Ruddy
Duck

Relationship between habitat layers and guilds.

206

Marshes

waders, dabbling ducks ). Shallow marshes may har-
bor nesting king rails (Rallus elegans), purple gallin-
ules (Porpbyrula martinica) or common moorhens
(Gallinula chloropus), and species of blackbirds
that vary by geographic area. More persistent and
taller vegetation of deep marshes support larger spe-
cies that need cover, support for nests, and the foods
induced by more permanent water (diving ducks,
swans, cranes, some geese, American coots [Fulica
atra], bitterns, and wader colonies).

Because many vertebrates may use the general
structural layer of vegetation, ecologists often use
the term guild to reflect taxonomically unrelated
species that use the same habitat in the same way
(Short and Burnham 1982). In marshes, bird nests
may be built at three layers: ( 1 ) in tall and robust
emergents, (2) in short and nonpersistent emergents,
and ( 3 ) at the water's surface. For feeding, birds may
forage in all three of the nesting layers, and divers
may seek foods in the water column or on the
bottom.

Surface feeding birds that specialize on foliage
include gadwall (Anas strepera) and American wid-
geon (Anas americana). Northern shovelers (Anas
clypeata) can feed in the open because they can
strain plankton or in shallow water where they may
eat snails or seeds. Phalaropes prey on plankton
forms by swimming and picking. Blue-winged teal
(Anas discors), however, are typical of many dab-
blers that have generalized bill structure and food
habits. These species can take invertebrates in sum-
mer when hens need protein for eggs and young
need it for body growth (Swanson, et al. 1979), or
seeds and foliage in nonbreeding periods. Predators
include mergansers, grebes and loons, and others
that feed on fish in open marshes. Mink may occur
around or in any of these wetland types, but dens
are most common by semipermanent water, presum-
ably because of the food supply.

CRITICAL HABITAT FEATURES

To appreciate the features of a marsh that make
it valuable to wildlife, it is essential to relate these
components to wildlife needs. Regardless of the cli-
matic conditions or the season, wildlife often seek
shelter from sun, wind, rain, and hail. Cover also is
essential for protection from predators. No wonder
predatory mink and weasels are long-bodied, and
aerial predators like the northern harrier (Circus cy-
aneus ) hover to follow and catch their prey.

The concentrated nutrients of the marsh pro-
duce high primary productivity in plants. The sec-
ondary' production of invertebrates produces an ar-
ray of potential food exceeded in few if any habitats.

The major herbivores in marshes are mammals

like muskrats which not only eat leaves and tubers of
various emergent plants but build lodges for winter
use, thereby cutting vast quantities of emergents.
Geese, especially Snow Geese (Anser coerulescens),
also may utilize tubers and rootstocks in wintering
areas. Swans, widgeons, gadwalls, and coots devour
huge quantities of submergent vegetation. Addition-
ally, many bottom (benthic) invertebrates utilize de-
composing plant matter, or feed on tiny organisms
like bacteria and fungi that live on old stems. Inver-
tebrate abundance must be directly related to plant
productivity at an earlier stage, but there has been
little effort at quantitating such energy flow in fresh-
water marshes.

Marsh vegetation is vital for many species of
birds that nest — often colonially — in marshes: ibis,
egrets, herons, grebes, ducks, gulls, etc. Diverse veg-
etation, whether it be for food, cover, or nest sub-
strate, clearly influences the potential richness of
species found in the system by its various heights,
coverage, foods potential, and food-chain support for
other lower organisms.

Water Chemistry and Plant Communities

Water that enters a marsh is a product of pre-
cipitation and hydrologic processes. Much of the
water may enter the system from surface runoff and
slow moving underground flows at various depths.
The timing of water availability (hydroperiod) and
depth determine the character and type of marsh
through influence on chemical characteristics of the
water and subsequently on vegetation and inverte-
brates used by wildlife.

Typically, rainwater carries with it few dissolved
chemicals except where it collects wind-driven soil
particles or salt spray. Water that flows over soil or
rock strata or through porous substrates may dis-
solve and carry a wide range of dissolved solids
which may serve as plant nutrients or, in extreme in-
stances, be so concentrated as to be toxic to life.
Thus, general characters of the water in marshes and
lakes reflect the character of the watershed as shown
by Moyle (1945) in Minnesota and Metcalf (1931)
in North Dakota. Moyle's work spanned more diverse
water chemistry and plant types, which he classified
as ( 1 ) soft waters (low dissolved salts) of northeast-
ern Minnesota Laurentian Shield country, (2) hard
water morainic lakes of central and southern areas,
and ( 3 ) alkali lakes of western Minnesota. He
pointed out the plant community that reflects each
of these types is distinctive (Table 1 ), but it is a
product of individual tolerance to chemical charac-
teristics, as shown by Bourn (1935) and others on
salt tolerance. Significant measured features were to-
tal alkalinity, sulfate ions, and hydrogen ions (pH).

Soft water flora are limited due to poor nu-
trients and are characterized by underwater leaves

Marshes

207

Table 1. Some examples of lake waters' influence on aquatic plants of Minnesota (selected from Moyle 1945).

Soft water: alkalinity < 40 ppm
sulfates < 5 ppm
pH < 7.5

► reflect poor nutrients

Lobelia dortmanna Scirpus subterminalis
Sparganium sp. Eriocaulon septanquatare
Lsoetes braunii Sagittaria latifolia

Hard water: alkalinity < 1 50 ppm
sulfates < 40 ppm
pH 8.0-8.8

► richer, attractive to
waterfowl

Chara sp. Ceratophylum demersum
Phragmites communis Lemna minor
Potamogen pectinatus Myrsophyllum exalbescens
Scirpus validus Lemna trisulca
Potamogen natans Nuphar variegatum
Scirpus acutus Spirodela polyrhiza
Elodea canadensis Numphaea tuberosa
Typha latifolia Utricularis vulgaris

Alkali water: sulfates < 50 ppm
pH 8.4-9.0

► reduced species, but
dense growth

Ruppia sp. Juncus torreyi
Typha augustifolia Scirpus paludosus
Najas marina Potamogeton pectinatus

and rosettes. Floating plants are rare. Hard water
floras are richer, more abundant, and are attractive
to waterfowl. Alkali floras have reduced species but
often have dense growth of a few well-adapted
species.

Stewart and Kantrud ( 1972) found that specific
conductivity measured dissolved solids simply and
reliably to indicate alkalinity or salinity of prairie
marshes. Means ranged from fresh (about 300 mi-
cromhos per cm at 25° C) to brackish (6,300) to sa-
line (37,500) in their North Dakota ponds. Plant spe-
cies composition in relation to conductivity readings
are reproduced in modified form in Figures 2
through 5 for wet meadow, shallow marsh, and deep
marsh communities. Despite broad ranges of toler-
ance, common emergent hydrophytes tended to be
associated with the following mean dissolved solids
measured by Stewart and Kantrud (1972): Slender
bulrush {Scirpus heterochaetus) 350 mh/cm, cattail
1000 mh/cm, and alkali bulrush (Scirpus paludosus)
3,500-32,000 mh/cm (Figure 6).

Salinity is clearly one of the more important
chemical factors of coastal and certain western
marshes. For example, the effects of salinity in estab-
lishment and growth of widgeongrass (Ruppia mar-

tina) are well known (Bourn 1935; Joanen and Glas-
gow 1966). Inland saline waters support many salt-
tolerant species (halophytes), and their salt concen-
trations strongly influence plant species composition
and growth as shown by Rawson and Moore (1944).
Experimental studies of vegetation in the Great Salt
Lake basin and efforts to freshen such areas to in-
duce more suitable plants for waterfowl have pro-
vided excellent insights into plant species and water
quality influences in these extreme conditions (Nel-
son 1954; Christiansen 1970). Additionally, observa-
tions of marshes and marsh hydrophytes associated
with warm, salt springs in Utah provide excellent
data on species composition and the role of water ta-
bles in salt deposition so important to wetland plants
(Bolen 1964).

Key to Figures 2 thru 5:

Width of bars indicates relative abundance.

Thick bars = frequently common or abundant.

Thinner bars = common, occasionally abundant
to occasionally fairly common.

208

Marshes

PRIMARY SPECIES:

False-aster

Boltonia latisquama

Fowl bluegrass

Poa palustris
Silver sedge

Carex praegracilis

Lowland white aster

Aster simplex

Northern reedgrass

Calamagrostis inexpansa

Prairie cordgrass

Spartina pectinata

Baltic rush

Juncus balticus
Foxtail barley

Hordeum jubatum
Inland saltgrass

Distichlis stricta

1

t
1

Fresh

Slightly
brackish

Moderately
brackish

Brackish

Subsaline

1

1

1

,

1-

1

-

1

i

1

1

Figure 2. Characteristic plant species of wet meadows arranged from fresh to salt water tolerance.

PRIMARY SPECIES:

Fresh

Slightly
brackish

Moderately
brackish

Brackish

Subsaline

Tall mannagrass

Glyceria grandis

Giant burreed

Sparganium eurycarpum

American waterplantain

Alisma triviale

Marsh knotweed

Polygonum coccineum

American sloughgrass

Beckmannia syzigachne

Awned sedge

Carex atherodes

Whitetop rivergrass

Scolochloa festucacea

1

Common spikerush

Eleocharis palustris

1

Narrowleaf waterplantain

Alisma gramineum

1

American bullrush

i i

Scirpus americanus

i 1

Nuttall alkaligrass

_i I

Puccinellia nuttalliana

1 ■

1

Samphire

Salicornia rubra

1

Figure 3. Characteristic plant species of shallow marsh emergent vegetation.

Marshes

209

PRIMARY SPECIES:
Slender bulrush

Scirpus heterochaetus

River bulrush

Scirpus fluviatilis

Hybrid cattail

Typha latifolia x angustifolia

Tule bulrush

Scirpus acutus

Alkali bulrush

Scirpus paludosus

SECONDARY SPECIES:
Softstem bulrush

Scirpus validus
Common cattail

Typha latifolia
Common reed

Phragmites communis

Narrowleaf cattail

Typha angustifolia

Fresh

Slightly
brackish

Moderately

brackish

Brackish

Subsaline

,

1

Figure 4. Characteristic plant species of deep-marsh emergent vegetation.

210

Marshes

PRIMARY SPECIES:
Aquatic liverwort

Riccia fluitans
Grassleaf pondweed

Potamogeton gramineus

Common waterstarwort

Potamogeton gramineus
Aquatic liverwort

Ricciecarpus natans
Common marestail

Hippuris vulgaris
Baby pondweed

Potamogeton pusillus
Star duckweed

Lemna trisulca
Common bladderwort

Utricularia vulgaris
Richardson pondweed

Potamogeton richardsonii
Common hornwort

Ceratophyllum demersum

Spiked watermilfoil

Myriophyllum exalbescens
Aquatic moss

Drepanocladus sp.
White buttercup

Ranunculus trichophyllus
Common duckweed

Lemna minor
Western widgeonweed

Ruppia occidentalis
Horned poolmat

Zannichellia palustris
Muskgrass

Charu sp.
Fennelleaf pondweed

Potamogeton pectinatus
Common widgeonweed

Ruppia maritima

Fresh

Slightly
brackish

Moderately
brackish

Brackish

Subsaline

Saline

1 1

1

3

| 1

1 1

1 1 I
i 1

1 1 1

i

1 1 1

1 i

Figure 5. Characteristic plant species of submerged and floating vegetation.

Marshes

211

PLANT COMMUNITY
(in order of increasing
specific conductance)

Scirpus heterochaetus

Scirpus fluviatilis

Typha sp.

Typha sp.

— Scirpus acutus

Scirpus acutus

Scirpus acutus

— Scirpus paludosus

Scirpus paludosus

— Scirpus acatus

Scirpus paludosus

SPECIFIC CONDUCTANCE IN MICROMHOS PER CENTIMETER AT 25 DEGREES CENTIGRADE
100 316 1000 3160 10,000 31,600 100,000

1

EXPLANATION

Geometric mean

Range
± Standard Deviation

Standard Error of the mean

2.00 225 2.50 2.75 3.00 3.25 3.50 3 75 4.00 4.25 4.50 4.75 5.00

SPECIFIC CONDUCTANCE (LOG ,0 SCALE)

Figure 6. Relationship of dominant deep-marsh vegetation to specific conductance.

Many water characteristics are influenced by
processes within a basin. These include the years of
alternate flooding and drying which may concentrate
deposits, the rate of water flow through the system
that carries away dissolved materials (including nu-
trients), and the processes of storage and decomposi-
tion of dissolved chemicals. Marshes also vary sea-
sonally in concentrations of dissolved solids.

Community Structure

Life Form and Zonation. Wildlife of marshes re-
spond visually to plant structure and by trial and er-
ror to the associated cover and food. Diversity of
structure is produced by various species of plants
adapted to marshes and also by zonation resulting
from basin configuration.

Plants of marshes have Life forms, growth, and
reproductive strategies that result from adaptation to
water depth and hydroperiod. Their mechanisms of
germination, growth, and spread also influence how
they are used as food and cover by wildlife. The
dominant life form, rooted emergents, represents the
easiest adaptation of hydrophytes to the semiaquatic
environment. Seeds typically germinate on mud flats
or in very shallow water, deriving their nutrients

( 1 ) emergents

212

Marshes

from soil or organic substrate. Emergents arc di-
verse and complex. The continuum can be subdi-
vided several ways. Nonpersistent, soft and often
short grasses, sedges, and forbs decompose usually in
less than one growth season. Larger, more robust,
and persistent emergents like cattail, some bulrushes,
and reeds may stand for several years. This category
normally does not include woody plants like willow
(Salix sp. ) which are classified as shrubs or trees.

Still more truly aquatic plants are the submer-
gents, most of which are rooted. They have much
dissected or linear leaves efficient at tapping reduced
sunlight at considerable depths. Such plants, like the
pondweeds (Potamogen sp. ), may have flowers and
seedheads that reach the water's surface. Others like
stonewort {Cham sp. ) are bottom plants with totally
submerged fruiting bodies.

Plants adapted to deeper water may have float-
ing leaves rising from long stems resting on the
water's surface ("pad" plants). They reproduce by
means of large seeds or well-established tubers (ex-
ample: water lily ). Such floating-leaf plants can tol-
erate considerable water fluctuation in spite of their
general preference for deeper water.

The most aquatic plants are floating plants that
may have roots dangling in the water for nutrients,
such as duckweed (Lemna sp. ) or the introduced
water-hyacinth (Eichornia crassipes). Another spe-
cies of Lemna, star duckweed (L. trisnlca), drifts in
the water column like the many species of micro-
scopic plankton.

Marshes are characterized by concentric zones
of vegetation, often intergrading or changing in a
continuum. Sometimes they change abruptly, which
adds diversity of structure horizontally as well as
plant species richness. Thus, a deep-marsh which re-
sults from semipermanent water has a shallow-marsh
zone around the perimeter. It may grade into a wet-
meadow or damp situation where water occurs in
spring but dries out in early summer (Figure 7). The
shallow-marsh or wet-meadow plants are those one
finds in smaller basins which hold shallow water
only temporarily in spring, but these lack the vegeta-
tive diversity in structure and layering found in a
deep marsh with gradual shorelines (Figure 7).
Nonetheless, some concentric zonation is expected
in any marsh, providing a horizontal change in struc-
ture which also induces vertical or layering diversity.
Few marsh basins have perfect symmetry, and other
diversity of pattern is created by flow-through chan-

Life forms of marshes: (1) emergents, (2) submergents, (3) floating-leaf plants, (•*) floating plants.

f?a&* £>tv4t*'f(d-

(2) submergents

(3) floating-leaf plants

(4) floating plants

Marshes

213

A. Three examples of small marshes with water regimes varying from ephemeral to shallow to deep,
resulting in: wet-meadow, shallow-marsh, and deep-marsh vegetation.

,~^^Alk^ljlW), ,\lUr\)W~ ••-*-«

&-.-' ?1

B. Typical deep marsh displaying zonation with deep-marsh vegetation in center, shallow-marsh
peripherally, and wet-meadow at the edge.

C. Open-water marsh resulting from loss of vegetation (due to continued high water or muskrats).

■U i III. I .1 yij

D. Germination phase at low water with a mix of annuals and perennials, persistent, and
nonpersistent plants.

•j**#m**ulM«^r.\^J)\j(,M

ikuJk

ii i i/i

^wmmmmmmwJl^d

Figure 7. Common vegetational patterns of marshes.

214 Marshes

Concentric zones of marsh vegetation.

nels, islands of soil, clumps of trees, or other unique
vegetation established at a lower water level.

Patterns of Vegetation and Water. The above com-
ments might suggest that the more abundant and di-
verse the vegetation, the better; but this is not al-
ways true for breeding birds. Many bird species that
are attracted to vegetation also need open water for
landing and swimming. Openings in emergents cre-
ate edge, providing access to vegetation. Openings
also allow sunlight penetration essential to submer-
gents and some animals. Thus, interspersion of cover
and water is best. Various species respond to differ-
ent cover-water ratios. Weller and Fredrickson
(1974) showed an increase in species richness of
marsh birds and numbers of individuals with increas-
ing numbers of openings or pools in nearly pure cat-
tail up to the point where too little vegetation re-
duced nest sites and food resources. Various cover-
water ratios supported the idea that ducks were
more attracted to 50-50 cover- water patterns (Ka-
minski and Prince 1981; Murkin et al. 1982), al-
though the reasons are more complex than earlier
thought.

Marsh Size and Wetland Complexes

Wetland size is an important influence of the
presence or absence of wildlife species and richness
only because larger areas tend to have more habitat
types (i.e., diversity). Certain birds (red-winged
blackbirds Agelaiiis phoniceus, blue-winged teal, and
some rails) are well adapted to small units or are
ubiquitous, and use these as well as perimeters of
large marshes. Other species (ruddy ducks, Osyura
jamaicensis, yellow-headed blackbirds, X. xanthoce-
phalus) occur mainly in larger, deeper areas with
persistent emergents and submergents. Thus, clusters
of diverse wetlands may function like larger and
more complex wetlands in having more bird species
in the community (Smith 1971; Weller 1981a).

Marsh Succession and Vegetation Dynamics

Because marshes are shallow, dynamic, water
bodies, they undergo changes in depth, chemical
characteristics of the water, vegetation (species, den-
sity, and coverage), and associated wildlife. Water re-
gimes over long periods dictate most or all of these

Marshes

215

changes; and the presence, richness, success, and du-
ration of the plant community is in response to this
regime. Hence, it is vital to have records of water
depth and extent of water coverage of the marsh
bottom, and to document seasonal or year-to-year
variation in hydroperiods to interpret responses by
vegetation and wildlife. Moreover, the seed bank of
long-lived seeds is a product of this water history
(Van der Valk and Davis 1978). The lack of vegeta-
tive response in a management setting may mean
that seeds are lacking (Pederson 1981 ). Logically,
shallow or temporary marshes have seeds of pioneer-
ing, annual, or short-lived plants. They are adapted to
quick response and maturation that enhance the
seed bank. Ducks and other seed eaters respond to
this stage as well, but sufficient seeds survive that
the species is not eliminated locally. When condi-
tions such as mud flats are again available, these
seeds will germinate. In deep, open waters domi-
nated by submergents, there may be fewer seeds of
emergents, although this situation varies by site.
More importantly, terrestrial sites flooded artificially
may lack the seed bank entirely, and establishment
of vegetation may take several years.

The usual pattern in natural shallow marshes is
that diverse seeds of annuals and perennials are pres-
ent. With natural drought at various times of the
year, different seeds respond (Van der Valk 1981 ).
Harris and Marshall ( 1963) showed that early draw-
down with fairly moist soil induces germinations of
persistent, perennial emergents (e.g., cattail). Later
and drier drawdown conditions may result in quick-
maturing and short-lived annuals like smartweed.
They still produce seeds that survive over winter —
and often for many years — until conditions are
suitable (see Weller 1981a).

If water levels in subsequent years remain low,
vegetation tends to be annual plants or a few peren-
nials characteristic of shallow marshes: sedges,
rushes, and softstem bulrush. If the water level in-
creases, it is the perennials that prosper. Cattail, tule
bulrush, common reed, and other plants tolerant of
deep water survive. In addition to the development
of masses of seeds, some floating-leaf and submergent
plants reproduce by tubers or bulbs. They store en-
ergy in one year that allows vegetative growth in the
next season — before the availability of suitable seed
germination substrate such as mudflats. These peren-
nials may go for years without germination of seeds,
as an influence on the local population of the plant.
However, seeds spread by wind or birds may influ-
ence the establishment of these species in other
areas.

Several studies have been cited demonstrating
the impacts of successional changes on wildlife of
single marshes. Because of long-term water cycles or
random changes, wetland habitats in large regions
are sometimes adversely affected for wildlife. Yeager

and Swope ( 1956), Smith ( 1971 ), and Pospahala et
al. (1974) all observed major population responses
to droughts. Several workers have shown how water-
fowl populations in wetter or more stable areas in-
crease when drought strikes the interior grassland
marshes (Derksen and Eldridge 1980).

Macroinvertebrates

Diverse and often abundant invertebrates pro-
vide needed protein for reproduction and growth of
many fish, amphibians, birds, and some mammals. In-
vertebrates occupy every habitat in marshes and ex-
ploit every food resource. They have evolved life his-
tory strategies attuned to dynamic water regimes,
seasonal temperature cycles, and differing salinities
and acidities. Life cycles are intimately tied to water
regimes. Marshes characterized by deeper and more
constant water levels will have populations of dra-
gonflies and other long-lived insects that may require
several years to mature. Shallow and more temporary
bodies have annual crops of invertebrates (fairy
shrimps, mosquitoes, some midges) that hatch from
drought-tolerant eggs and mature rapidly, to leave
eggs for another time (Wiggins et al. 1980).

Population fluctuations among invertebrates are
dramatic. The emergences of annual crops of insects
like midges and caddisflies, or the "blooms" of plank-
tonic crustaceans such as Daphnia or Cyclops repre-
sent discrete seasonal energy flow patterns as well as
reproductive cycles. The abundance and species
richness of invertebrates vary with marsh type or
successional stage (Voigts 1973), and are influenced
by nutrient availability and shifts in food abundance.
Vertebrate consumers respond promptly. An emer-
gence of midges in mid to late summer feeds fish,
amphibians and ducks in the water, ducks and terns
at the water's surface, plus gulls and swallows in the
air column up several hundred feet. Plankton blooms
are exploited by shovelers with a straining bill, phal-
aropes that pick out individual targets, and fish and
amphibians in the water column. The predators and
scavengers too must be adaptive, because emer-
gences or blooms may last a few days to a few
weeks, and other foods may appear that require dif-
ferent feeding methods.

216

Marshes

Wetland biologists now recognize the vital role
that macroinvertebrates play in dictating the pres-
ence or absence of wildlife that need these for food.
Some managers are even manipulating water levels
to induce high invertebrate populations to attract
waterfowl and shorebirds. Most biologists who have
worked with invertebrates use them as indicators of
marsh condition, but few techniques have been de-
veloped to monitor populations or interpret such
data for rapid assessments in the field.

INVENTORY AND MONITORING SYSTEMS
FOR MARSH HABITATS
AND THEIR WILDLIFE

As discrete, often small and self-contained eco-
systems, marshes usually are more easily inventoried
than more extensive terrestrial habitats. But evalua-
ting them or monitoring changes can present diffi-
cult problems. First, biologists are often not as famil-
iar with the plants and invertebrates or their use by
wildlife as in terrestrial ecosystems. Second, biolo-
gists lack technology, manpower, and precise evalua-
tion systems. To examine these problems and possi-
ble solutions, I will first define the systems. I will
then outline the ways in which one might want to
use these systems or devise procedures to meet par-
ticular goals.

Inventory

Inventory is the determination of number, size,
and distribution of marshes. Mapping involves the

Wetland maps, such as these, are available from the U.S.
Fish and Wildlife Service.

geographic aspect of location and distribution pat-
terns. The most extensive national wetland inventory
ever attempted is in progress. It is a multiagency
Federal effort to record the number, size, and distri-
bution of all wetlands of the contiguous United
States. Much of the survey is already complete. Some
maps have been published and are available from the
U.S. Fish and Wildlife Service. Other areas are still
being mapped or have not been mapped in detailed
scale. Vegetation zones may be gross or lacking on
these maps. A general summary of wetland changes
over time has resulted from these inventories
(Frayer et al. 1983). They suggest a loss in vegetated
palustrine wetlands (marshes) of 94,770 ha (234,000
a.) per year from the 1950s to the 1970s. The data
on wetland number, area, and modifications thereof
are stored in computer data bases, and more detailed
analysis can be expected in the future.

On a regional or local scale, the inventory pro-
cess may produce greater detail, not only in identify-
ing smaller areas but in perceiving changes due to
dynamic water regimes. Additionally, on-the-ground
vegetation and water depth measurements can be
taken.

On a very gross scale, the number of wetlands
has been related to waterfowl populations over vast
areas of the northern U.S. and Canada. In addition to
the standard annual aerial waterfowl censuses, ob-
servers have recorded the number of prairie wet-
lands to reflect year-to-year climatic conditions. Cor-
relations of mallard population size and distributions
with pond numbers reflect the impact of habitat
quality on breeding conditions for mallards (Pospa-
hala et al. 1974). On a smaller scale, a simple posi-
tive correlation between pond numbers and canvas-
back pairs was noted by Sugden ( 1978) over a series
of years of observations.

Cover Mapping

Cover mapping produces a detailed analysis of
the vegetation patterns found within a marsh or
other wetland. Coupled with these is classification,
which describes vegetation, basin, and water regimes
to place marshes or other wetlands into identifiable
types or classes. Vegetation mapping is a vital first
step in understanding potential for wildlife use. It es-
tablishes baseline information necessary for monitor-
ing. Mappers gain a familiarity with the area that as-
sists in devising census methods or observation sites.
On a national scale, the National Wetlands Inventory
is using aerial photographs almost exclusively with
ground truthing in various wetland types (Montanari
and Townsend 1977). Periodic black-and-white
photo series by the Soil Conservation Service usually
are available for baseline maps. Current color photos
have excellent potential for indicating significant
characteristics in wildlife habitat evaluation (Cowar-
din et al. 1981 ). Local overflights can provide views

Marshes

217

of marshes that add great appreciation and under-
standing as well as an opportunity for photographs.
Out-of-the-window obliques are useful for records of
vegetation and water patterns, but are very mislead-
ing for mapping due to the dramatic distortion that
even a small angle produces. A plane with a "belly"
port is ideal for low-level (9 15- 1,830m; 3,000-6,000
ft) 35mm pictures. Where this equipment is not
available, out-of-the-window verticals can be taken
with an externally-mounted camera (Meyer and
Grumstrup 1978).

A series of aerial photographs and the resulting
maps can be used to demonstrate seasonal or yearly
changes in vegetation and coverage, cover-water ra-
tios, and numbers of muskrat lodges in a given wet-
land. Vegetation surveys are necessary in the marsh
to provide detailed data on species richness or diver-
sity, stem density, water depth, submergent plant dis-
tribution, etc. Moreover, work in the marsh is vital
to understanding wildlife-plant-water relationships.
There is no substitute for a good pair of hip boots or
chest waders!

Monitoring Habitat and Wildlife

Monitoring implies biological assessment with a
series of timed observations for a stated purpose.
Objectives can include assessment of habitat and
wildlife values to prevent habitat deterioration, mea-
surement of fluctuation in populations or vegetation
as natural entities of the system, or establishment of
a baseline for evaluating the consequences of human
activities on wetland quality. All of these activities
are essential for making sound resource decisions.
Ordinarily, the process requires measurement of hab-
itat and wildlife populations. Many of these tech-
niques are covered elsewhere. Here it is important
only to consider the combination of individual tech-
niques to form a "system" that will satisfy your par-
ticular needs or goals. The following are possible fea-
tures of the habitat, and use by wildlife and fish that
might be measured.

Water Depth and Fluctuation. Water level changes
are important to understanding vegetation germina-
tion, growth, and loss. Water levels also affect nests
of birds or litters of muskrats, potential predation by
terrestrial predators, and the survival of duck broods.
Where practical, gauges should read in tenths and
hundredths of a foot in relation to known mean sea
level. This relation is an engineering standard that
has practical mathematical and communication ad-
vantages when working with managed or artificial
impoundments.

Water depth and extent of coverage in a marsh
determines wildlife use and seasonally influences
plant germination, growth, and seed production.
Mud flats may be viewed negatively at times of nest-
ing, but may be beneficial in the long run because
the resulting germination creates emergent vegeta-
tion suitable for nesting in subsequent years.

Vegetation. The species composition, density,
height, distribution, and cover-water interspersion
patterns by water depth are among the features of
the plant community that might be assessed. How-
ever, time and financial constraints prevent measure-
ment of all these details unless the goals of the moni-
toring demand them. Moreover, there is significant
vertical layering in marsh vegetation as a result of
the four dominant life forms of aquatic plants: rooted
emergcnts, floating-leaf, rooted submergents, and
free-floating (surface or water column). There also
are some epiphytic filamentous algaes that can be
abundant, suitable food for fish and ducks, and sub-
strates for macroinvertebrates. Sampling systems
must be suitable for this variety.

Point-counts are especially useful because they
are rapid and can be used to assess all life forms of
plants above, at, and below water level as well as
water depth (Weller and Voigts 1983). However,

218

Marshes

.

they provide only a erude index to density of less
common plants. Additional sampling systems such as
quadrats or releves (Mueller-Dumbois and Ellenberg
1978) may be necessary. Several studies have dem-
onstrated dramatic year-to-year changes in vegetation
that strongly influenced wildlife populations (Weller
and Fredrickson 1974).

Wildlife Census. Investigators censusing wildlife in
marshes use strategies similar to those for terrestrial
wildlife, but are complicated by access and visibility.
This is why aerial census of waterfowl and large
wading birds of open water developed early, and is
used in place of an index to population change
(Henny et al. 1972). Boats, canoes, air boats, and
marsh buggies also have been used when aerial sur-
vey is either financially impossible or the size of the
area is too small. Air boats and marsh buggies are ex-
pensive to buy and maintain, and can only be justi-
fied in large marsh units. Boats often will not go
through emergents or submergents and thus do not
perfect survey techniques. Canoes have the advan-
tage of being quiet, but are slow and sometimes awk-
ward, and like walking, visibility may be impaired.
Hence, biologists often use subsamples, indexes, and
indirect estimates. Certainly no simple technique can
be advised for all areas.

Censusing wildlife in marshes requires prelimi-
nary experimentation because each area differs. Con-
stancy in technique will then provide an index even
when abundance or density figures cannot be ob-
tained. No one technique will work for all birds or
mammals. Fish, amphibians, and reptiles are still
more difficult because of poor visibility, and require
subsampling techniques that will produce different
levels of success.

Fish. Seines are not useful in marshes because
of the vegetation and organic debris. Kushlan ( 1974)
used lift-traps, drop-traps, and bottom nets to get es-
timates of species richness and species composition.
Rotenone and electrofishing also can be useful, but
no technique is equally suitable for all species and
all vegetation types. Submergents are a particular
hindrance. Some minnows and smaller species in
prairie marshes have been studied by trapping
(Payer and Scalet 1978). Much needs to be learned
about census techniques that assess communities of
fish in marshes.

Amphibians. I know of no accurate method of
censusing various kinds of frogs in various habitats.
Bullfrogs are conspicuous at night with a flashlight,
but leopard frogs and small peepers are difficult to
count. Incidence observed per unit time by habitat
category is possibly the best available index.

Birds. I have been well satisfied with counts of
territorial male songbirds on quadrats from observa-
tion towers or raised shorelines. Red-winged and yel-

low-headed blackbirds are especially conspicuous.
Even less visible but vocal species like marsh wrens
can be pinpointed by triangulation and repeated sur-
veys. At least three early morning surveys at the
peak of breeding activity are required for statistical
reasons as well as to compensate for variation in ac-
tivity of birds. In small marshes and ponds where ac-
cess is enhanced by walking on knolls, biologists
were able to survey blue-winged teal and other pairs
of ducks and get fairly consistent results on each of
three annual pair counts (Weller 1979c).

Rails represent one of the most difficult of all
birds to census in any habitat. Except for vocaliza-
tions, and an occasional nest accidentally discovered,
rails can be present in large numbers and remain un-
detected. Some workers have been successful in
stimulating calling by playing back recorded vocali-
zations (Zimmerman 1977). Nest density of larger
species has been recorded as a product of very in-
tensive effort (Tanner and Hendrickson 1956). Least
Bitterns belong in the same difficult-to-census class.

The above examples demonstrate the difficulty
of finding one technique that works for species using
several different strata of vegetation and several dif-
ferent reproductive strategies. The line transect
method of sighting and vocalization is now in fairly
widespread use and holds promise for providing in-
dexes to various species (Mikol 1980). Whether the
method is equally functional in a marsh in early-sea-
son low vegetation and after a full season's growth is
still uncertain.

Nest counts are favored for conspicuous colony
birds such as ibises, egrets, and herons. They have
also been used for ducks where chain drags and
other devices are used to arouse hens from their in-
cubation (Higgins et al. 1969). In some instances,
post-breeding season surveys have been used to cor-
rect for partial counts or counts in dense vegetation.

Mammals. Censuses in marshes have focused
around sign or dens as opposed to visual count,
since many species are nocturnal or hidden in dense
vegetation. Muskrats are censused by numbers of
"lodges" with a fall per-house number of five com-
monly used to estimate totals (Dozier 1948). Num-
bers and home range size of swamp rabbits in south-
east Missouri were derived by driving with men and
dogs and by grid trapping (Toll et al. I960).

Evaluation

Goals of Evaluation. Evaluation of marshes is com-
plex. It involves some method of correlating charac-
teristics of the marsh with a value in which there is
special interest, such as wildlife populations or vari-
ety. Qualities that will attract an endangered species
are also evaluated. There are no well-defined proce-
dures. Biologists are usually dependent on simple

Marshes

219

correlations or associations rather than cause-and-ef-
fect relationships described in quantitative terms.
Dependent upon objectives, such evaluations may be
simplified by comparing two wetlands on qualitative
rather than quantitative terms. Thus, assessing the
quality of a marsh for wildlife is a skill that is still
evolving. Unfortunately, we have a small data base,
and our parameters usually are qualitative rather
than quantitative. Hence, it is imperative that we
clarify the objectives and the parameters of an
evaluation.

Often we limit our interest and evaluation to a
single species (e.g., whooping cranes, as an endan-
gered species) or a group of recreational importance
(waterfowl) or a nuisance group (blackbirds). Using
individuals per hectare (or square mile on vast
areas), or duck-use days, we identify marshes that
are highly productive or much used versus those
with low numbers. If we use a wetland classification
system, we associate number of species and individu-
als with types of wetlands as did Stewart and Kan-
trud ( 1973) in North Dakota. Or we may map the
cover of zones or areas of a single marsh to denote
where nesting occurs, where feeding occurs, etc. We
can then determine how much of any one habitat is
needed to attract and maintain a single species.

There are dangers in species or group-oriented
assessments in that we ignore the community of ver-
tebrates as a whole and bias our effort toward spe-
cies of our choosing. Other species may be of impor-
tance to other persons for other reasons. Another
danger rests with a comparative approach of choos-
ing between two marshes because two wetlands in
the same area may have different water regimes and
be in different phases of the drawdown-reflooding
cycle. Their potential to produce may be just as
great but each is on a different time schedule. With-
out long-term observations, or experience in a cer-
tain climatic-geomorphic area, a highly erroneous
conclusion might be reached. Generally, such rela-
tions between maximum wildlife populations and
some habitat feature are based on one or few stud-
ies, so confidence levels may be low. Clearly, addi-
tional work is needed.

Tools for Evaluation. For marsh wildlife, we look
positively on certain features as being important en-
hancements in marshes and use them as parameters
of a quality habitat.

Complexity or diversity of vegetative structures
is measured by cover-water ratio, vertical layering,
and possibly plant species richness. Cover-cover
edges or cover-water edges were identified as attrac-
tive to nesting birds by Beecher ( 1942), and maxi-
mum species richness of birds was noted with a ratio
of 30% cover to 70% open water ( Weller and Fred-
rickson 1974). Using experimental approaches,

workers found the 50-50 cover-water ratio most at-
tractive to feeding dabbling ducks as well (Kaminski
and Prince 1981; Murkin et al. 1982). Vertical layer-
ing (strata) has been suggested as important for nest-
ing marsh birds by Weller and Spatcher ( 1965) and
Short (1982). These measures of structural diversity
suggest that increased variety and numbers of birds
would be expected with increased water, cover ra-
tio, or number of strata, but that low numbers and
different species may be expected in the extremes of
either a dense monoculture or sparse vegetation
(open marsh).

A wetland has characteristic water regimes due
to local climate, surface drainage patterns, and sub-
surface hydrology. Determining what is a favorable
regime is a moot issue, and dependent on the values
and functions attributed to the wetland. For wildlife,
water level stability during breeding seasons is ideal
to prevent flooding of duck nests (Low 1945) or
drowning of young muskrats (Errington 1937). Peri-
odic drawdowns are essential to establishment of di-
verse wetland plants in reestablishment of vegetation
after a flood-out or an eat-out by herbivores. Musk-
rats or nutria are more vulnerable to predators in
dry or shallow marshes (Errington 1961 ). Permanent
changes producing higher water level may be nearly
as detrimental to wildlife as is drainage of a wetland.
Such "flooding" reduces submergent growth (Robel
1962), availability of benthic organisms to some spe-
cies (Boyer and Psejek 1977), and covers established
beds of emergents (Weller 1981b).

Shoreline gradient and vegetative cover can be
very important to suitability of an area for wildlife.
Steep shores reduce emergent growth and induce
more terrestrial plants. Such shorelines may be ideal
loafing sites for ducks or muskrats, but are less suita-
ble for invertebrates that favor shallow water and
aquatic plants.

Turbidity reduces growth of important submer-
gents like sago pondweed, P. pectinatus (Robel
1961 ). Current or wave action tends to reduce perim-
eter emergents and takes out nutrients from the sys-
tem. Nutrient levels in water and soil can be mea-
sured by various chemical means. Usually, some sim-
ple measurements of alkalinity are used to reflect
nutrient-rich waters (eutrophic conditions) as op-
posed to less enriched (oligotrophic) water. Conduc-
tivity, an electrical measure of dissolved solids, is a
common, quick technique to assess whether water is
extremely pure, highly alkaline, or saline due to con-
centrations of minerals. Extreme levels of salt con-
tent may inhibit plant growth.

Nutrient cycling also is involved as nutrients
tend to be tied up quickly (seasonally) by plants,
and returned to water and substrate reservoirs by
decomposition of plants. Where decomposition is ex-
tensive in late stages of succession or following eat-

220

Marshes

outs, extensive surface floating plants or algal blooms
may be common. Again, this is not easily measured
but one must be conscious that fluctuation in float-
ing duckweed or other seasonal events is not neces-
sarily negative. Floating duckweed can be valuable as
food itself, and for the snails and amphipods it har-
bors. But continuous spring to fall crops suggest ex-
cess nutrients and duckweed will screen out growth
of submergents that produce seeds and tubers valua-
ble to waterfowl later in the year. Filamentous algae
can have the same effects but is less valuable as a
food.

Food resources are a product of water regimes
plus plant and invertebrate communities. Generally,
there are positive correlations between vegetative
complexity and bird populations (Weller and Fred-
rickson 1974) and between invertebrate abundance
and duck populations (Whitman 1976; Eriksson
1978;Joyner 1980). The presence of seed-producing
plants like smartweed and sago, or of foliage-produc-
ing plants like sago and widgeongrass, is of high
value for wildlife (Fredrickson and Taylor 1982). But
data are not available for all foods or all wildlife spe-
cies. What causes the increase in populations of am-
phibians like frogs and salamanders is complicated
by complex food chains of algae, phytoplankton, zoo-
plankton, and small invertebrates, and cannot be
quickly or easily assessed. The presence of fish is a
strong influence on quality of habitat for mergansers
and grebes, but large predatory fish may prey heavily
on ducklings (Lagler 1956).

The presence of macroinvertebrates is positive.
Cover-water ratios do not always enhance inverte-
brate numbers (Murkin et al. 1982), but variation in
type and numbers of invertebrates varies with
successional stage of the marsh (Voigts 1973). Be-
cause invertebrates are not easily assessed by any
simple and easily obtained measurement, taxon rich-
ness is being explored as the most simple way to as-
sess relative importance. However, one must be cau-
tious when associating abundance with availability as
deep water can prevent some ducks from obtaining
food (Boyer and Psujek 1977).

Nest sites and materials are a product of vegeta-
tion types and patterns. Plant density could influence
suitability of cover for nests — although data are not
available to clearly support this assumption. Sparse
vegetation may serve coots and grebes that carry
dead vegetation for nests and need only the sparse
material to hold the nest in a wind. But ducks do not
carry vegetation to build nests. Dead and live mate-
rials on the site must be sufficient to create a nest
bowl and sometimes a canopy. Songbirds seem to
need vertical stalks of vegetation that are sufficiently
robust that the weight of the nest, eggs, female, and
young do not cause collapse. The stalks must be suf-
ficiently close that weaving of nest materials be-

tween the uprights provides support and cover
(Weller and Spatcher 1965).

Muskrats carry lodge material, but build large
lodges where material is abundant, and small plat-
forms in suboptimal habitats. Breeding sites for fish,
frogs, and most reptiles really have not been evalu-
ated in marshes; it is believed that they too are at
peak levels when the entire ecosystem is responding
to nutrient availability, successional stages, inverte-
brate variety, and water depths that allow comple-
tion of life cycles.

Evaluation Systems. Evaluation of wetlands for
wildlife is not new, and several systems are currently
in use. The Habitat Evaluation Procedures (HEP)
developed by the U.S. Fish and Wildlife Service
(USFWS 1980) utilizes individual species models
called Habitat Suitability Index ( HSI ) models.
These models are used to build a composite of key
species within the habitat ( Edwards and Twomey
1982; Schroeder 1982). HSI models are not available
for all species but are being prepared by USFWS.
Such a system is used in mitigation but also has wide
applicability to biological survey work in marshes.
The system has been criticized for being species
rather than community oriented. This and other sys-
tems are described in the chapter on habitat evalua-
tion systems.

The U.S. Army Corps of Engineers has a more
ecosystem-oriented method ( Habitat Evaluation
System), but it has not reached the same level of
development and now deals mainly with terrestrial
systems ( U.S. Army Corps of Engineers 1 980 ).

The need for evaluation is so great for so many
diverse fields that less rigorous approaches are being
developed. A system in progress by the Federal High-
way Administration evaluates wetlands in their en-
tirety for all values and functions, with qualitative
ratings of high, moderate, and low for some 75 pre-
dictors (Adamus 1983; Adamus and Stockwell 1983).
Waterfowl and other water birds are used as the ma-
jor indicator groups so that much of the wildlife
community is not represented, but there are several
important predictors related to values for wildlife.
These include contiguity between wetlands, size,
vegetation form and density, salinity-conductivity,
pH, flooding, water depths, edge, waterfowl food val-
ues of plants, invertebrate density, alkalinity, and eu-
trophication. Collectively, these include many of the
important factors that have been identified as contrib-
uting to quality habitat for marsh wildlife. But no
one has derived a mechanism for rating these and
developing a cumulative rating without going to the
detail used in the HEP.

Several states (Michigan, New York, and Wiscon-
sin) also have wetland evaluation or rating systems
that emphasize wildlife and fish values, but thev tend

Marshes

221

to have only regional applicability. Numerous indi-
viduals and groups are experimenting with group-
specific evaluation systems incorporating current
ecological community factors such as species rich-
ness and species diversity in addition to population
data (Williams 1980). Thus, at the present stage of
development, the options are --

• make qualitative judgments based on items of
special interest to the evaluator (done consis-
tently and comparatively),

• use HEP if the HSI models are available and
seem to fit the needs, or

• make a series of gross to detailed surveys and
measurements that provide indices to those
items of special interest or importance at the
site or sites in question.

DISCUSSION

Compared with other habitats, marshes are so
dynamic they create difficult situations for assess-
ment as wildlife habitats. Marshes require - -

• long-term observations or use of past aerial pho-
tos to aid in interpretation,

• careful selection of parameters that fit the local
situation,

• testing of measurements for reliability as indica-
tors of wildlife habitats,

• standardization of times and techniques when
reassessing habitats,

• experimental efforts to improve technologies,
and

• data gathering on localized areas to reflect local
water regimes and wildlife populations.

Marshes are exciting because of their diversity
and dynamics, but they can be disappointing and de-
ceiving as well. A rich habitat once lost causes out-
cries from an uninformed public for correction of a
problem that may not exist. Public information on
the "boom and bust" dynamics of wetlands is essen-
tial to prevent wasted and needless action on areas
best left alone. Concurrently, gradual changes can be
hidden among the extremes, and only long-term
monitoring will indicate trends that induce concern
and investigation.

The accuracy of predictors, qualitative judg-
ments, or even detailed measurements is poorly
known. It is difficult to encourage extensive effort

when the variability of the system may mask impor-
tant events. Yet, there is no option. The alternative
of no evaluation, census, or monitoring is too likely
to result in irreparable damage to the system. The
level of accuracy can only be set by the needs, and
even detailed evaluation can ignore obvious truths. If
goals are established first, and the level of detail set
by need, many of the otherwise difficult decisions
become obvious.

Above all, do not assume someone else knows
all the answers. Instead, learn the system by repeated
visits and careful observation, search for the obvious
truths, use common sense in devising an evaluation
system, then stick with it as much as possible to pro-
vide comparable data between observations.

LITERATURE CITED

ADAMUS, PR. 1983. A method for wetland functional as-
sessment, Vol. II. FHWA assessment method. U.S. Dep.
of Transportation. Federal Highway Admin. Rep.
FHWA-IP-82-24. 138pp.

PR. and L.T. STOCKWELL. 1983- A method for

wetland functional assessment: Vol. I. Critical review
and evaluation concepts. U.S. Dep. of Transportation.
Federal Highway Admin. Rep. FHWA-IP-82-23. 181pp.

BEECHER, WJ. 1940. Nesting birds and the vegetative
substrate. Chicago. Chicago Ornith. Soc. 69pp.

BERGMAN, R.D., R.L. HOWARD, K.F. ABRAHAM, and M.W.
WELLER. 1977. Waterbirds and their wetland re-
sources in relation to oil development at Storkersen
Point, Alaska. U.S. Dep. Inter., Fish and Wildl. Serv. Re-
sour. Publ. 129. 38pp.

BOLEN, E.G. 1964. Plant ecology of spring-fed salt

marshes in western Utah. Ecol. Monogr. 34:143-166.

BOURN, W.S. 1935. Sea-water tolerance of Ruppia mari-
tima Contrib. Boyce Thompson Inst. 7:249-255.

BOYER, R.L. and MJ. PSUJEK. 1977. A comparison of wet-
land bird aggregations and macrobenthos. Trans., 111.
State Acad. Sci. 70:332-340

CHABRECK, R.H. 1972. Vegetation, water and soil charac-
teristics of the Louisiana coastal region. Louisiana St.
Univ! Bull. 66-4. 72pp.

CHRISTIANSEN, J.E. 1970. Water requirements of water-
fowl marshlands in northern Utah. Utah Div. of Fish
and Game. Publ. 69-12. 108pp.

COWARDIN, L.M. and D.H.JOHNSON. 1973. A prelimi-
nary classification of wetland plant communities in
north-central Minnesota. U.S. Dep. Inter., Fish and
Wildl. Serv. Special Sci. Rep— Wildl. 168. 33pp.

, D.S. GILMER, and L.M. MECHLIN. 1981. Character-
istics of central North Dakota wetlands determined
from sample aerial photographs and ground study.
Wildl. Soc. Bull. 9:280-288.
-, V. CARTER, F.C GOLET, and E.T. LAROE. 1979.

Classification of wetlands and deepwater habitats of
the United States. U.S. Dep. Inter., Fish and Wildl. Serv.
FWS/OBS-79/31- 103pp.
DERKSEN, D.V. and WD. ELDRIDGE. 1980. Drought-dis-
placement of pintails to the arctic coastal plain, Alaska.
J. Wildl. Manage. 44:224-229.

222

Marshes

DOZIER, H.L. 1948. Estimating muskrat populations by
house counts. Trans. North Am. Wildl. Conf. 13:372-
392.

EDWARDS, E.A. and K. TWOMEY. 1982. Habitat suitability
index models: common carp. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-82/10.12. 28pp.

ERIKSSON, O.G. 1978. Lake selection by goldeneye duck-
lings in relation to the abundance of food. Wildfowl
29:81-85.

. 1979. Competition between freshwater fish and

goldeneyes {Bucephala clangula ) for prey. Oceologia
41:99-107.

ERRINGTON, P.L. 1937. Drowning as a cause of mortality
in muskrats. J. Wildl. Manage. 18:497-500.

. 1961. Muskrats and marsh management. The Stack-
pole Co., Harrisburg, PA and The Wildl. Manage. Inst.,
Washington, DC. 183pp.

and T.G. SCOTT. 1945. Reduction in productivity

of muskrat pelts on an Iowa marsh through depreda-
tions of red fox. J. Agric. Res. 71:137-148.

ESHER, R.J., R.L. WOLFE, and J.H. LAYNE. 1978. Swimming
behavior of rice rats ( Oryzomya palustris ) and cotton
rats (Sigtnodon hispidis). J. Mammal. 59:551-558.

FISLER, G.F. 1961. Behavior of salt marsh Microtiis during
winter high tides. J. Mammal. 42:37-43-

FORNEY, J. L. 1968. Production of young northern pike in
a regulated marsh. N.Y. Fish and Game J. 15:143-154.

FRAYER, W.E., DC. BOWDEN, FA. GRAYGILL, and T.J.
MONAHAN. 1983- Status and trends of wetlands and
deepwater habitats. U.S. Dep. Inter., Fish and Wildl.
Serv. 31pp

FREDRICKSON, L.H. and T.S. TAYLOR. 1982. Management
of seasonally flooded impoundments for wildlife. U.S.
Dep. Inter., Fish and Wildl. Serv. Resour. Publ. 148.
29pp.

GOLET, F.C. and J.S. LARSON. 1974. Classification of fresh-
water wetlands in the glaciated Northeast. U.S. Dep.
Inter., Fish and Wildl. Serv. Resour. Publ. 116. 56pp.

HARRIS, S.W. and W.H. MARSHALL. 1963. Ecology of
water-level manipulations on a northern marsh. Ecol.
44:331-343.

HENNY, C.J., DR. ANDERSON, and R.S. POSPAHALA. 1972.
Aerial surveys of waterfowl production in North
America, 1955-71. U.S. Dep. Inter., Fish and Wildl.
Serv. Special Sci. Rep.-Wildl. 160. 48pp.

HIGGINS, K.F., L.M. KIRSCH, and I.J. BALL. 1969- A cable-
chain device for locating duck nests. J. Wildl. Manage.
33:1009-1011.

JOANEN, J.T. and L.L. GLASGOW. 1966. Factors influenc-
ing the establishment of widgeongrass stands in Loui-
siana. Proc. Ann. Conf. SE Assoc. Game and Fish
Comm. 19:78-92.

JOYNER, D.E. 1980. Influence of invertebrates on pond se-
lection by ducks in Ontario. J. Wildl. Manage. 44:700-
705.

KAMINISKI, R.M. and H.H. PRINCE. 1981. Dabbling duck
and aquatic macroinvertebrate responses to manipu-
lated wetland habitat. J. Wildl. Manage. 45:1-15.

KRAPU, G. 1974. Feeding ecology7 of pintail hens during
reproduction. Auk 91:278-290.

KUSHLAN, J.A. 1974. Quantitative sampling offish popula-
tions in shallow, freshwater environments. Trans. Am.
Fish. Soc. 103:348-352.

LAGLER, K.F. 1956. The pike, Esox lucius Linnaeus, in re-
lation to waterfowl on the Seney National Wildlife
Refuge, Michigan. J. Wildl. Manage. 20:114-124.

LOW, J.B. 1945. Ecology and management of the redhead,
Nyroca americana, in Iowa. Ecol. Monogr. 15:35-69.

MARTIN, AC, N. HOTCHKISS, F.S. UHLER, and W.S.

BOURN. 1953. Classification of wetlands of the United
States. U.S. Dep. Inter., Fish and Wildl. Serv. Special
Sci. Rep.-Wildl. 20. 14pp.

McILHENNY, E.A. 1976. The alligators life history. Soc. for
the Study of Amphibians and Reptiles. 117pp.

McKNIGHT, D. and J.B. LOW. 1969. Factors affecting wa-
terfowl production on a spring-fed salt marsh in Utah.
Trans. North Am. Wildl. Nat. Resour. Conf. 34:307-
314.

METCALF, F.P. 1931. Wild duck foods of North Dakota
lakes. U.S. Dep. Agric. Tech. Bull. 221. 72pp.

MEYER, MP. and P.D. GRUMSTRUP. 1978. Remote sensing
applications in agriculture and forestry. Univ. of Min-
nesota, St. Paul. IAFHE RSL Res. Rep. 78-1. 60pp.

MIKOL, S.A. 1980. Field guidelines for using transects to
sample nongame bird populations. U.S. Dep. Inter.,
Fish and Wildl. Serv. FWS/OBS-80/58.

MONTANARI, J.H. and J.E. TOWNSEND. 1977. Status of
the national wetlands inventory. Trans. North Am.
Wildl. Nat. Resour. Conf. 42:66-72

MOYLE, J.B. 1945. Some chemical factors influencing the
distribution of aquatic plants in Minnesota. Am. Midi.
Naturalist. 34:402-420.

MUELLER-DUMBOIS, D and H. ELLENBERG. 1974. Aims
and methods of vegetation ecology. John Wiley &
Sons, Inc. 545pp.

MURKIN, H.R., R.M. KAMINSKI, and R.D. TITMAN. 1982.
Responses by dabbling ducks and aquatic inverte-
brates to an experimentally manipulated cattail marsh.
Can. J. of Zool. 60:2324-2332.

NELSON, N.F. 1954. Factors in the development and resto-
ration of waterfowl habitat at Ogden Bay Refuge, We-
ber County, Utah. Utah Dep. of Fish and Game. Publ.
6. 87pp.

NEXON, S.W. 1982. The ecology of New England high salt
marshes: a community profile. U.S. Dep. Inter., Fish
and Wildl. Serv. FWS/OBS-81-55. 70pp.

PAYER, R.D. and C.G. SCALET. 1978. Population and pro-
duction estimates of fathead minnows in a South Da-
kota prairie marsh. Prog. Fish-Cult. 40:63-66.

PEDERSON, R.L 1981. Seed bank characteristics of the
Delta Marsh, Manitoba: applications for wetland man-
agement. Pages 61-82 in Richardson, B., Selected Pro-
ceedings of the Midwest Conference on Wetland Val-
ues and Management. 660pp.

PENFOUND, W.T. and E.S. HATHAWAY. 1938. Plant com-
munities in the marshland of southeastern Louisiana.
Ecol. Monogr. 8:1-56.

POSPAHALA, R.S., DR. ANDERSON, and C.J HENNY.
1974. Population ecology of the mallard: II Breeding
habitat conditions, size of breeding populations, and
production indices. U.S. Dep. Inter., Fish and Wildl.
Serv. Resour. Publ. 1 1 5. 73pp.

RAWSON, D.S. and G.T. MOORE. 1944. Saline lakes of Sas-
katchewan. Can. J. of Res. 22:141-201.

ROBEL, R.J. 1961. The effects of carp populations on the
production of waterfowl food plants on a western wa-
terfowl marsh. Trans. North Am. Wildl. Nat. Resour.
Conf. 26:147-159.

ROBEL, R.J. 1962. Changes in submersed vegetation fol-
lowing a change in water level. J. Wildl. Manage.
26:221-224.

SCHROEDER, R.L. 1982. Habitat suitability index models:
yellow-headed blackbird. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-82/ 10.26. 12pp.

SHAW, S.P. and C.G. FREDINE. 1956. Wetlands of
the United States: their extent and their value to

Marshes

225

waterfowl and other wildlife. U.S. Dep. Inter., Fish and
Wildl. Serv. Circular 39. 67pp.

SHORT, H.L. 1982. Use of northern prairie and wetland
habitat by breeding birds. U.S. Dep. Inter., Fish and
Wildl. Serv. 27 pp.

and K.P. BURNHAM. 1982. Technique for structur

ing wildlife guilds to evaluate impacts on wildlife
communities. U.S. Dep. Inter., Fish and Wildl. Serv.
Spec. Sci. Rep.-Wildl. 244. 34pp.

SMITH, AG. 1971. Ecological factors affecting waterfowl
production in the Alberta parklands. U.S. Dep. Inter.,
Fish and Wildl. Serv. Resour. Publ. 98. 49pp.

STEWART, R.F. and HA. KANTRUD. 1971. Classification of
natural ponds and lakes in the glaciated prairie region.
U.S. Dep. Inter., Fish and Wildl. Serv. Resour. Publ. 92.
57pp.

. 1972. Vegetation of prairie potholes, North Dakota,

in relation to quality of water and other environmen-
tal factors. U.S. Dep. Inter., Fish and Wildl. Serv. and
Geological Survey. Geological Survey Professional Pa-
per 585-D. 36pp.'

. 1973. Ecological distribution of breeding water-
fowl populations in North Dakota. J. Wildl. Manage.
37:39-50.

SUGDEN, L.G. 1978. Canvasback habitat use and produc-
tion in Saskatchewan parklands. Can. Wildl. Serv. Occ.
Pap. 34. 32pp.

SWANSON, G.A., G.L. KRAPU, and JR. SERIE. 1979. Foods
of laying female dabbling ducks on the breeding
grounds. Pages 47-57 in Bookhout, T.A., ed. Proc.
1977 Symp. Waterfowl and Wetlands. The Wildlife So-
ciety. Madison, WI.

TANNER, W.D. and G.O. HENDRICKSON. 1956. Ecology
of the king rail in Clay County, Iowa. Iowa Bird Life.
26:54-56.

TOLL, J.E., T.S. BASKETT, and C.H. CONAWAY. I960.
Home range, reproduction, and foods of the swamp
rabbit in Missouri. Am. Midi. Naturalist. 63:398-412.

U.S. ARMY CORPS OF ENGINEERS. 1980. A habitat eval
uation system for water resources planning. U.S. Army-
Corps of Engineers, Lower Mississippi Valley Division.
1 58pp.

U.S. FISH AND WILDLIFE SERVICE. 1980 Habitat evalu
ation procedures (HEP). U.S. Dep. Inter., Fish and
Wildl. Serv. ESM 102. 27pp.

VAN DER VALK, AG. 1981. Succession in wetlands: a
gleasonian approach. Ecol. 62:688-696.

and C.B. DAVIS. 1978. The role of seed banks in

the vegetation dynamics of prairie glacial marshes.
Ecol. 59:322-335.

VOIGTS, D.K. 1973- Food niche overlap of two Iowa
marsh icterids. Condor. 75:392-399.

WELLER, M.W. 1979a. Wetland habitats. Pages 210-234 in
Greeson, P.E., J.R. Clark and J.E. Clark, eds. Wetland
Functions and Values: the State of Our Understanding.
Am. Water Resour. Assoc. 674pp.

. 1979b. Birds of some Iowa wetlands in relation to

concepts of faunal preservation. Proc. la. Acad. Sci.
86:81-88.

. 1979c. Density and habitat relationships of blue-
winged teal nesting in Northwestern Iowa. J. Wildl.
Manage. 43:367-374.

. 1981a. Freshwater marshes: ecology and wildlife

management. Univ. Minn. Press. 146pp.

. 1981b. Estimating wildlife and other wetland

losses due to drainage and other perturbations. Pages
337-346 in Richardson, B., ed. Selected Proceedings of
the Midwest Conference on Wetland Values and Man-
agement. Minn. Water Planning Board, St. Paul, MN.
660pp.

and L.H. FREDRICKSON. 1974. Avian ecology of a

managed glacial marsh. Living Bird. 12:269-291.

and C.E. SPATCHER. 1965. Role of habitat in the

distribution and abundance of marsh birds. Iowa State
Univ. Agric. and Home Econ. Exp. Sta. Spec. Rep. 43.
31pp

and D.K. VOIGTS. 1983. Changes in the vegetation

and wildlife use of a small prairie wetland following a
drought. Proc. Iowa Acad. Sci. 90:50-54.

WHITMAN, W.R. 1976. Impoundments for waterfowl. Can.
Wildl. Serv. Occ. Pap. 22. 22pp.

WIGGINS, G.B., R.J. MACKAY, and I.M. SMITH. 1980. Evo-
lutionary and ecological strategies of animals in annual
temporary pools. Arch. Hydrobiol. Suppl. Bd. 58.
206pp.

WILLIAMS, G. 1980. An index for the ranking of wildfowl
habitats, as applied to eleven sites in West Surrey,
England. Biol. Conserv. 18:93-99.

YEAGER, L.E. and H.M. SWOPE. 1956. Waterfowl produc-
tion during wet and dry years in north-central Colo-
rado. J. Wildl. Manage. 20:442-446.

ZEDLER, J.B. 1982. The ecology of southern California
coastal salt marshes: a community profile. U.S. Dep. In-
ter., Fish and Wildl. Serv. FWS/OBS-81/54. 1 10pp.

ZIMMERMAN, J. L. 1977. Management of migrator)' shore
and upland game birds. Pages 45-46 in Sanderson,
G.C., ed. Virginia Rail (Rallus limicola). International
Assoc. Fish and Wildlife Agencies, Washington, DC.
358pp.

224

Marshes

11

STREAMS

Paul Cuplin

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

"I love any discourse of rivers, and fish, and fishing.
— Izaak Walton, Compleat Angler

Editor's Note: This chapter is the first of two on
aquatic habitats, and covers moving water habitat
(rivers and streams). The emphasis is on the
smaller streams typical of the western United States,
as opposed to large systems such as the Mississippi
River. As with all other natural systems, the streams
cannot be understood or measured without looking
at the associated wetlands and uplands. Therefore,
stream inventory and monitoring involves measur-
ing or estimating factors such as stream bank sta-
bility and dominant riparian vegetation. Never-
theless, the focus of a study is often the stream and
its potential for supporting fish and other aquatic
vertebrates. This chapter covers the basic measure-
ments and measurement systems for describing
stream habitat.

INTRODUCTION

Eighty-five thousand miles of fishing streams
occur on public lands managed by the U.S. Bureau of
Land Management ( BLM ) in the 1 1 western States
and Alaska (Table 1 ). Streams provide recreational
opportunities ranging from white-water boating to
family picnicking. Important wildlife habitat associ-
ated with streams include riparian vegetation which
provides food and cover for many wildlife species.
Fresh-water fishing in terms of sportsfishing dollars is
estimated to be $11 per fisherman per day. About
36.4 million Americans made 620.5 million fishing
trips in 1980. Total dollars expended annually for
fresh-water fishing in the U.S. is estimated to be $7.8
billion (U.S. Department of the Interior, Fish and
Wildlife Service and U.S. Department of Commerce
1982). Salmon and steelhead originating from public
land streams that are taken in commercial fisheries of
Alaska and Oregon totaled 58,358,000 pounds in
1983 (U.S. Department of the Interior, Bureau of
Land Management 1983). Commercial fisheries catch
statistics can be related to dollar values by fish
species.

Stream habitat management requires an aware-
ness of land uses that can improve, stabilize, or de-
stroy stream habitat. A starting point is a baseline
description of existing habitat conditions. Changes in
habitat conditions are detected by monitoring spe-
cific attributes of the stream over time. Stream habi-
tat features, i.e., stream bank stability, streambed
siltation, stream channel stability, shade, water qual-
ity, stream width and depth, and other features are
monitored to detect changes in stream habitat
conditions.

STREAM CLASSIFICATION

Although classification of streams has received
some attention in the U.S., there is no universally
accepted method of classification. The U.S. Fish and

Streams

225

Wildlife Service has developed a wetland and deep-
water classification for a national wetland inventory
(Cowardin et al. 1979) to comply with the Clean
Water Act of 1977. This classification is very useful
for broad-base wetland inventory and mapping, but it
does not provide information on condition that is
needed by the public land resource specialist or
manager.

A stream classification would be useful for com-
paring streams that are similar in size, water quality,
elevation, and geomorphology. Common classifica-
tion terminology would allow for predicting aquatic
habitat potential relative to aquatic habitat
improvement.

Stream order, which provides a concept of
stream size and characteristics for the biologist and
manager, was first defined by Horton (1945). He
designated unbranched tributaries as first-order
streams; streams that receive first-order tributaries, as
second-order streams; those that receive second- or
first- and second-order streams as third-order
streams, and so on until the mouth of the stream is
reached. Strahler (1952) modified Horton's stream
order and called all unbranched tributaries first-order
streams; two first-order streams join to make a sec-
ond-order stream, and so on downstream to the
stream mouth (Figure 1).

Figure 1. Stream order by A.N. Strahler ( 1952).

Table 1. Estimated miles of existing fishing streams, and acres of lakes and reservoirs on public
lands administered by U.S. Bureau of Land Management ( 1983).

State

Lakes1

Reservoirs1

Streams2

Alaska

3,874

0

65,000

Arizona

1

26

439

California

19

2

818

Colorado

1

19

1,822

Idaho

10

39

3,577

Montana

16

22

996

Nevada

23

5

1,134

New Mexico

2

3

176

Oregon,

Washington

29

18

7,136

Utah

1

8

2,300

Wyoming

7

32

2,537

Total

3,983

174

85,935

1ln thousands of acres.

includes only miles of fishable streams.

226

Streams

Third-order and larger streams generally have
gradients less than 0.5%, and streambed is usually
rubble, gravel, sand, silt, and organic detritus. These
streams are meandering with abundant riparian
shrubs and trees; stream stability is generally fair.
Sediment discharge is higher than the smaller first-
and second-order streams. Fishery habitat is good.

Map scale influences determination of stream or-
der. For example, a l:500,000-scale map would not
show most first- and second-order streams. Although
l:24,000-scale maps will show first-order and larger
streams, Boehne and House (1983) suggested
1:12,000 as a minimum scale for mapping stream
orders of Oregon streams.

Stream order descriptions vary within physio-
graphic regions. Streams in mountainous regions are
used as an example for a general description of
stream orders.

In my experience, first- and second-order
streams generally have steep gradients ( 3% or more )
and large coarse streambed and streambank material.
Stream meander is moderate; riparian shrubs and
trees are present; and debris jams are common. Fish-
ery habitat is suitable in some reaches for spawning
and rearing.

STREAM HABITAT FEATURES

Stream habitat feature is a convenient term used
to identify the variables that make up stream habitat.
Stream habitat features are streamflow patterns,
streambank stability, stream channel stability, ripar-
ian vegetation, riffles, pools, streambed, stream
depth, stream width, stream gradient, stream diver-
sity, stream water chemistry, and macroinvertebrates.

Streamflow Pattern

Streamflow pattern significantly affects the biotic
life of the stream. The most desirable streamflow
pattern for aquatic production is one of perennial
flow with moderate spring runoff. Perennial streams
can provide year-round spawning, rearing, and feed-
ing requirements for fish. However, extreme drought
may cause drying of perennial streams. The type of
spring runoff will govern habitat stability.

Intermittent streams flow only part of each year.
Generally streamflow occurs during the rainy season
or during snow melt; otherwise the streambed is dry.
Intermittent streams are indicated by a dashed line
on U.S. Geological Survey (USGS) 7.5-minute quad
maps. Intermittent streams often provide spawning
and rearing habitat for fish before the streambed
dries up during the summer months.

Ephemeral streams are those that flow only in
response to a rain shower. The ephemeral stream
provides marginal habitat for aquatic life but it may
be very important to amphibians.

Instream Flow Needs

The amount of water needed to maintain stream
habitat on a year-round basis is termed "instream
flow needs." It can be determined by several meth-
ods as summarized by Cuplin et al. ( 1979). Where
detailed, precise, legally defensible information is
required, the multiple transect-incremental flow
method is recommended. This method is described
in detail in Bovee and Milhous ( 1978).

For rapid assessment, a single transect method is
available (Cuplin et al. 1979). This technique re-
quires less time for field work and data analysis but
it produces less precise results than the incremental
flow method.

Computer programs have been developed for
analyzing stream channel cross-section survey data
collected in conjunction with instream flow assess-
ments (Parsons and Hudson 1985). A single transect
method can be used for rapid assessment. A com-
puter analysis program is available for BLM offices
through the Denver Service Center, Division of Re-
source Systems (Parsons and Hudson 1985).

The flow of a stream is measured in cubic feet
per second (cfs) to provide information on stream
water volume. Water flow is estimated to the nearest
cfs during stream inventory. The U.S. Geological
Survey maintains permanent streamflow measuring
weirs and annually publishes surface water reports
for each State. These reports include data on stream-
flow and water chemistry of selected streams.

Streambank Stability

Stable streambanks and abundant riparian vege-
tation provide good habitat for aquatic and terrestrial
wildlife. Aquatic systems depend on the condition
of the adjacent land. Removal of riparian vegetation
through timber harvest, road construction, overgraz-
ing, or excessive recreational use can cause unstable
streambank conditions. Eroding streambanks allow
sediment to enter the stream and degrade the
streambed. Silt and sediment fill the spaces between
streambed gravel and rubble and prevent the flow
of water and oxygen necessary for the survival of fish
eggs, larval fish, and macroinvertebrates.

Stream Channel Stability-
Stream channel stability is related to the water
flow of the stream and the sediment load carried by
the stream. Material that enters the stream from the

Streams

227

watershed or streambanks is either washed down-
stream or becomes part of the bed material ( Bovee
1982). Schumm and Meyer (1979) related the shape
of the stream channel to the amount and kind of
load it carries. Channels can be classified as sus-
pended load (silt and clay), mixed load, or bed load
channels (sand and gravel). Five channel patterns
have been identified which correspond to each of
the types of load and transitions between load types
(Figure 2; Shen et al. 1981 ).

Riparian Vegetation

Woody riparian vegetation, shrubs and trees, are
most important to smaller streams. Streamside

woody vegetation provides shade that prevents ex-
cessive warming of water during summer months.
Overhanging trees provide cover for fish and the
leaves and twigs provide stream energy through
nutrient cycling. Riparian vegetation provides a
buffer from upland activities as well as a filter for
overland soil erosion.

A method of predicting riparian vegetation po-
tential has been developed by Crouse and Kindschy
(1981). This method provides guidelines for predict-
ing the riparian vegetation that can exist under var-
ious conditions of soil and water flow if protection
through fencing or grazing management system is
applied to a stream/riparian area.

CHANNEL TYPE

Suspended
Load

Mixed Load

Bed Load

h-

>

*, *»=!*

X
CD

CO

CD
3

T.

<

§ £

3;

CO

O

DC

o o

I

h-
co

-J _l

CD
Co

-^£^*£

Z
DC

i((&£) /£

\

CD CO
O 3

2

CD
CO

3

m

LU
(-

<

CD

o

CO
DC _

ajw\

^1 fe

^

o

c

a.

CD

r-
>

<

LL

DC

_c ^

' IK 1 i

m

LU

±3 CD

1 Q)w. J 1

\ V

_J

Q

m-O

V <*ttB^ )

\ v-

05

LU

Z

<

Ll2

v^y

V.

o

S

>

z
<

X

o

LU
2

L

CD ^t

o

c

CO

3
Q.
CD

CO

r~

H

-<

Q

>

O

3

I —

^-——+. ^-^-~"^

LU

D

<

DC
CD

<

c

o'

3

3
3
CD

CO

3;

O

Small

Sediment Size

Large

Small

Sediment Load

Large

Low

Flow Velocity

High

"" ""*— — Chanr

el Boundary

Low

Stream Power

High

,-*•"" "*> Flow

(3% >) Low

Bed Load Total Load Ratio High (> 11%)

tfTgg^ Bars

LOW

RELATIVE STABILITY HIGH

Figure 2. Channel classification showing relative stability and types of hazards encountered with each pat-
tern (from Shen et al. 1981 ).

228

Streams

Typical riparian vegetation along a foothills' stream.

Low gradient stream.

Riffles

Riffles are the steeper gradient, high-water ve-
locity stream sections. Macroinvertebrates reproduce
in the riffle, and fish feed on them at the base of
riffles in either pools or runs. Steep gradient streams
that are largely composed of riffles provide food
and some spawning area but not the necessary rest-
ing and shelter mix to provide good habitat for sal-
monids. Other fishes, e.g., certain suckers and
minnows, are almost exclusively found over riffles.

Pools

Pools are the deeper, slow-moving-water por-
tions of streams. Pools can be classified by width and
depth (Duff and Cooper 1978; Hamilton and Berger-
sen 1984) and they provide resting areas for fish.

Runs

Runs are shallow troughs that generally have
sand or gravel bottoms. Runs are smooth, laminar
flows of slow to moderate velocity, intermediate
between a pool and a riffle.

Streambed

The streambed consists of various amounts and
sizes of material. These are classified by particle size
(Table 2). The amount of different-sized particles

that compose/make up the streambed determines the
uses of the stream bottom by fish, macroinverte-
brates, and other aquatic life. For example, coarse
gravel (2.5-7.5 cm [1-3 in. diameter]) is most com-
monly used by salmonids for spawning, whereas
small rubble (7.5-15.0 cm [3-6 in. diameter]) free of
silt and sand, provides good habitat for macroinver-
tebrates. Silt and sand inhibit the flow of water and
oxygen through rocks in the streambed and reduce
the success of fish egg incubation and aquatic orga-
nism survival.

Table 2. Classification of streambed material (Duff
and Cooper 1976).

Material

Particle Size

Bedrock

Exposed solid rock

Boulder

12" (305 mm)

Large rubble

6-12" (152-305 mm)

Small rubble

3-6" (76-152 mm)

Coarse gravel

1 -3" (25-76 mm)

Fine gravel

0.1-1.0" (2.5-25 mm)

Sand

(0.074-2.5 mm)

Clay

(0.074 mm)

Sapropel

Reduced organic matter

Detritus

Particulate organic matter

Streams

229

Stream Width and Depth

The average width of a stream is the average
distance between the water's edges, perpendicular to
the stream flow (Hays et al. 1981). Stream depth is
the average vertical distance from the surface to the
bottom (Hays et al. 1981).

Wide shallow streams provide reduced habitat
for aquatic life. They usually lack resting cover for
fish and allow water temperature to rise beyond the
tolerance limits of some fishes. Narrow deep streams,
in comparison, maintain lower water temperatures
because of less surface area exposure to solar warm-
ing, and the deep water provides good resting cover
for fish.

Stream Gradient

A steep, torrential mountain stream provides
very little habitat for fish. A moderate gradient will
provide some of the requirements for food produc-
tion, spawning, and rearing of fish. A very low gra-
dient, deepwater stream provides the greatest
stability and warmer water. Large, low gradient
streams, free of pollution, provide the greatest habi-
tat diversity and the largest number of aquatic
species.

Stream Water Chemistry

Suspended Solids and Turbidity. Suspended
solids in the form of silt, clays, and other fine

materials can cause temporary to permanent
turbidity or murkiness.

Prolonged turbidity caused by runoff or distur-
bance in normally clear streams can cover the
streambed with silt which can smother macroinver-
tebrates, cover spawning areas, and reduce photosyn-
thetic rates. Turbidity can cause feeding problems
for sight-feeding trout and gill irritation to most
fishes with the exception of those that are adapted
to year-round turbidity.

Turbidity caused by suspended solids can be
judged in terms of visibility of objects in water, such
as muddy water (0.5-1 foot of visibility), murky
water (1-5 feet of visibility), and clear water (5 feet
of visibility). Turbidity can be measured and re-
ported in Nephelometric Turbidity Units (NTUs),
which are measures of light transmittance in a water
sample. Turbidity in excess of 25 NTUs is judged
to be too high for sight-feeding fish.

Water Temperature. Water temperature range is
an important criterion of the resident fish and
aquatic species in a stream. Species that are
acclimated to various maximum water temperatures
in their life cycle are listed in Table 3 (U.S.
Environmental Protection Agency 1971). An increase
in temperature of a few degrees can upset the
natural balance in a stream. Fish and other organisms
exposed to maximum temperatures for long periods
may suffer from unusual stresses, disease, and
reduced appetite.

Table 3. Maximum temperature probably compatible with the well-being of various species of fish and their
associated biota in °C and °F (U.S. Environmental Protection Agency 1971 ).

Temperature

Taxa

34°C (93.2°F)

Growth of catfish (Ictaluridae), gar (Lepisosteus sp.), white bass (Morone chrysops),
yellow bass (M. mississippiensis), spotted bass (Micropterus punctulatus), buffalo
{Ictiobus sp), carpsucker (Carpiodes sp), threadfin shad (Dorosoma petenense), and
gizzard shad (D. cepedianum)

32°C (89.6°F)

Growth of large-mouth bass (M. salmoides), drum {Aplodinotus grunniens), bluegill
(Lepomis macrochirus), and crappie (Pomoxis sp.)

29°C (84.2T)

Growth of pike (Esox lucius), perch {Percina sp), walleye {Stizostedion vitreum), small-
mouth bass (M. dolomieui), sauger (S. canadense), California killifish (Fundulus parvi-
pinnis), and topsmelt (Atherinops affinis)

27°C (80.6°F)

Spawning and egg development of catfish, buffalo, threadfin shad, gizzard shad, Cali-
fornia grunion (Leuresthes tenuis), opaleye (Girella nigricans), and northern swellfish
(Sphoeroides maculatus)

24°C (75.2°F)

Spawning and egg development of largemouth bass, white and yellow bass, spotted
bass, sea lamprey (Petromyzon marinus), alewife {Alosa pseudoharengus), and striped
bass {Morone saxitilis)

19°C (66.2°F)

Growth of migration routes of salmonids and for egg development of perch, smallmouth
bass, winter flounder (Pseudopleuronectes americanus), and herring (Clupea sp.)

12°C (53.6°F)

Spawning and egg development of all salmonidae (other than lake trout [Salvelinus
namayacush])

9°C (48°F)

Spawning and egg development of lake trout, walleye, northern pike, sauger, and Atlan-
tic salmon {Salmo salar)

230

Streams

Chemical Variables. The three most important
chemical measurements in stream water are
dissolved oxygen, pH (hydrogen ion concentration),
and specific conductance which can be converted
to represent total dissolved solids in milligrams per
liter (mg/L).

Dissolved Oxygen. The dissolved oxygen con-
tent of water is an indicator of the biochemical con-
dition of water. It also indicates the balance between
oxygen-consuming and oxygen-producing processes.
Fish and other desirable clean-water biota require
relatively high dissolved oxygen levels at all times.
Dissolved oxygen levels below 6 mg/L are dangerous
to sensitive or higher elevation fish. Desert fishes
can survive and withstand low oxygen tensions to 1
mg/L.

Streams with large loads of organic material may
have oxygen-consuming and inorganic reactions
that deplete oxygen to levels unfavorable for the
clean-water species.

pH. The pH is the abbreviation for the negative
base 10 log of the hydrogen ion concentration.
Stream water not influenced by pollution has a pH
between 6.5 and 9.0, which is the acceptable range
of pH for fish.

Specific Conductance. Specific conductance is
the measure of the ability of water to conduct an
electrical current and is expressed in micromhos per
centimeter at 25°C (7°F). It can be used for approxi-
mating the dissolved solids in water by using the
formula —

Specific conductance x (0.65 ± 0.05) = mg/L
dissolved solids or total dissolved solids
(see below).

The formula should be verified by comparing
specific conductance with total dissloved solids de-
termined by a chemist.

Total Dissolved Solids Total dissolved solids
(TDS) are all of the dissolved materials present in
natural waters and consist of carbonates, bicarbon-
ates, chlorides, sulfates, phosphates, and other sub-
stances. Most productive fresh water has a TDS
above 350 mg/L. The maximum safe level is about
1,500 mg/L TDS in fresh water. Chemical laboratory
assistance will be needed for the analysis of heavy
metals, TDS, pesticides, and pollutants.

Macroinvertebrates

Macroinvertebrates are aquatic organisms that
can be seen with the naked eye, such as mayfly, stone-
fly, and caddis fly larvae; Gammarus sp., aquatic
worms, beetles, clams, and snails. These animals pro-
vide an important food supply for fish.

Macroinvertebrate production in streams is re-
lated to water temperature, water chemistry, and
habitat diversity. A stream that is low in TDS and has
low average temperatures will produce few macroin-
vertebrates per square foot of stream bottom.

Rock and rubble 7.5-15 cm (3-6 in.) in diameter
provide good habitat for macroinvertebrates. Smaller
rocks from 2.5-7.5 cm (1-3 in.) provide less desir-
able habitat, hence fewer numbers of macroinverte-
brates. Sand and silt provide little habitat except
for tubificid worms if nutrient levels are high.

Plant-eating macroinvertebrates consume algae,
detritus, diatoms, wood, and leaves. Some macroin-
vertebrates are carnivorous and prey on other
stream dwellers.

Stream Diversity

The pools, riffles, runs, boulders, streambed
gravel and rubble, shade, downed trees, and gradient
all provide habitat diversity in a stream. Streams
with good habitat diversity provide the essential re-
quirements of food, cover, spawning, and rearing for
fish.

Stream habitat improvement enhances diversity
and provides the missing habitat features such as
pools, streambed spawning gravel, cover, and stream-
bank stabilization.

An example of diversity destruction, which was
once a common practice, is the removal of all
downed trees and logs from a stream. It is now rec-
ognized that the natural tree fall into a stream adds
to the diversity and to the flow of energy to support
the food chain.

STREAM INVENTORY AND MONITORING
SYSTEMS

Stream Habitat Inventory

Streams are inventoried to determine existing
and potential aquatic and terrestrial wildlife habitat
related to existing or planned land uses. A general
system that can be used for photointerpretation and
on-the-ground assessment is the stream habitat in-
ventory developed by the Oregon State Office, BLM
(Cuplin 1981). This method uses five criteria (shade,
riparian zone condition, streambank stability, stream
channel stability, and sedimentation of streambed) to
classify a stream as poor, fair, good, or excellent.
Each of these attributes is rated in one of three or
four categories and the ratings are then summed to
give an overall stream condition rating (Figure 3).
This classification has limited applicability on south-
ern desert erosional streams.

Streams

231

CO

b

c
g

S
"cd

Q.

a>
c

o

o

.c

Q.

o

o
c

CD
>

0

■D
CD
CO

CD

JD

o

E
o

CO

CD

co

T3

CD

CD
CO

E
to

CD

-C
Q.
CO

w_

o

I I *

> en .5?

C- c >-

3 CD CD

CO -J <

2 O

s z

CD

■4— •

CO

E

CO
CD

CO

c

cn

c
'c
c
jo

CL

>>

CD

3

CO

32

CD
LL

CD

CD

CD

CD

CO

>h_

k~

co

O

o

o

O

0

E

E

E

E

o

o

o

o

o

0s

^5

•4>

O

LO

f—

CD

CD

«fr

CM

■"3"

CNJ

CM

CM

C\J

CNJ

CJ

CO

CO

^5

.9

^5

^5

O

ID

k_

o

LD

LO

CD

CM

^—

CM

■

1

i

'

O

CD

O

T

T—

"*

■<t

co

CO

co

CO

co

CO

CO

CO

CO

sS

^5

_CD

0

o
co

ID

o

O

o

0^

o

CD

o

CD

O

CD

CM

■<*

■*

^t

■t

CO
CO

co

CO

CO

CD

CD

CO

o^

+
o

CO

o

o

1

o

O
ID

CD
-Q

o

1-
c
E

3

o
O

C

£>_

|

o

-^ "*"—

CO

"6

c
o

Is

CO §

CO

o

1 — >

o

CO CO

CD O

^E,

c-2

c
o

O O

C (0

2o

c- CO

COQ
CD^

CD
TJ

CO

SZ
CO

NCO
C CD

co CD

■9-^p
QC2^

CO

c

CD —

^5

CO 2^

co£^

co^

CD
O

co

Q.

w

£

.5?

Q.

O

Q.

Q.
CO

c

LLI

CD
00

_g
co
>

LU

E

CO
CD

co

o

c

CD

c
ro

DC

c
o

t5
c
o
o

E

CO
CD

00

Is- CD

*~ V co

c "* V cn

fl CJ >_ (-)

CO

luOllq.

cd

CD
CO

jo
o

c
£
o

co
co

"3

n

cn

c

T3
CO

>>

n

co 2

CD (0

. CL

0) t
"D CD

b t!

-I

CM 2

_T "CO
CO ^^

^- o

i_ c

o c

— CO

c °

co c
■c o

Q_ CO

E ^

— a>

CO CD

— >

CD c

^■g

§1

CD <D

y 53

CD -C

232

Streams

Habitat Quality Index (HQI)

Binns (1979) developed a Habitat Quality Index
(HQI) which relates habitat quality to trout biomass
in streams. He field-tested 1 3 physical habitat attri-
butes, 5 water chemistry attributes, and 4 biological
attributes and chose 9 attributes for the HQI (Table
4). The attributes chosen for the HQI are late sum-
mer streamflow, annual streamflow variation, maxi-
mum summer stream temperature, water velocity,
cover, stream width, eroding banks, substrate, and ni-
trate nitrogen.

The HQI is given by the expression:

log10 (Y+ 1) = [( -0.903) + (0.807)log,„ (X, + 1 ) +
(0.877) log 10(X2 + 1) + (1.233)
log10(X3 + 1) + (0.631) log10(F +
1) + (0.182) log10(S + 1)]
[1.12085]

Where.

Y = Predicted trout standing crop (biomass)

X, = Late summer stream flow

X2 = Annual stream flow variation

X3 = Maximum summer stream temperature

F =Food index = X, (X4) (X9) (X10)

S = Shelter index = X7 (X8) (X,, )

X4 = Nitrate nitrogen

X7 = Cover

X8 = Eroding stream banks
X9 = Substrate
Xio = Water velocity
Xn = Stream width

Binns (1979) found a close correlation between
predicted and measured standing crop of trout in 36
Wyoming streams.

Riparian Aquatic Information Data System
(RAIDS)

A computer-based system for summarizing ripar-
ian and aquatic data has been developed by the U.S.
Bureau of Land Management, Service Center. This
system, called the Riparian Aquatic Information Data
System (RAIDS), can accommodate most stream
inventory methods. It is available to BLM offices;
more detailed information can be obtained from the
BLM Service Center, Division of Resource Systems.

DISCUSSION

A recurring problem in BLM is the failure of
biologists to record their stream inventory data in
the data base. Monitoring changes in a stream eco-
system requires that baseline data are available in an
analyzed and recorded format. Inventory data must
be on a specific area and have repeatable variables,
otherwise monitoring will be of little value. Variables
that can be measured with good to excellent repeat-
ability over time are stream width, stream depth,

Table 4. Stream habitat attributes selected for field testing during development of a stream Habitat Quality
Index (HQI) for trout (Binns 1979).

Attribute Class

Physical

Chemical

Biological

Late summer stream flow1

Nitrate nitrogen

Streambank vegetation

Annual stream flow variation

Total alkalinity

Fish food abundance

Maximum summer stream
temperature

Total phosphorous

Fish food diversity

Water velocity

Total dissolved solids

Fish food type

Turbidity
Cover

Hydrogen ion

Stream width

Stream depth

Stream morphology
Eroding banks

Substrate

Bed material

Silt deposition

Attributes underlined were ones selected for

use in the HQI.

Streams

233

steambank undercut, and streambank angle. Variables
that have poor repeatability are percentage pools,
pool quality, percentage riffle, bank-to-bank width,
high stream water width, and stream rock content
(Table 5).

LAND-USE IMPACTS ON STREAM HABITAT

Land uses such as livestock grazing, timber har-
vest, road construction, mining; and recreation can
adversely affect stream habitat if good management
practices are ignored.

Timber harvest impacts are reduced by leaving
buffer strips of uncut trees in a 75-ft. or wider zone
on each side of the stream. Stream habitat can be
destroyed by snagging, log dragging, crossing with
heavy equipment, and slash and sawdust dumping.

Livestock grazing in the riparian zone can cause
streambank sloughing, reduced shrub and tree repro-
duction, compacted soils, and increased bacterial
count in the water.

Hydraulic mining affects all stream habitat fea-
tures as well as the riparian zone, especially if
streamside settling ponds are required. Mining explo-
ration should be conducted a safe distance away
from a stream to prevent erosion and degradation of
the streambed.

Recreation activities such as picnics and over-
night camping facilities should be carefully designed
to prevent water pollution and deterioration of the
riparian zone.

Road construction should be confined to areas
other than riparian zones. Culvert design must con-
form to accepted practices for fish passage.

Stream channelization destroys stream diversity
and causes long-term damage to stream productivity.
The impact of each of these activities on stream
habitat features is summarized in Table 6.

Streams are extremely important on arid lands.
The variety of uses received in stream/riparian areas
are compounded by the attractiveness of these areas
to all users. Livestock and wildlife congregate in
stream/riparian areas because of the availability of
food, cover, and water. Recreationists enjoy boating
and aesthetic values, and fishermen enjoy the pursuit
of their sport. Road construction is the least costly
in the riparian zone. Hydraulic mining must of neces-
sity be carried on in the streambed.

These competing uses require biologists to have
a data base on stream habitat conditions as well as
a knowledge of the impacts from all land uses on
such conditions.

Unregulated logging can destroy stream habitat.

Hydraulic mining residues can pollute main stem streams
in addition to side streams.

234

Streams

Table 5. Average expected repeatability, precision, and confidence intervals of water column and streambank
measurement means from six selected streams in Idaho and Nevada. Confidence intervals are at the 95%
level and expressed as percentage of the mean (from Platts 1981).

Confidence

Repeatability

Item

Interval

Precision

over Time

Water Column

Stream width

5.4

9

9

Stream depth

8.2

9

©-•

Streamside water depth

16.6

e

e-9

Pool (percent)

10.3

9

o

Pool (quality)

8.0

9

o-e

Riffle (percent)

12.5

e

o

Sun arc angle

1.1

•

9

Bank-to-bank width

Very wide

o

O

High water stream width

Very wide

o

o

Streambank

Soil alteration

12.3

e

Q-9

Vegetative stability

3.1

•

e-9

Undercut

18.5

e

9

Angle

4.4

•

9

Rock content

Very wide

o

o

Excellent 9 = Good 0 = Fair Q = Poor

Table 6. Land uses in stream/riparian zones and their impacts on stream habitat features (from American
Fisheries Society, Western Division 1982).

Stream Habitat Feature

Livestock
Grazing

Mining

Road
Construction

Stream
Channeli-
zation

Timber
Harvest

Recreation

(Fishing,

Picnicking)

Stream flow pattern,
increased runoff

Vegetation removal,
shade and cover
reduced

Stream temperature
increased

Streambank stability
reduced

•

Channel stability
reduced

Streambed siltation
increased

•

Stream width increased

Stream depth
decreased

Bacterial count
increased

•

Suspended solids
increased

•

•

•

Stream diversity
reduced

•

•

•

Streams

235

LITERATURE CITED

AMERICAN FISHERIES SOCIETY, WESTERN DIVISION.
1982. The best management practices for the man-
agement protection of western riparian stream eco-
systems. Riparian Habitat Committee, Western
Division, Am. Fisheries Soc. 45pp.

BINNS, N.A. 1979. A habitat quality index for Wyoming
trout streams. Fishery Research Report Monograph
Series, 2. Wyoming Game and Fish Dep., Cheyenne.
75pp.

BOEHNE, P.L. and R.A. HOUSE. 1983. Stream ordering: a
tool for land managers to classify western Oregon
streams. U.S. Dep. Inter., Bur. Land Manage. TN OR 3,
Oregon State Office, Portland. 6pp.

BOVEE, K.D. 1982. A guide to stream habitat analysis
using the instream flow incremental methodology.
Instream Flow Information Paper 12. Cooperative In-
stream Flow Service Group, U.S. Dep. Inter., Fish
and Wildl. Serv., Ft. Collins, CO. FWS/OBS-82/26.
248pp.

and R. MILHOUS. 1978. Hydraulic simulation in

instream flow studies: theory and techniques. IFIP 5,
Cooperative Instream Flow Service Group, U.S. Fish
and Wildl. Serv., Ft. Collins, CO. 1 30pp.

COWARDIN, L.M., V. CARTER, F.C GOLET, and ET.

LAROE. 1979. Classification of wetlands and deepwa-
ter habitats of the United States. U.S. Dep. Inter., Fish
and Wildl. Serv., Washington, DC. 103pp.

CROUSE, M.R. and R.R. KINDSCHY. 1981. A method for
predicting riparian vegetation potential of semiarid
rangelands. Pages 110-116 in Acquisition and Utiliza-
tion of Aquatic Habitat Information. Western Div. of
Amer. Fisheries Soc. Portland, OR.

CUPLIN, P. 1981. The use of large scale, color-infrared
photography for stream habitat and riparian vegeta-
tion inventory. U.S. Dep. Inter., Bur. Land Manage.,
Denver, CO. 57pp.

, B. VAN HAVEREN, R. BOROUICKA, J. ERDMANN,

L. LEE, R. MCQUISTEN, and M. WHITTINGTON.
1979. Instream flow guidelines. U.S. Dep. Inter., Bur.
Land Manage., Denver, CO. 57pp.

DUFF, DA. and J.L. COOPER. 1978. Techniques for con-
ducting stream habitat surveys on national resource
land. U.S. Dep. Inter., Bur. Land Manage. Denver, CO.
Tech. Note 283- 72pp.

HAMILTON, K. and E.P. BERGERSEN. 1984. Methods to
estimate aquatic habitat variables. U.S. Dep. Inter.,
Bur. Reclamation. Eng. Res. Center. Denver, CO.
1984.

HAYS, R.L., C. SUMMERS, and W. SEITZ. 1981. Estimating
wildlife habitat variables. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-81/47. 111pp.

HORTON, RE. 1945. Erosional development of streams
and their drainage basins — hydrophysical approach to
quantitative morphology. Bull. Geol. Soc. Amer.
56:275-370.

PARSONS, S.C and S. HUDSON. 1985. Stream channel

cross-section surveys and data analysis. Pam. 228. U.S.
Dep. Inter., Bur. Land Manage. Denver, CO. 48pp.

PLATTS, W.S. 1981. A plea for fishery habitat classification.
Fisheries 5( 1 ):2— 6.

SCHUMM, S.A. and D.F. MEYER. 1979. Morphology of
alluvial rivers of the Great Plains. Pages 9-14 in
Agric. Council Publ. 91. U.S. Dep. Agric, For. Serv.
Fort Collins, CO.

SHEN, H.W., S.A. SCHUMM, J.D. NELSON, DO. DOEHR-
ING, MM. SKINNER, and GL. SMITH. 1981. Assess-
ment of stream-related hazards to highways and
bridges. Federal Highway Admin., Washington, DC.
24 lpp.

STRAHLER, A.N. 1952. Dynamic basis of geomorphology.
Bull. Geol. Soc. Am. 63:923-938.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF
LAND MANAGEMENT. 1983. Public land statistics,
Vol. 168. U.S. Govt. Print. Off. Washington, DC.

U.S. DEPARTMENT OF THE INTERIOR, FISH AND

WILDLIFE SERVICE and U.S. DEPARTMENT OF COM-
MERCE. 1982. 1980 National survey of fishing, hunt-
ing, and wildlife-associated recreation. U.S. Govt.
Print. Off. Washington, DC.

U.S. ENVIRONMENTAL PROTECTION AGENCY. 1971.
Freshwater biology and pollution ecology training
manual. Training Manual, Water Quality Office, Wash-
ington, DC.

236

Streams

12

LAKES

James F. LaBounty

U.S. Bureau of Reclamation
Denver, CO 80225

"A lake . . . forms a little world within itself — a mi-
crocosm within which all elemental forces are at
work and the play of life goes on in full, but on so
small a scale as to bring it easily within the mental
grasp."

— Stephen A. Forbes

Editor's Note: This second chapter on aquatic habi-
tats covers bodies of standing water, including
lakes, ponds, and reservoirs. These habitats are as
important to the many aquatic and terrestrial spe-
cies that depend on them as flowing water is in
the arid western States. However, the techniques for
inventorying and monitoring these habitats and
their fauna are often different from those used for
flowing bodies of water. Entire books and complete
courses have been devoted to the study and meas-
urement of aquatic life in lakes. This chapter sum-
marizes this complex subject and provides
guidance for inventorying and monitoring of lakes.

INTRODUCTION

In a very simple sense, lakes are bodies of stand-
ing fresh water. In reality, they are complicated sys-
tems of living and nonliving features that influence
each other immensely. Lakes may be shallow or
deep, large or small, but the features that give them
their most distinctive characteristics are all related
to their standing water. A lake basin can be thought
of as a trap or catchment area, both for water and
materials from within the watershed. Typically, lakes
have inlets and outlets so that some of the water
and materials that enter a lake are discharged from
the lake at some future time.

Even though one can go back to a particular
lake year after year and it appears to remain the
same, it is constantly changing. Lakes are only tem-
porary features on the landscape. Most lakes in the
U.S. were formed between 10,000 and 12,000 years
ago after the last retreat of the glaciers ( Figure 1 ;
Odum 1971). Others were formed as a result of
volcanic activity or a change in the earth's crust
(e.g., earthquakes). Lakes will eventually fill with the
materials that accumulate and become a portion of
the terrestrial landscape. The life span of lakes varies
considerably, but the outcome is always the same.
The stage of development of a lake is very important
in preparing management strategies. Generally, there
is an increase in the complexity and diversity of life
within a lake as it ages.

Most consider reservoirs (Figure 2) to be just
another kind of lake, but there are two major differ-
ences between reservoirs and natural lakes that
should be considered. First, reservoirs are man-made
and their water levels are controlled. Many lakes
have control structures placed on their outlets. How-
ever, the degree of control between reservoirs and
lakes is very different. The deepest point in a reser-
voir is generally at its outlet — the dam. Therefore
it could be drained. On the other hand, the deepest
point of a lake is somewhere near its center. The
lake's natural outlet is logically the shallowest point

Lakes

237

\^*> . ..- *

Figure 1. A mountain lake at 9,200 feet in elevation in Colorado.

Figure 2. Seminoe Reservoir located in the high plains of southern Wyoming.
238 Lakes

in the body of water. Thus, without some major
excavation or catastrophic event such as an earth-
quake, a lake could not be emptied. These two dif-
ferences are very important considerations in putting
together management strategies for a body of water.

Ponds (Figure 3) can, in a small way, exhibit
all of the characteristics of a lake or a reservoir.
They play important roles in watershed management
in the western U.S.

Figure 3. A pond located along Lake Creek in
central Colorado.

CLASSIFICATION SYSTEMS

Lakes (Figure 1 ) possess a great variety of com-
binations of properties (Odum 1971). This makes
it difficult to select any one basis to classify lakes.
Therefore, one can find all sorts of terminology ap-
plied to lake types. To illustrate this, Hutchinson
(1957) listed at least 75 lake types based on geo-
morphology and origin. In this chapter, Odum's

1971 use of three-lake categorization schemes will
introduce the most commonly encountered termi-
nology as applied to lake types:

( 1 ) the oligotrophic-eutrophic series of ordinary,
clear-water lakes based on productivity;

(2) special lake types; and

(3) impoundments.

Oligotrophic — Eutrophic Series

These terms, implying productivity or fertility,
are probably the most commonly used to describe
lakes. The classic oligotrophic lake is one that has
low primary productivity and is relatively deep. In
contrast, eutrophic ("good foods") lakes are shal-
lower and have a greater primary productivity. The
term mesotrophic implies moderate productivity,
meaning the lake falls in a category somewhere be-
tween oligotrophic and eutrophic.

Other categories — ultraoligotrophic, oligomeso-
trophic, and hypereutrophic — modify the above-
mentioned terms. However, emphasis should remain
on using the oligotrophic, mesotrophic, and eu-
trophic terms because they will suffice when refer-
ring to a lake's productivity. Many times the term
eutrophic is improperly used as a synonym for pol-
luted. Pollution does not always correlate with the
process of eutrophication when referring to increas-
ing the nutrient (nitrogen and phosphorus) content
of a lake from sewage effluent. A lake can, in fact,
be "polluted" by a heavy metal or other toxic sub-
stance and become more oligotrophic. Care and
thought should then be made before using the term
eutrophic in describing a lake, even though in-
creased eutrophication of a lake frequently results
from untreated or little-treated effluent, such as run-
off from a cattle feed lot or nearby town.

In relatively natural situations, lakes become
more eutrophic as they age. Another general rule is
that there is a trend of increasing productivity (or
eutrophication) with decreasing depth of a lake.
Therefore, a rule of thumb is that eutrophic lakes are
shallow and have greater primary productivity, and
oligotrophic lakes are deep and have lower
productivity.

Special Lake Types

Odum ( 1971 ) listed seven special lake types:

( 1 ) Dystrophic lakes generally have high concen-
trations of humic acid.The water appears
brown and the lakes eventually develop into
peat bogs.

Lakes

239

(2) Deep, ancient lakes contain animals that are
found nowhere else on earth.

(3) Desert salt lakes occur in sedimentary drain-
ages in arid climates where evaporation ex-
ceeds precipitation. They are sumps for
drainages where salts build up that have no
outlets such as the Great Salt Lake. Communi-
ties of few species are the rule for these
lakes.

(4) Desert alkali lakes occur in igneous drainages
in arid climates. Being alkaline, their pH val-
ues and concentrations of carbonates are
high. Pyramid Lake, Nevada is an example.

( 5 ) Volcanic lakes occur in areas with active
volcanoes and may be acid or alkaline.

(6) Chemically stratified meromictic or partly-
mixed lakes become permanently stratified
due to such things as intrusion of saline
water or salts liberated from sediments. This
sets up a permanent density difference be-
tween surface and bottom waters. The
boundary between the upper, circulating
waters and the lower, non-circulating layer is
termed a chemocline. The bottom layers are
free of any organisms. Big Soda Lake, Nevada
is an example of a meromictic lake. In addi-
tion, Flaming Gorge Reservoir on the Green
River of Utah and Wyoming had a chemo-
cline from the time it was filled in 1967 until
1982. The chemocline resulted from salts
being liberated from the sediments.

(7) Polar or alpine lakes have surface tempera-
tures that remain below 4°C (39°F) except
for brief periods during the ice-free summer.
Plankton populations grow rapidly and store
fat for the long winter. These lakes are not
productive enough to obtain optimum
growth of fish.

Impoundments

Impoundments (Figure 2) are artificial lakes
created by placing a dam on a river or stream. Bea-
ver ponds are the smallest example in this category.
A tall concrete dam such as the Hoover and Glen
Canyon dams on the Colorado River are examples of
the other extreme. The limnological characteristics
of an impoundment vary according to the nature
of the drainage, the climate, and probably most of
all, the operation of the reservoir. Wright (1967)
and Odum ( 1971 ) listed a number of effects of
dams. Some of the more notable include the fact that
water is released from the surface of natural lakes

and from the bottom of most reservoirs. Therefore,
the releases from the deep layers of a reservoir will
have higher salinities, be colder during the ice-free
period, have a lower concentration of dissolved oxy-
gen, be higher in some of the reducing compounds
(such as hydrogen sulfide), and have higher nutrient
concentrations than would normally be stored in a
natural lake. The advent of using multilevel outlets
on dams has resulted in the moderation of some
of the above effects. Nevertheless, water levels of
reservoirs can fluctuate greatly whereas lake levels
remain relatively stable. This operational difference,
along with all the ramifications, puts impoundments
in a category of lakes that require different manage-
ment techniques than natural lakes.

CRITICAL HABITAT FEATURES

The physical feature that exerts the most direct
control on a lake's characteristics is its thermal strati-
fication. The maximum density of pure water is just
under 4°C (39°F). All fresh water at any other tem-
perature is lighter and found on the surface. Thus,
ice floats and in summer, the upper layer of a lake is
warmest. The thermal regime of a lake definitely
reflects its ambient temperatures as they fluctuate
over a day and over a season.

The following is a scenario of a temperate lake.
The example will also be used to define some terms.
Consider a deep lake whose temperature in late
spring is fairly uniform from top to bottom at about
4°C (39°F). As the day lengthens, solar radiation is
absorbed by the water, mostly by the upper layers of
the lake. Light intensity decreases with increasing
depth until the point is reached where complete
darkness occurs. Wind mixes the surface water
rather thoroughly but does not reach down to the
deeper water. As heating and mixing of the surface
layers continue to midsummer, a typical pattern
of temperature distribution develops (Figure 4). The
warm top layer, which is heated by the sun and
mixed by the wind and other currents, is called the
epi limn ion. The bottom layer (neither heated by
the sun nor mixed by the wind) is called the hypo-
limnion. The transition between the two is the
metalimnion or thermocline. The thermocline is
generally recognized where the water temperature
drops at a rate of 1°C per m of water depth, al-
though this is a textbook definition and may in real-
ity be better described as a biological barrier.

Once a thermocline has formed, the two layers
are virtually separated and no water is exchanged
between them; even the current patterns within
them are different. These layers are maintained
strictly because warmer water is lighter than colder
water. As fall progresses, heat is lost from the epilim-
nion to the atmosphere faster than it is absorbed
and the temperature drops. At the same time, the

240

Lakes

4° Temperature

u

CD

5

/ Epilimnion

^ Metalimnion, or
^_^--~^^~^ Thermocline

o

c

a
CD

2

Hypolimnion

Figure 4. Typical pattern of temperature distribu-
tion in a deep lake in the summer zone during
the summer.

thermocline continues to sink. The difference be-
tween water temperatures at the surface and at the
bottom steadily decreases until the thermocline no
longer exists and the lake "turns over" as the wind
mixes it.

As winter intensifies, stratification, in reverse of
summer, is setup. Ice forms on the surface of the
water at about 0°C (32°F) and temperatures at the
bottom are about 4°C (39°F). In typical temperate-
zone lakes, the winter stratification is not as strongly
developed as the summer stratification; and in some
areas ice never forms and the lakes, more or less,
continually turn over the entire winter due to wind
driving their circulation.

In spring, the lake again absorbs heat from the
sun and its temperature rises until it is uniform at
about 4°C (39°F) from top to bottom. The cycle
begins again. Lakes that undergo two turnovers per
year are called dimictic. Those that turn over only
once and never manage to form ice are called mon-
omictic. There is a range of lakes found in each of
the two categories.

The determining factors for the complexity of
life in a lake are a combination of the physical and
chemical factors in and around a lake. Factors other
than temperature include the basin configuration,
depth of lake, water clarity, dissolved oxygen and
dissolved solids, inflow-outflow volume versus lake
volume (flushing rate), nutrient (e.g., nitrogen, phos-
phorus, and silica), and a myriad of other chemical
and physical factors including many caused by man.

Lakes are ephemeral features, considering the
geologic time scale. The length of their life is ex-
tremely variable, but they all progress toward the eu-
trophic state. This rate of change is influenced
tremendously by man's activities. Water is a precious
commodity in the western U.S. and numerous factors
compete for the limited supply. Therefore, man has
varying degrees of influence on all the lakes, ponds,
and reservoirs in the western U.S. The watershed
that supplies water to lakes may have cattle grazing,
which causes added nutrient input and erosion of
streambeds from vegetation loss. Added nutrient
input tends to increase eutrophication of a lake
whereas erosion of streambeds adds turbidity. Min-
ing activities in a watershed may add heavy metals to
a lake. Acid mine pollution can essentially kill a lake.
In some areas, acid mine effluent from old mine
shafts or tailing piles has been continually polluting
area streams and lakes for over 100 years. The ulti-
mate influence of this acid mine effluent is to ac-
tually suppress a lake's development, keeping its
living component at a minimum.

Proper management of the watershed will avoid
disasters and reduce problems that may later be
too expensive or impractical to fix. The paramount
consideration is that a lake is only part of the entire
watershed; therefore, proper management of the lake
alone will not be enough.

Different lakes and even different parts of a lake
have different communities of animal and plant life.
These differences are functions of many things, in-
cluding lake size and depth, substrate, light, thermal
stratification, and geographic location. All of these
factors should be considered in the management
scheme of a lake.

MAIN LIFE ZONES OF A LAKE

Figure 5 diagrams the main life zones within a
lake. All open water, where the bottom is too deep
to be inhabited by living plants, is termed the pe-
lagic zone. Within the pelagic zone there are two
subzones: the limnetic, where there is enough light
for photosynthesis to occur, and the profundal,
the dark bottom layers. The layer of most biological
importance in the pelagic zone is above the compen-
sation point where light is just sufficient to produce
exactly what is used up (e.g., photosynthesis = de-
composition). Above this level, autotrophs (such
as algae) can produce food through photosynthesis.
These autotrophs form the base of the food web
as they are the main source of primary production
for the lake. Diversity of life in the limnetic zone
is greatest. Below the compensation level is the pro-
fundal zone. Here autotrophs cannot produce food
(lack of enough light) and the main source of energy
is detritus that "rains" out of the limnetic zone. All
organisms in the profundal zone are called hetero-
trophs and are either detritus feeders or carnivores.

Lakes

241

Eulittoral

Infralittora

Figure 5. The main life zones of a lake.

One final way a lake is partitioned is by the
bottom. The lake bottom may be of original rock in
younger lakes; in older lakes the bottom is covered
with sediments to form a uniform substrate of mud
or sand. The area of bottom below the compensation
level is the profundal benthic zone. The area from
the lower limit of rooted vegetation to the shore is
known as the littoral zone. The littoral zone has the
lake's greatest diversity both in habitat and commu-
nity. The littoral zone is further broken down into
the infralittoral zone, which is always under water
and contains rooted vegetation; and the eulittoral
zone, which is nearest the shore and may be cov-
ered by water only part of the time. Much erosion
can occur in this zone as agitation is greatest.

MAJOR SPECIES GROUPS

Lakes are used for many forms of recreation.
These include swimming, boating, water skiing, sail-
ing and, of course, fishing. As the population of the
western U.S. increases, so does the fishing pressure
on its limited number of lakes, ponds, and reservoirs.
The fishery must then be closely managed by stock-
ing and other techniques. Above all other forms of
life, fish are generally the primary concern. To put
this into a proper perspective, along with the fish
themselves, other aspects about a particular lake
govern the fishery and must be considered and
known.

The variety of animal and plant assemblages
found in lakes is as great as the variety of different
kinds of lakes. In addition, different parts of a lake
have different characteristics and therefore have
different animal and plant communities. Probably the
most influential factor in determining the kind of
plant and animal communities that occur in a lake is

temperature. One commonly encounters the terms
cold-water fishery, cool-water fishery, and warm-
water fishery. The term two-story fishery is also com-
monly used by fishery biologists. The following are
rules of thumb: A cold-water fishery in the western
U.S. usually means a trout fishery. Therefore, it could
include any kind of native or introduced salmonid
such as rainbow (Salmo gairdneri), brown (5. trutta),
or brook trout (Salvelinus fontinalis), or a variety of
salmon. A cool-water fishery includes walleye (Sti-
nostedion rutreum), sauger (5. canadense), yellow
perch (Perca flavescens), northern pike (Esox lu-
cius), and muskellunge (Esox masquinongy) (Tran-
dahl 1978). A warm -water fishery in the western U.S.
usually includes spiny-rayed fishes, such as sunfish
(Centrarchidae) and bass (Micropterus sp.), and also
catfish, bullheads (Ictaluridae), and carp (Cyprinus
carpio). A two-story fishery is one that includes both
cold-water and warm-water fishes.

The situation for a two-story fishery is special in
that thermal stratification during the summer is
strong enough that the epilimnion is generally above
20° to 25°C (68° to 77°F) in midsummer while the
hypolimnion (lower layer) is around 10° to 15°C
(50° to 59°F). For a successful two-story fishery,
thermal conditions during winter cannot be so se-
vere as to violate the ecological requirements of
warm-water species. In addition, the proper kinds
and amounts of food must be available for the cold-
water and warm-water fish. Thus, considerable infor-
mation on a particular lake must be available to de-
termine if it will sustain a two-story fishery. All of
the fisheries mentioned are found throughout the
western U.S.

Both recreationally and commercially, fish are
the most important living component in a lake. How-
ever, other living components within the lake dictate

242

Lakes

what the fishery is and how successful it is (or could
be if managed correctly). The differences in commu-
nities between lakes and within any particular lake
are functions of many other physical and chemical
factors besides temperature. These include substrate,
light, cation-anion concentrations, and geographic
position of communities within the lake. The com-
munities within the different life zones (e.g., lim-
netic, profundal, pelagic, and littoral ) of a lake are
not completely separate, as some species overlap
two or more zones.

The upper portion (limnetic zone) of the open
water (pelagic) area is of the most obvious biologi-
cal importance, as it is where autotrophs can pro-
duce food through photosynthesis. The simplest
form of life in the food chain of a lake would most
likely be found in its upper zone. The autotroph
is algae that converts the sun's energy and inorganic
material to living material. There are many kinds of
algae, ranging from diatoms dominating in lakes of
low production to blue-greens that dominate more
productive lakes during midsummer. The algae are
grazed on by microscopic animals collectively called
zooplankton. These range from relatively small roti-
fers (10 microns) to fresh-water shrimp (20-30 mm
[1 in.] long) that feed on the smaller zooplankton.

Zooplankton provide the food base for fish.
Some species of fish only eat zooplankton for a short
time in their life cycle, while they are very small.
Other fish (commonly called filter feeders because
they filter the water) feed on zooplankton their en-
tire life. The next predator in this simple food chain
is a fish that feeds on other fish. An example is a
largemouth bass feeding on minnows. Man, other
fish-eating mammals, or some fish-eating birds would
be the final link in this food chain. The following
illustration is a sample food chain.

Inorganic Material

I

Light -► — Algae

I

Zooplankton (small)

I

Zooplankton (large)

T

Fish (small)

i

Fish (large)

t

Mammals or Birds

In reality, the community of a lake is a food
web. That is, there are many kinds and species of
plants and animals, and a variety of conditions and
factors that result in a web pattern versus the simple
chain described above. The following illustration is
a sample food web.

"^.Inorganic Material^

ALGAE
(Diatom)
7 species

T

ALGAE

(Green)

4 species

ALGAE

(Yellow-brown)

2 species

~1
ALGAE
(Blue-green)
2 species

* i ir

ZOOPLANKTON
(Rotifers)

ZOOPLANKTON
(Copepods)

tTI!

ZOOPLANKTON
(Cladoceran)

1 f

ZOOPLANKTON
(Shrimp)
1 species

-FISH-

t\-

-^FISH---

1
-FISH

(Kokanee)

(Lake Trout)-

(Rainbow Trout)—1

I

-MAN

The benthic and littoral components of the
community also play significant roles in the food
web. The benthic community may consist of insect
larvae, true worms, and clams. These animals break
down the organic material (e.g., dead fish) that falls
to the bottom into the inorganic material that reen-
ters the food web when the lake circulates at spring
and fall turnover. The littoral zone contains an even
greater diversity of insect larvae, worms, clams, and
rooted and floating aquatic plants. The littoral zone
often provides the most diverse aquatic habitat
within a lake; it also may have the most complex
food web of a lake.

INVENTORY AND MONITORING METHODS
FOR LAKES

A number of useful manuals and books can be
consulted for methodology of studying lakes. They
range from sections of limnology and fishery text
books to entire manuals devoted to methods. Some
of the more common manuals or textbooks that
include methods are listed in Appendix 1 .

The U.S. Geological Survey publication entitled
National Handbook of Recommended Methods
for Water Data Acquisition ( 1977) is considered
to be the "bible" by users and collectors of data
relating to hydrology, chemical quality, groundwater,
sediment, biological quality soil water, evapotranspi-
ration, and hydrometeorology. It should be referred
to by anyone surveying or monitoring lakes, ponds,
reservoirs, or streams. Compiled by agencies that
acquire information on or related to water in the
U.S., it contains standardized methodology to assure

Lakes

243

greater comparability, compatibility, and usability
of water data. It is easily obtainable from the Office
of Water Data Coordination, U.S. Geological Survey,
Department of the Interior, Reston, VA 22092.

Levels of Inventory or Monitoring

There are, of course, several ways (and levels of
effort) to conduct limnological or fishery surveys of
lakes. The goals and needs should first be identified,
which will dictate both the type of survey and the
level of effort. If, for example, several hundred lakes
need to be characterized as eutrophic versus oligo-
trophic, a remote sensing survey with a minimum of
ground-truthing during midsummer could provide
the information necessary. It should be cautioned,
however, that in almost every situation there would
not be enough information provided from a survey
of this type to make judgments as to the fishery.
Also, a lake one year could be classified as eutrophic,
whereas another year it could not.

The most difficult level of effort would include a
long-term (5-year), twice-monthly to monthly, sur-
vey of the physical, chemical, and biological limnol-
ogy and frequent netting or creel census data
collection of each lake or lakes. The long-term sur-
vey could involve expenditures of more than
$200,000 per year, whereas the remote sensing sur-
vey may cost only a few hundred dollars. Although a
prudent manager is likely to opt for the remote sen-
sing survey, the needs and long-term usefulness of
the data should be considered. The following are
descriptions of several levels of effort that can be
chosen or blended to fit a needed situation. Level I is
the least amount of effort; Level IV the most.

The first thing that needs to be determined is
why a survey is needed. When that question is an-
swered, the tests to be run and samples to be col-
lected can be planned. Upon completion of the
reconnaissance survey, when the data are in hand, it
is important for the manager to realize the limitation
of the data. That is, a one-shot collection of data
during midsummer reveals little of the lake's ecology
during winter or even for the month before and
after the collection of data. Nevertheless, remember
that by doing the reconnaissance survey, managers
can know things about the lake not previously
known. As long as the data are kept in proper per-
spective, such a survey is useful.

Level I — Minimum Effort Inventory. This level,
which is a simple inventory, can also be described as
a reconnaissance survey. Reconnaissance surveys
are usually used to determine the general ecological
trend of a system and should be conducted in
midsummer. They may also be used to identify areas
in which more detailed or synoptic surveys should
be done.

A Level I survey can be approached in various
ways, depending upon the need. If there is need
to simply characterize a lake as to cold water versus
warm water, measuring depth and surface and bot-
tom temperatures during midsummer may be suffi-
cient. If there is need to characterize the trophic
status of a lake or several lakes, remote sensing may
be used, or carefully collected and properly handled
water samples may be analyzed for phosphorus and
nitrogen concentrations.

If there is a need to know the quality of a lake's
fish food base (i.e., plankton), a top to bottom plank-
ton sample needs to be collected with a proper
plankton net and analyzed. If some information on a
lake's fishery needs to be known, creel-censusing
(interviewing fishermen) or spot-netting with appro-
priate gill nets is necessary.

All of the above are examples of information
that would be obtained at a "one-shot" reconnais-
sance level for a specific need. It may be that several
things need to be known about a lake; therefore, a
reconnaissance survey would include many of the
above tests. It must be remembered that there are
several approaches to obtaining information for any
one particular need. I recommend that the National
Handbook of Recommended Methods for Water-
Data Acquisition (U.S. Geological Survey 1977) be
consulted for each test.

Level II — Maximum Level Inventory. This level
would also be classified as a reconnaissance survey,
with several tests done to obtain more knowledge of
the subject lake. For example, suppose a manager
wanted to know the general characteristics of four
or five lakes so that future studies could be planned
to develop an intensively managed fishery. An
approach I recommend would involve not more than
four workdays per lake, preferably on the same day.
Each lake's survey would be in two parts — fishery
and limnology. The limnology would involve
choosing the deepest part (or parts if it is relatively
large ) of the lake and collecting some critical profile
data and samples there. A profile of temperatures
each 1-2 m (3. 3-6.6 ft) from surface to bottom is
essential; dissolved oxygen concentration, at least at
the bottom, is desirable.

Also, some information on other water quality
variables is desirable. A determination must be made
as to what kind of water quality data are needed
and what is to be done with them. This will deter-
mine how many samples are collected and how
these samples are handled. At a minimum, a sample
or samples of water should be collected from at least
three depths — surface, mid-depth (thermocline),
and near-bottom.

Nitrogen and phosphorous concentrations are
always important data to collect as they indicate

244

Lakes

trophic status of a body of water. The samples for
these nutrient analyses must be collected in clean,
approved plastic bottles, with no contamination from
handling. Usually 0.5 L (0.5 qt) is a sufficient sample
for these analyses. After collection of the samples,
they need to be analyzed by a qualified laboratory or
preserved by freezing until they are sent to a
laboratory.

Next, a sample of the lake's plankton should be
collected with a sock-like plankton net which is
pulled from bottom to top or within layers of the
lake (e.g., 5 m [16 ft] to surface; 10 m [33 ft] to sur-
face), depending on the equipment available and
the need for a detailed study. These samples provide
information on the food base for fish. The samples
need to be placed in acceptable bottles (0.5 L [0.5
qt]) and preserved with several drops of formalin or
other preservative. The samples are then analyzed
qualitatively and quantitatively, if desired, in the
laboratory. A binocular, low-power microscope,
standard counting cell, and simple pipette are
needed to do these analyses.

Next, benthic samples of the lake could be col-
lected. This involves using one of several types of
dredges, a standard No. 30 sieve bucket, and some
forceps. The bottom mud is collected, washed in the
sieve bucket, and the sample plus residue preserved
in a bottle with a few milliliters of formalin. These
samples will reveal many things about the lake, such
as the benthos food base, past stratification strength,
hypolimnion chemistry, and general status of the
lake. These samples would also be analyzed in a
laboratory.

Finally, in this brief reconnaissance limnological
survey, some indication of light penetration is desira-
ble. The simplest way to do this is with a Secchi
disc. This is a circular metal plate, 20 cm (8 in.) in
diameter, the upper surface of which is divided into
quarter sections painted black and white. It is at-
tached to a line by a center eyebolt. The point at
which the eyebolt can no longer be seen upon low-
ering is the Secchi depth. This simple piece of infor-
mation is extremely valuable since light is one of the
important qualities of the lake's limnology. The lake
in question can be easily compared to others by this
method.

The second portion of this reconnaissance is to
examine the fishery. If no data exist, a gill-net survey
is a good place to start. This involves first having
the proper collecting permit, which should be ob-
tained well ahead of time from the state game and
fish agencies. Assistance from the agencies is desira-
ble and often essential. They can assist in using
proper equipment, such as gill nets, shocking boats,
and beach seines. Without their involvement, a sur-
vey of the fishery will most likely lack some impor-
tant knowledge.

Often, a one-night set of experimental gill nets
in some strategic locations will yield information
that will help managers make some preliminary deci-
sions as to what kind of a fishery exists and what
needs to be done. Sometimes beach seining during
late summer or early fall can give fishery biologists
and ichthyologists a good indication of the type of
fishery existing in a lake where little is known. The
Level II survey involves much preparation and fol-
low-up to be done properly.

Level HI — Minimum Effort Monitoring. This
level of effort involves part or all of the tasks
described under Level II except they are performed
on a seasonal basis, that is, four times a year at mid-
season. Again, details of the study depend on the
questions being asked. The Level III studies are the
first stage of synoptic or detailed studies. Instead
of an inventory, a monitoring study is being
performed. These synoptic or detailed studies are
done to determine ecological baselines from which
change can be determined in the future. In general,
the premise on which synoptic surveys are based
is that an evaluation of kinds, numbers, and relative
abundance of species enables one to establish the
biological health of a lake, pond, or reservoir (or any
other ecological system ). As with the other levels of
effort, this monitoring should be attempted only
after careful planning and preparation. If several
lakes are involved, some coordination is beneficial so
that inter-lake comparisons can be made.

If the needs dictate a Level III effort, costs can
add up to tens of thousands of dollars per lake per
year. However, if there is only a need to monitor
lake temperatures, all that is needed is a boat and
motor, a thermometer probe with enough cable to
reach the bottom, and a readout. If plankton species
need to be documented, all that is needed is a boat
and motor, a plankton net, a collection bottle, and
some preservative. Seasonal samples of plankton
would then reveal not only if certain kinds required
for certain species of fish are present, they would
reveal many things about the lake's health.

Level IV — Maximum Effort Monitoring. Level IV
should be chosen only when a highly synoptic set of
data is needed. Monitoring studies at this level
become more involved and more frequent. The main
thing is that they are more synoptic, and routine
field sampling dates need to be selected. An example
would be every other month, or monthly during
turnover, and every third month at other times. A
resource specialist or manager needs to ask why the
data are needed, when the most significant
information can be collected, and how much and
what kind of data are needed to solve the problem.
It cannot be overemphasized that the studies must
constantly be adjusted and fine-tuned to keep them
in a proper perspective. It is common for a resource

Lakes

245

specialist to convince a manager of the need to
perform a certain study, yet the manager may not
understand the utility of collecting a lot of data. Both
resource specialists and managers should agree how
the data are to be used by their agency.

For example, the need may exist to monitor a
lake that is involved in a water resources develop-
ment scheme. Suppose that its ecosystem is likely to
be modified considerably. For this example, consider
that the lake is a natural subalpine lake with a sur-
face area of about 700 ha (1,750 a.). Furthermore,
consider that the lake is to be raised behind a newly
constructed dam. In addition, a pumped storage
power plant will be constructed that will pump lake
water 150 m (495 ft) up to a newly constructed
reservoir. During the generation phase, water will
pass from the upper reservoir through the power
plant back to the natural lake. Current environmental
laws require that the environmental effects be recog-
nized and some values protected. Therefore, the
ecosystem must be understood. Limnology and fish-
ery investigations would be necessary before, during,
and after changes were made. The lake's food chain
should be studied in enough detail to be able to first
predict, then detail and quantify changes to the lake
that are accountable to the development project.

The actual level of detail of each study will vary
from location to location and will be dictated by
needs identified by resource specialists. The manuals
and textbooks listed in Literature Cited should be
referred to as a guide in determining how the out-
lined needs will be met. One of the more difficult
tasks for a manager is knowing how much and what
kind of information are necessary for a particular
situation. Many times, not enough information is col-
lected to provide managers or resource specialists
with the evidence they need to make a critical deci-
sion. This not only can do an injustice to the re-
source, it can be embarrassing for the decisionmaker.
On the other hand, some studies become "gold-
plated" and the information may not apply to the
need. An example may be that an adequate amount
of chemistry data is available on a particular lake for
the manager to determine if the inflow actually is
contributing to the lake pollution. However, these
data need summarization by a resource specialist so
that managers can use them in their decisions. Sup-
pose the investigators selected to do the summariza-
tion are involved deeply in computer modeling and
their product is a detailed and highly technical
model of the lake. There have been instances when
the manager has not been able to use this product
because it is too technical.

The more practical way to handle this situation
would be to have the summarization done by a re-
source specialist who will also know the practical

need. The manager may be less technically qualified
than a professional statistician or modeler, but the
manager's grasp of the situation will allow the data
to be useful and allow a decision to be made. This
does not mean that models are not useful. The point
is that there will always be difficulty in knowing
exactly how much information is needed. A consen-
sus of resource personnel and management should
be sought as to how much information is necessary
to make a wise decision regarding a lake, pond, or
reservoir.

Field Techniques

This section includes a summarization of some
of the accepted field methods and the types of gear
needed for the various studies that are done for the
levels described in the previous section. Three as-
pects of a survey are used to facilitate this presenta-
tion: physical-chemical limnology, biological
limnology, and fishery.

For all the studies listed below, a boat, outboard
motor, field book, and pen or pencil are required.
At all times, approved life jackets or vests are re-
quired equipment for a survey. The following sum-
marizes the requirements for specific tests.

Physical-chemical surveys. Collecting profile
information on a lake requires some instrument s)
that measures the desired features. The most
primitive way is to lower a water bottle (described
below) to each desired depth; collect a sample;
bring it to the surface; and measure the temperature,
conductance, and dissolved oxygen (DO) of water
in the bottles with a simple hand-held thermometer
and other portable gear. This method is acceptable if
errors of up to several degrees temperature can be
tolerated. Also, the methodology for DO
measurement requires clear glass bottles that hold
about 300 ml (9 oz) and have glass stoppers (called
BOD bottles), and several chemicals. In many
situations, this sampling method is adequate, but in
others, more sophisticated gear is needed. This gear
may be very complex electronic equipment that
measures five or six variables at once (temperature,
pH, DO, conductance, oxidation-reduction potential,
and depth) and stores the data on internal memory
chips.

Practically speaking, a multiparameter probe
(see Figure 6) is a fairly basic piece of equipment
where routine lake surveys are being done. These in-
struments cost from $4,000 to $8,000, but are accu-
rate and save much time in the field and laboratory.

Water chemistry data generally require that
water samples be collected, placed into acceptable

246

Lakes

Figure 6. A multiparameter probe is used to measure profile data (courtesy of Hydrolab Corp.).
( 1 ) Digital Display Unit, (2) Sohole Unit, (3) Circulatory Assembly, (4) Cables, (5) Battery Pack,
and (6) Digital Field Datalogger.

containers, preserved, and taken to a qualified labora-
tory for analyses. Portable water chemistry kits are
available and the data obtained from their use are
acceptable for some measurements when the analyst
is qualified. However, for synoptic studies or where
data are required for critical decisions, the samples
should be carefully collected and analyzed by a qual-
ified laboratory.

Ideally, the sample bottles should be provided
by the laboratory that does the analyses. The labora-
tory should be consulted as to the amount of water
they need for the requested analysis. A water bottle
of the Van Dorn or Kemmerer types (see Figure
7) is necessary if more than surface samples are to
be collected. The water bottle must be attached to a
reel of cable of sufficient length. Along with the
water bottle, a messenger or something to trigger or
close the bottle at the desired sampling depth is
needed. Samples for chemical analyses should in-
clude both inflow and outflow water to determine
quantitatively the chemical elements that flow in and
out of a lake or reservoir and those that are retained.
Finally, the proper preservative is needed. For nor-
mal cation and anion analysis, only refrigeration is
needed; for nitrogen and phosphorus nutrients, freez-
ing is best; and for most heavy metal analyses, lower-
ing the pH with nitric acid below a value of 2 is
recommended.

Figure 7. A Kemmerer water sample bottle in the
open position.

Lakes

247

The amount of light that enters a lake is an im-
portant quality. It can be measured easily with a
Secchi disc. If more quantitative information on light
is desired, instrumentation called limnophotometers
may be used. These cost between $3,000 and $5,000
and actually measure light in units as the photocell
is lowered into the water.

Secchi disc

The following list of equipment will be sufficient
to perform a physical-chemical survey:

• map of the lake (optional);

• reel with enough cable or rope to reach desired
depth;

• thermometer, multiparameter probe, or what-
ever is appropriate to measure temperature,
dissolved oxygen, conductance, pH, redox po-
tential, and depth;

• water sample bottle (Kemmerer or Van Dorn
type);

• appropriate containers to hold samples for
water analysis;

• appropriate preservative; and

• messenger (heavy metal weight made especially
to trigger sampling gear).

Biological Limnology. The selection of equipment
here is as broad as the kinds of data that need to be
collected. The following is by no means exhaustive.
A limnological text should be consulted before
proceeding. Some examples are nevertheless given.

Primary production is commonly measured by
two methods — chlorophyll or light versus dark bot-
tle studies. An analysis of chlorophyll pigments in
a known volume of lake water will provide an indi-
cation of the algae biomass. This then gives direct
evidence of the amount of primary products in the
lake. The entire process involves access to a labora-
tory with a filtration system, spectrophotometer, test
tubes, and acetone. Water is collected and stored
for a short period until the samples are filtered (usu-

ally with a pressure filter ) through recommended
fiberglass filters. These are kept frozen until they are
put into a test tube of acetone and crushed. The
test tubes are centrifuged and the decanted sample is
put into a spectrophotometer. Light penetration at
wavelengths 630, 645, and 667 is recorded and the
results put into equations found in Wood (1975)
or Maitland (1978).

An actual measurement of photosynthetic rate
may be done in several ways. Most involve compar-
ing activity in light versus dark BOD bottles left in
the lake at desired depths for specified times. One
common method is to measure the DO before and
after a predetermined time period in both dark and
light bottles. The difference in dissolved oxygen
between the light and dark bottle then represents an
indication of the amount of primary production.
Specific methodology for this is available in Wood
(1975). Another commonly used method is even
more direct and involves spiking the light and dark
bottles with a known concentration of radioisotope
tagged source of carbon (NaH ^C03). The theory
is that algae during photosyntheses use carbon- 1 2
and by putting tagged carbon- 14 in, filtering the
algae later, and measuring the amount of radioactiv-
ity, the amount of carbon uptake by algae will be
known. The amount of carbon assimilation then indi-
cates the primary productivity rate at the time of
the survey. The carbon- 14 methodology requires ex-
pensive laboratory equipment, a radioactive source,
and individuals licensed by the Nuclear Regulatory
Commission. Nevertheless, when done properly, it is
the best indication of a lake's primary production
rate.

The next step is to collect the plankton (both
phytoplankton and zooplankton) and measure densi-
ties (i.e., number per volume). Again, there are nu-
merous ways to accomplish this task. Probably the
most common method is to use a closing net (Figure
8). This is a sock-like net that is lowered into the
water column to various depths. It strains the water
as it is raised through the water column. A cup at
the lower end is emptied into a collection jar and
preserved with a few drops of formalin, alcohol, or
other preservative.

A subsample (e.g., 1 ml) of the plankton is put
into a counting cell for counting and identification
under a microscope. The densities of each kind of
plankton are then calculated. Other commonly used
methods include towing a metered Clark-Bumpus
plankton net horizontally a short distance across the
lake or collecting a known volume of water with
the water bottle and pouring it through a net. All of
these involve counting and identifying the samples in
the laboratory by qualified personnel. A note of cau-
tion: some of the zooplankton may not be there
when and where sampling is done. For example, the
freshwater shrimp (Mysis relicta) that live in cold-

248

Lakes

water lakes are nocturnal and will not be collected
during daylight hours. Therefore, adequate sampling
techniques would include spatial and temporal
considerations.

The benthos of a lake are collected by gathering
a known area of bottom mud and straining it
through a sieve. Sampling gear includes Ponar, Peter-
son, and Ekman dredges (Figures 9 and 10). They
all have their most appropriate uses — an Ekman for
shallow or soft mud, Peterson for hard lake bottom,
and Ponar for deep lakes of either type.

Figure 9. Ponar dredge in the open position.

Figure 8. The closing style plankton net —

(b)

detatchable bucket

(cs)

canvas sleeve

(h)

headpiece

(lr)

lower ring

(m)

messenger

(r)

release

(s)

sleeve of silk bolting cloth or other

suitable material

(ur)

upper ring

Figure 10. Ekman dredge, open and closed
positions.

Lakes

249

Samples are preserved in formalin and analyzed
in the laboratory (identification and enumeration).
Qualified personnel are required for the laboratory
procedure.

One aspect of a lake's limnology that is fre-
quently ignored is the emergent and submergent
plants (macrophytes). This is because it is difficult to
collect information on these, especially underwater.
Usually found in the shallow waters of littoral zones
of lakes, these plants may be free-floating or at-
tached. They serve as habitat for a variety of other
aquatic organisms. The contribution of macrophytes
to the primary production of lakes, ponds, and even
reservoirs has been grossly underestimated or ig-
nored in many synoptic studies. However, rooted
aquatic plants are becoming increasingly significant
in the study of aquatic communities, supplementing
data obtained on algae (Patrick and Martin 1974).
Collection of these plants is normally done by har-
vesting. Species identification, biomass, and regrowth
measurements are done in the laboratory. A surface
plot is harvested and the regrowth rate is deter-
mined as a function of nutrients and productivity. A
qualified aquatic botanist with access to laboratory
and greenhouse facilities should be employed to
accurately determine the macrophytes of a lake.

Bacterial counts may be important in some sur-
veys. They give an indication of the amount and type
of pollution in the water. These surveys are particu-
larly important where water is used for contact

sports, drinking water, or water in which fish live.
There are a number of techniques available for de-
termining bacteria counts (see Am. Public Health
Assoc. 1971). A qualified bacteriologist should be
employed to perform the study and analyze the data.

Figure 1 1 lists the equipment or alternatives
that will be sufficient to perform biological limnol-
ogy studies. Appendix II lists sources of equipment.

Fishery Studies. Fish are considered by many to be
the most important component of a lake's food
chain. This may or may not be true, but under any
circumstance, the entire food chain needs to be
considered. Emphasis is usually on the fish because
of fishermen's demands. There are many ways to
approach a fishery survey and, therefore, the
equipment needed also varies. As previously
mentioned, the state game and fish agency should be
aware and involved in many of these studies. In all
states, a collection permit is required before any
work begins. A creel census can take many hours
because fishermen need to be frequently
interviewed. On the other hand, a netting survey can
require very expensive, sophisticated gear.

There are basically two categories of methods
for actually collecting fish — those that trap and hold
the fish and those that physically incapacitate them.
Those two can further be broken down by the var-
ious habitats in which each individual method is
applicable. No one method is the best for all sizes

Studies

Reel with

Cable or

Rope

Water

Sample

Bottle and

Messenger

Sample
Bottles

Preservative

Laboratory

Facility

Includes

Other
Equipment

Primary
Productivity

•

•

•

•

— NRC licensed
personnel for
C'4 techniques

— BOD light and
dark bottles
with clips

— Anchor with
float

Plankton

•

•

•

•

— Binocular
microscope

— Counting cell

— Suitable
plankton net

Chlorophyll

•

•

•

— Filtration

system
- Fiberglass

filters

— Cooler

— Spectrophoto-
meter

Benthos

•

•

•

— Binocular
microscope

— Wash pans

— Forceps

— No. 30 sieve
wash bucket

— Ponar, or
similar dredge

Macrophyte

— Greenhouse

facilities
— Aquaria/bell

jars

— Tape measure
and plot frame

— Plastic bags

— Digging device

Figure 11. Equipment recommended for biological limnology studies.

250

Lakes

.

and species of fish in all locations. Trawls ( Figure
12) are useful where water depth and bottom com-
position are favorable; trawls are most successfully
used at night. The boat must tow the trammel net
efficiently, so sufficient water depth is needed. The
only fish not sampled with a trawl are the very fast
swimmers and very small ones.

A variety of nets or seines are available for tak-
ing fish. Nets are usually referred to as fixed or pas-
sive sampling devices. Seines are usually indicative of
an action device, one that is moved through the

water and traps fish. Gill nets (Figure 13) are station-
ary and fish swim into them and become captured;
hoop or trap nets are those that fish swim into and
are held in the enclosed device until removed. Hori-
zontal gill nets are most commonly used, but vertical
gill nets are also used to determine depth of water
occupied by fish. The best nets for reconnaissance
surveys are termed "experimental gill nets" since
they include a variety of mesh sizes sewn together.
Gill nets are most commonly fished overnight. After
reconnaissance is done and the desired size of fish to
be studied is determined, the size mesh to be used
is selected.

Figure 12. The Beam Fish Trawl (a) and the Otter Fish Trawl (b).

Figure 13. Gill net (a) and hoop net (b).

Lakes

251

Seines commonly used are the straight or min-
now seine. Two individuals wading (or one wading
and one in a boat) pull it near the shore. These are
good for littoral zone sampling. The minnow (or
haul) seine has a trailing bag. There are several
methods of incapacitating fish. One is the use of
piscicide chemicals. The most commonly used of
these is rotenone. Creol and cyanide are also used.
All of these require, in some states at least, personnel
that are licensed piscicide applicators. Use of these
piscicides can cause serious damage to fish and other
biota. Care in their application should be taken.

Another way of incapacitating fish is the use of
electrofishing techniques. This method works well in
very shallow lakes or portions of lakes. Where
aquatic vegetation precludes efficient seining, elec-
trofishing gear is very applicable. The gear can be
either portable or mounted on a boat. Night is the
best time to electrofish in lakes, ponds, and
reservoirs.

The last method is the use of hook and line.
This method, however, is highly selective to pre-
ferred game species. After fish are collected, species
identification, number of fish captured per unit ef-
fort, length, weight, sex, and age should all be re-
corded. These data can usually be collected without
harm to the fish. Some individuals are collected and
preserved for both verification purposes and also
for further analysis (e.g., food habit studies).

DISCUSSION

Each lake is a world within itself. One lake may
be similar to another lake, but each is still unique.
That is, two lakes may both have rainbow trout,
freeze every winter, be too cold for people to ever
swim in comfortably, contain sparkling clear water,
have a stream flowing in and a stream flowing out,
and even have pine trees up to their shoreline. A
closer look reveals, however, that one lake is twice
as deep as the other and has stronger thermal stratifi-
cation; twice the abundance of plant nutrients; three
times the densities of plankton; a relatively large,
shallow littoral area in its upper end; and almost four
times more fish caught in any year.

From the above facts, a biologist can see that
one lake probably needs a little closer attention and

perhaps a good management plan, whereas the other
lake with richer flora and fauna should be left alone.
They then are different internally. As knowledge is
gained from their similarities, knowledge also is
gained from their differences.

As alternative management decisions regarding a
resource are being considered, it is advisable that as
many concerned resource specialists and managers
as possible be consulted for their opinions. This is, of
course, a difficult task. I know from experience that
as the number of interests (agencies and people)
in natural resources increases, the complexity of any
decision increases and the reality of achieving a wise
decision seems more remote. However, biologists
and resource managers should not shy away from
seeking all opinions possible, no matter how diverse
they may become, so that decisionmakers can make
decisions based on the best information.

The questions are always how much data is
needed and how many times do things need to be
measured? These questions should be asked and
answers sought from resource personnel from all
interested resource agencies. The decision that is fi-
nally made may strongly influence the future of a
particular resource. For example, some decisions
may result in irreversible changes to a system (e.g.,
introduction of a new species of game fish ), whereas
others may be reversed (e.g., change in the annual
flow regime from a reservoir). Having the most and
best information possible adds to the credibility of
these important decisions.

The final important step that is neglected most
of the time is to follow up on the actual influence
and success of the decision. A good rule for the re-
source's sake is to see that follow-up monitoring
investigations are performed. For example, suppose
it had been decided, based on scientific evidence,
that a forage species of fish was desirable in a certain
lake because walleye did not have enough to eat
after they became 15 cm (6 in.) long. It would be
desirable to monitor whether the introduction of
forage fish was successful or not, also to know if
forage fish are reproducing naturally in sufficient
numbers to provide an adequate food base for wall-
eye in the future. Certainly monitoring this situation
would make the next similar decision easier to make.
Post-decision monitoring then makes each future
decision easier to make.

252

Lakes

LITERATURE CITED

AMERICAN PUBLIC HEALTH ASSOCIATION and AMER-
ICAN WATER WORKS ASSOCIATION. 1971. Stand-
ard methods for the examination of water and
wastewater. Water Control Bull. 1971. 13th ed. Am.
Pub. Health Assoc. Washington, DC. 874pp.

HUTCHINSON, G.E. 1957. A treatise on limnology, Vol-
ume I, geography, physics and chemistry. John Wiley
& Sons, Inc. New York, NY. 1015pp.

MAITLAND, PS. 1978. Biology of freshwaters. John Wiley
& Sons, New York, NY. 244pp.

ODUM, E.P 1971. Fundamentals of ecology. W.B. Saunders
Co. Philadelphia, PA. 574pp.

PATRICK, R. and D.M. MARTIN. 1974. Biological surveys
and biological monitoring in freshwaters. Number 5.
Contributions from Dep. Limnology, Acad. Nat. Sci.
Philadelphia, PA. 63pp.

TRANDAHL, A. 1978. Preface, Page x in Kendall, R.L. ed.,
Cool-water Fishes of North America. Special Publica-
tion 11, Am. Fish. Soc. Washington, DC. 437pp.

U.S. GEOLOGICAL SURVEY, Office of Water-Data Coordi-
nation. 1977. National handbook of recommended
methods for water-data acquisition. Two Volumes.
U.S. Dep. Inter. Reston, VA.

WOOD, R.D. 1975. Hydrobotanical methods. Univ. Park
Press. Baltimore, MD. 173pp.

WRIGHT, J.C. 1967. Effects of impoundments on produc-
tivity, water chemistry, and heat budgets of rivers.
Pages 188-199 in Reservoir Fishery Resources Sym-
posium. Am. Fish. Soc. Washington, DC.

Lakes

253

APPENDIX I.

APPENDIX II.

Manuals and Textbooks with Inventory
and Monitoring Methods for Lakes.

Sources of Limnological and Fishery
Equipment.

CAIRNS, J., Jr. and K.L. DICKSON. 1973. Biological meth-
ods for the assessment of water quality. ASTM Special
Publication 528. Am. Soc. for Test & Mat. Philadel-
phia, PA. 256pp.

EDDY, S. and J. UNDERHILL. 1978. The freshwater fishes.
Wm. C Brown Co. Dubuque, IA. 215pp.

EVERHARDT, W.H., AW. EIPPER, and WD. YOUNG.
1975. Principles of fishery science. Cornell Univ.
Press. Ithaca, NY. 288pp.

HUTCHINSON, G.E. 1967. A treatise on limnology, Vol-
ume II. Introduction to lake biology and the limno-
plankton. John Wiley & Sons, Inc. New York, NY.
1115pp.

LAGER, K.F. 1956. Freshwater fishery biology. Wm. C
Brown Co. Dubuque, IA. 421pp.

LIETH, H. and R.H. WHITTAKER, eds. 1975. Primary pro-
ductivity of the biosphere. Ecological Studies 14,
Springer-Verlag, New York, NY. 339pp.

LIND, O.T. 1979. Handbook of common methods in lim-
nology, Second Edition. The C.V. Mosby Co. St. Louis,
MO. 154pp.

WELCH, P.S. 1948. Limnological methods. McGraw Hill,
Inc. New York, NY. 381pp.

This list implies no endorsement of product or
service. A more complete listing is "Sources of Lim-
nological and Oceanographic Apparatus and Sup-
plies" (special publication No. 1, third revision)
available from the American Society of Limnology
and Oceanography, 1530 12th Avenue, Grafton,
Wisconsin 53024.

Test kits

Hach Chemical Co.
P.O. Box 389
Loveland, CO 80539

General Sampling & Field Equipment
Wildlife Supply Company

(Wildco)
2200 S. Hamilton St.
Saginaw, MI 48602

Kalh Scientific

Instrument Co.
P.O. Box 1166
El Cajon, CA 92022

Fish Nets and Fish Sampling Gear
Nylon Net Company
7 Vance Avenue
P.O. Box 592
Memphis, TN 38101

Cofelt Electronic Co.,

Inc.
3910 So. Windermere St.
Englewood, CO 80110

Radio Isotopes

Amersham-Searle Corp.
2636 Clearbrook Dr.
Arlington Heights, IL
60005

New England Nuclear
575 Albany St.
Boston, MA 02118

Multiparameter Instruments Submarine Photometers
Hydrolab
P.O. Box 50116
Austin, TX 78763

Kalh Scientific

Instrument Co.
P.O. Box 1 166
El Cajon CA 92022

254

Lakes

Ill SPECIES GROUPS

13 Fish

14 Amphibians and Reptiles

i

15 Songbirds

16 Raptors

17 Marsh and Shorebirds

18 Waterfowl

19 Colonial Waterbirds

20 Upland Game Birds

21 Rodents and Insectivores

22 Lagomorphs

23 Carnivores

24 Bats

25 Ungulates

'. ■

13
FISH

Paul Cuplin

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

Editor's Note: This chapter is the first of a series in
which the focus is on species groups. In this book,
species are lumped into groups based upon similar-
ity of the techniques used to inventory and monitor
their habitat or populations. Fish are of course a
diverse and widespread group, and fisheries man-
agement is often considered as a separate discipline
from wildlife management. However, inventory
and monitoring of an area of land would be in-
complete without taking into consideration the
aquatic resources, of which fish are extremely im-
portant. One chapter cannot cover, except in the
briefest manner, the massive field of fisheries biol-
ogy. However, it can serve as a starting point for
the biologist with little experience in this area The
references cited provide further guidance.

INTRODUCTION

The purpose of this chapter is to assist biologists
in identifying information needed to inventory and
monitor fish-producing waters including streams,
lakes, ponds, and reservoirs. Fishery biology is a
well-developed profession with academic curricula
and degrees, professional societies, and scientific
publications. A single chapter on fish cannot cover
the entire subject of fish habitat and population tech-
niques in any depth. This chapter emphasizes the
techniques appropriate for inland fisheries investiga-
tions in the western U.S., particularly for salmonids
and for smaller streams and ponds in arid lands. For
more detailed information, consult Nielsen and John-
son (1983) or the papers cited in this chapter. Carl-
son and Gifford (1983) provide guidance on locating
and accessing fisheries literature.

State fish and game agencies have historically
managed fish populations; federal land agencies man-
age habitat on public land. There must of necessity
be some overlapping of responsibility since federal
agencies cannot evaluate the habitats properly with-
out knowing something about existing fish popula-
tions. Likewise, State agencies need to know habitat
variables in order to manage fish populations. Nor
can critical habitat for threatened or endangered fish
species be preserved without knowledge of habitat
variables needed to sustain listed fish species.

AQUATIC HABITAT FEATURES

Important habitat features for fish are described
in this section. Methods for measuring many of the
following attributes or components are described by
Hamilton and Bergersen (1984).

Soils

The ability of the resource specialist to deter-
mine potential of riparian zones through the analysis

Fish

257

of soil and water flows will direct rehabilitation ef-
forts to riparian areas with the highest potential.
Streams in improved riparian habitat will support
larger fish populations.

Crouse and Kindschy (1981) described soil
conditions required by various riparian plant species
along streams in southeastern Oregon. In the west-
ern U.S., highly alkaline soils along perennial streams
will predictably produce alkali bullrush (Scirpus
paludosus), greasewood (Sarcobatus sp.), buffalo-
berry (Shepherdia sp.), and salt cedar (Tamarix sp.);
less alkaline soils will produce sedges (Carex sp.),
forbs, grasses, cattails (Typha sp.), and few woody
species. Soil that is extremely rocky will support
willow (Salix sp. ), mock orange (Philadelphus sp. ),
chokecherry (Prunus sp.), and sparse stands of
grasses and forbs. Fine textured soil will support tree
willow, cottonwood (Populus sp.), alder (Alnus
sp.), aspen, dogwood {Cornus sp.), mock orange and
other shrubs, dense stands of grasses, sedges, and
forbs.

Physical Features

Physical features of streams and lakes such as
stream gradient, may be correlated with the capabil-
ity of the water to support or produce fish. These
are described in more detail in the Lakes and
Streams chapters. Methods for measuring such attri-
butes are described in the Aquatic Physical Features
chapter.

Vegetation

Along streams, the capability of the stream to
produce or support fish populations is often a func-
tion of or correlated with the condition of the ripar-
ian vegetation. Shrubs, trees, and other woody
riparian vegetation are needed to shade smaller
streams and control high summer water tempera-
tures for cold-water fishes such as the Salmonidae,
which include trout, salmon, and grayling. Stream-
banks are strengthened and stabilized by the root
structure of trees and shrubs. Streams with only
grass and forb cover are subject to streambank ero-
sion during high water flows.

Food of salmonids sometimes consists of as
much as 35% terrestrial insects during the summer
months. If woody riparian vegetation is absent, this
source of food is also absent.

Water Quality

The chemical and physical properties of water
are extremely important habitat characteristics for
fish. Three of the most important attributes are pH,
dissolved materials, and dissolved oxygen. Other
properties are described in the chapters on Streams
and Lakes. Methods for measuring water quality are
described in the chapter on that subject.

pH. Most freshwater fishes can accommodate pH
ranging from 6.0 to 95, although some species can
adapt below 6.0 or above 9.5 pH.

Dissolved Materials. Total dissolved materials in
excess of 1,500 mg/L NaCl cannot be tolerated by
most fresh-water fish species. Most waters in the
West fall in a range of less than 400 mg/L except for
the lower Colorado River, saline seeps, and springs.

Dissolved Oxygen. For warm-water fish the
dissolved oxygen (DO) concentration should be 5
mg/L or greater. Cold-water fishes prefer the DO
concentration to be at saturation which varies with
water temperature and elevation. Under extreme
conditions the DO may range to 5 mg/L or lower
during the winter. The critical periods for DO are
during late summer and during long periods of ice
and snow cover during winter months.

Macroinvertebrates

Macroinvertebrates can be used to characterize
stream water quality by the diversity of species,
number of species, and by the presence or absence
of certain species ( see the Macroinvertebrates
chapter).

The simplified stream-aquatic organism relation-
ship is presented in Figure 1. Macroinvertebrates
play a major role in consuming aquatic plants and
detritus and providing fish with most of their food
supply. A wide diversity of macroinvertebrates is
produced in the stream substrate under optimum
conditions.

Food Supply

The flow of nutrients into the stream system
governs the number and species of macroinverte-
brates that are available to fish for food.

Shrubs and trees also provide much of the en-
ergy source for the stream from leaf fall and twigs.
Tree fall across the stream provides resting shelter
and stream habitat diversity. Healthy riparian vegeta-
tion acts as a filter to prevent soil from reaching
the stream during heavy rainstorms or snowmelt
(Odum 1978).

Fish feeding habits vary widely by family, spe-
cies, young of the year, juvenile, and adult. Examples
of planktivores, plankton feeding species; benthic or
bottom feeders; and piscivores or carnivorous fish
are shown in Table 1. Not listed are sockeye or ko-
kanee (non-migratory) salmon that are planktivores
throughout all life stages.

258

Fish

^A Sun (Energy)

Algae, aquatic plants

Riparian vegetation
(leaves, twigs, wood, detritis)

\

Herbivorous aquatic invertebrates
(mayfly, caddis, diptera larvae, snails)

\

Fi!

-h ri

Carnivorous aquatic insects
(dragonflies, caddis, some stoneflies)

an "*

Figure 1. Simplified stream-aquatic organism relationships.

Table 1. Examples of families and species of planktivores, benthic feeding fish, and piscivores (from Fritz et
al. 1980).

Group

Family

Species

Planktivores

Clupeids
Corregonids

Alewife (Alosa pseudoharengus)
Gizzard shad (Dorosoma cepedianum)

Lake whitefish (Coregonus clupeaformis)

Ciscoe (Coregonus sp.)

Round whitefish (Prosopium cylindraceum)

Benthic feeding fish

Catostomids

Cyprinids
Percopsids
Cottids
Ictalurids

Percids

White sucker (Catostomus commersomi)
Longnose sucker (C. catostomus)
Carp (Cyprinus carpio)

Spottail shiner (Notropis hudsonius)

Troutperch (Percopsis omiscomaycus)

Sculpins (Cottidae)

Catfish (Ictalurus sp.)
Bullhead (Ictalurus sp.)

Walleye (Stizostedion vitreum)
Yellow perch (Perca flavescens)

Piscivores

Salmonids

Osmerids
Esocids

Lake trout (Salvelinus namaycush)

Brown trout (Salmo trutta)

Rainbow trout (Salmo gairdneri)

Chinook salmon (Oncorhynchus tshawytscha)

Coho salmon (0. kisutch)

Rainbow smelt (Osmerus mordax)

Northern pike (Esox lucius)

Fish

259

The food supply may be an important habitat
component limiting fish populations or productivity.
This may be important primarily during one season
of the year or during one life stage. Bowen (1983)
describes methods for determining fish diets.

HABITAT OF IMPORTANT FISH FAMILIES

Families of fishes have developed under various
ecological conditions. The habitat requirements of
each family of fish are well known, and habitat stud-
ies continue to add to our knowledge.

The families of fishes inhabiting public land
waters in the western States and Alaska can be
grouped into cold-water, cool-water, and warm-water
fishes.

Cold-Water Families

Salmonidae. Family Salmonidae, for example, is
represented by trout, salmon, whitefish, and grayling.
This family represents important game and
commercial species. Their life requirements are
similar in that they inhabit cold water, and most
spawn in or on stream substrate in the fall or spring.
Anadromous members (migrating from the ocean
up a fresh water stream to spawn ), such as the
Pacific salmon (Oncorhynchns sp.), migrate to the
ocean in the first or second year of life and return to
fresh water to spawn and die after spending 1 to 3
years in saltwater. Unlike the Pacific salmon, the sea-
run rainbow trout (steelhead), sea-run cutthroat
trout (S. clarki), and Dolly Varden trout (Salvelinus
malma) return from the ocean to spawn in fresh
water but do not die after spawning. They can repeat
the ocean — fresh water spawning migration cycle.
Whitefish and grayling (Thymallus arcticus) spend
their entire life as residents of a lake or stream.

Cottidae. The family Cottidae, sculpins or cottids,
lives on the stream bottom of cold-water streams.
Sculpins or cottids inhabit much of the same stream
habitat occupied by trout and provide food for them.

Cool-Water Families

Catostomidae. The family Catostomidae, suckers,
uses some of the same habitat as trout. Suckers
migrate upstream to spawn during the spring months
and compete with trout for spawning space. They
are common in degraded trout streams which have
silted streambeds and higher than normal summer
water temperatures.

Escoidae. The family Escoidae, pike and pickerels,
inhabits streams, lakes, and reservoirs and is a native
east of the Rocky Mountains to Alaska. Introductions
have been made in western waters. The best known
representative is the northern pike (Esox lucius).
Preferred habitats are weedy areas of lakes and
reservoirs. Spawning takes place in the spring. The
eggs adhere to the stream substrate.

Percidae. Family Percidae, yellow perch (Perca
flavescens) and walleye (Stizostedion vitreum),
originally occurred in lakes and reservoirs east of the
Continental Divide. They have been widely
introduced into the western States. They prefer the
cool water in lakes and reservoirs.

Warm-Water Families

Centrarchidae. The family Centrarchidae, sunfishes,
is best known for bluegill (Lepomis macrochirus),
largemouth bass (Micropterus salmoides), and
crappie (Pomoxis sp. ). Most of these species prefer
warm-water lakes and reservoirs. Smallmouth bass
(M. dolomieui), however, inhabit the larger, warmer
streams. Only one member of this family, the

Grayling.

Northern pike (top) and sauger (bottom).

260

Fish

Sacramento perch (Archoplites interruptus), is
native west of the Continental Divide, primarily
found in the Sacramento and San Joaquin River
Basins in California (Simpson and Wallace 1978).

Cyprinidae. The family Cyprinidae, minnows and
carps, is best represented by the common carp
(Cyprinus carpio), squawfish (Ptychocheilns sp. ),
and the goldfish (Carassius auratus). Carp,
introduced from Asia, thrive in warm, shallow
streams and lakes with abundant aquatic vegetation.
The rooting habits of carp contribute to the high
turbidity of water.

Squawfish are found in a variety of habitats from
reservoirs to the largest western rivers. Squawfish
once achieved a very large size — as much as 27 kg
(60 pounds) in the Colorado River. The numbers of
Colorado squawfish have been severely reduced
due to dam construction which altered their re-
quired stream habitat and water temperatures for
spawning.

Cyprinodontidae. The family Cyprinodontidae,
killifishes, inhabits ponds, springs, and streams of the
southwestern deserts. The most well known of this
group is the Devil's Hole Pupfish (Cyprinodon
diabolis) which exists in the most restricted fish
habitat on earth, a spring-fed pond in Nevada with a
surface area not much larger than a bathtub; the
total population is 300 to 350 individuals. The
habitat for this and other similar fishes of the desert
is in jeopardy due to the use of water for crop
irrigation.

LolonraxLo oorM-euju~C*^Ks

~£<n~2-/W LlvuJr-

hUxAM^i^AcA^- iMjuJr-

Ictaluridac. The family Ictaluridae, catfishes,
originally found east of the Continental Divide, can
now be found in lakes, reservoirs, and streams
throughout the West. The family is represented by
black bullhead (Ictalurus melas), brown bullhead (/.
nebulosus), channel catfish (/. punctatus), flathead
catfish (Pylodictis al warts), and blue catfish (/.
furcatus). The channel, flathead, and blue catfish are
most commonly found in the large warmer ponds
and reservoirs and large warm-water streams.

Endangered Fish Species

Twenty-seven of the 44 U.S. fish species listed
under the Endangered Species Act of 1973 are en-
demic to the southwestern U.S. (Johnson and Rinne
1982), an indication of the rapidly disappearing
aquatic habitat of the southwestern deserts.

Dewatering of desert springs through ground-
water pumping and continued drying of desert
aquatic habitats has seriously reduced habitats for
desert fishes. Water impoundments on the Colorado
River have reduced water temperatures and there-
fore the reproduction capabilities of the four native
"large river" species, Colorado Squawfish, razorback
sucker (Xyrauchen texanus), humpback chub {Gila
cypha), and bonytail chub (Gila elegans).

POPULATION MEASUREMENT TECHNIQUES

Fish population studies are conducted to relate
fish biomass to habitat conditions, observe change in
relative fish numbers, determine habitat preference
by species, and determine use periods of the year at
various life stages. A method of sampling should be
selected to provide the required information at a
minimum cost.

Johnson and Nielsen (1983) compared relative
costs to obtain different types of information on
fish populations and communities (Table 2). The
lowest level (easiest to get but least valuable) of in-
formation about fish communities is the number of
species present and susceptible to sampling gear. At
the other end of the scale are radio tracking and
food habit information; both activities are very ex-
pensive, but the resultant information may be exten-
sive. Between these extremes are a series of
activities that will provide more information at in-
creasing cost. Table 2 lists these activities and an
estimate of their relative cost. Each activity is com-
pared to the first activity, and each activity is af-
fected by sampling bias. Since the relative cost in
dollars and personnel of different types of studies
vary considerably, the biologist should be careful not
to collect more information than is required for the
problem being addressed.

Fish

261

Table 2. A cost-benefit hierarchy for information about fish populations and communities (from Johnson and
Nielsen 1983).

Level

Activity

Information

Relative Cost

Comments

1

Species
enumeration

Number of species
present.

1

Useful in sampling.

2

Numbers of fish
caught of each
species

Relative abundance of the
species present.

x2

Usually the minimal level
of information needed.

3

Length of fish

Length distributions can
give relative year-class
strength, growth, and
mortality, especially for
young fish.

x4

A great deal of helpful
information added, partic-
ularly for fast growing,
temperate latitude.

4

Weight of fish

Length-weight curves,
condition factors, relative
weight.

x12

Most field scales are not
accurate enough. Must
construct special shelter
or move indoors. Condi-
tion factors need both
accurate lengths and
weights.

5

Age determination

More accurate than length
distributions for calculating
year-class strength, age
distribution, growth history,
and mortality.

x120

Must have accurate length
measurements; requires
extra handling of fish and
laboratory time for analy-
sis.

6

Radio or sonar
tagging

Exact information about
fish location and mea-
surements. May be com-
bined with information
about water depth, tem-
perature.

x1,200

Much information gained
about movements of rela-
tively few fish. Equipment
cost and maintenance can
be quite high.

Inventorying a small stream by electrofishing.

262

Fish

General Considerations

Life Stages of Fishes. The terms applied to the
developmental life stages of fish assist in describing
and understanding fishery reports and studies.

Developmental stage terms most commonly
used are as follows:

Egg — Fertilized ovum with developing embryo.

Eyed Egg — Embryo developed to the state

where the eyes can be readily ob-
served.

Larva — Prolarva still bear yolk sacs, sometimes
called sac fry. With postlarva, yolk sacs
have been completely absorbed.

Alevin — The newly hatched salmon, with yolk
sac still attached.

Fry — The same as postlarva; yolk sac has been
completely absorbed.

Fingerling — A fish in the first or second year of
life, 5 to 10 cm (2 to 4 in.) long.

Young of the Year — Young fish added to the fish
population during the cur-
rent year.

Smolt — One- or 2-year-old anadromous fish

physiologically prepared for migration
into saltwater.

Grilse — Early maturing salmon, mostly male fish.

Adult — Mature fish capable of spawning.

Kelt — Spent salmon or sea-run trout after spawn-
ing, in a weak, emaciated condition.

In the past, most fish population work has con-
centrated on mature stages of fish (from fry through
kelt). Awareness is growing of the need to collect,
identify, and gather data on larval fish. In many cases,
habitat for larval fish is quite different from that of
mature stages; it may be geographically separated
also. If larval habitat is deteriorating, then it should
be identified and monitored.

Laboratories are available for assisting in larval
fish sampling and identification. Since 1977 a larval
fish conference has been held annually in the U.S. or
Canada. Information on larval fish procedures are
described by Snyder (1983); more information on
the conferences and on procedures can be obtained
from the Larval Fish Laboratory, Colorado State Uni-
versity, Fort Collins, CO 80523.

Fish Collection. Fish specimens sometimes need to
be collected and stored for later identification or
tagged or marked for population surveys. Biologists
must comply with State and federal laws including
collection permit requirements when sampling or
collecting fish. Procedures for tagging and marking
fish are described by Wydoski and Emery (1983).

Fish Preservation. Specimens should be preserved
in 4% formaldehyde solution (10% formalin) con-
taining about 3 g of borax per liter (Greeson et al.
1977). Fish more than 3 inches long should be slit on
the right side to allow preservatives to penetrate
the body cavity.

Identify samples with date, species, location
where collection was made, and name of collector. If
individual fish from separate locations are to be
stored in the same container, they can be wrapped
in cheesecloth with an identification tag attached.

Capture Methods. Fisheries surveys beyond Level 1
(presence; Table 2) normally require fish capture.
Fish capture can also be used to determine presence.
Many methods for capturing fish are available.
Seining, electrofishing, fish toxicants, and hook-and-
line sampling are commonly used on small inland
waters.

Seining Small seines are useful in small ponds
but are less effective in small streams. Details about
seining techniques are described by Hayes
(1983.139-143).

Electrofishing. Electrofishing can be used to
determine fish presence, species composition, and to
make fish population estimates. Electrofishing is the
most effective method of fish population sampling
for a stream. The most effective time for electrofish-
ing is from midsummer to the fall months when
the streamflow is at a low point.

A 600-m (1/10-mi. ) sampling area is blocked
with block nets at the upper and lower ends of the
sampling area. Population estimates for important
game, sensitive, and endangered or threatened fishes
are made by the Seber-LeCren (1967) method. A
fish population estimate, relative abundance estimate,
fish length distribution, and pounds of fish can be
calculated for each fish population sample. Reynolds
( 1983) describes in detail procedures for
electrofishing.

Other. Many other techniques for capturing
fish are available including use of toxicants and
hook-and-line. Hubert (1983) and Hayes (1983)
describe procedures for passive and active capture
techniques, and Davies (1983) describes use of toxi-
cants.

Fish

263

Specific Techniques

Appropriate techniques that correspond with
each of the 6 levels of survey (Table 2) are de-
scribed below.

Presence (Level 1).

Counting Weirs and Spawning Ground
Surveys, Aerial Redd Counts. Observations of
salmon and other anadromous fish species in
spawning runs have often been made at counting
wiers and by the enumeration of dead fish on the
spawning grounds. The latter method is used to
determine numbers and species on specific streams.
Redd counts can be made during the spawning
season with helicopter or fixed-wing aircraft
(Neilson and Green 1981).

Visual Observation of Stream Fish. Stream
surveyors can recognize species by observing, deter-
mining presence, and estimating relative numbers
per measured distance of a stream.

Relative Abundance — Population Estimation
(Level 2). Relative abundance can be expressed as
the numbers of fish, by species, per unit of capture
effort, e.g., number of brook trout (S. fontinalis) per
miles of stream shocked.

For important species, population estimates can
be made using one of the techniques described be-
low. Express population estimates in numbers per
linear mile for streams or numbers per surface acre
for lakes, ponds, and reservoirs. For details of statisti-
cal analyses or information on alternative methods,
see Lagler (1956), Ricker (1958), Armour and Platts
(1983), Otis et al. (1978), or White et al. (1982).

Two-Catch Depletion (Seber-LeCren) Esti-
mator. Depletion methods require that the fish from
samples not be re-placed (put back into the water).
In the two-catch depletion method, a first sample
is taken (without re-placement) followed by a sec-
ond sample. Equipment and effort must be standard-
ized so that the expected proportion of the
population caught is the same for each sample.

The total population can then be estimated with
the formula (Seber and LeCren 1967):

C, - C2

Otis et al. ( 1978) provide a formula for calculat-
ing sampling variance. However, White et al. (1982)
recommended using three or more captures to test
the assumption of equal capture probability and to
increase precision. They describe assumptions and
statistical analyses for three or more samples.

Mark and Release. Mark and release tech-
niques have most often been used in lakes, ponds,
and reservoirs to study fish population numbers. Fish
populations can be estimated by capturing and re-
leasing a number of marked fish and then recaptur-
ing marked and unmarked fish.

The population can then be estimated by the
formula (Davis and Winstead 1980).

N =

Mn

m

Where N = estimated number of fish in population

Ci = number of fish caught in first sample

C2 = number of fish caught in second
sample

Where N = estimated number of fish in population

M = number of fish marked

n = total number of fish captured in
second sample

m = number of marked fish in second sample

Assumptions applied to this type of population
study are ( 1 ) fish do not lose their marks, (2)
marked fish are randomly distributed in the popula-
tion, ( 3 ) marked and unmarked fish are equally cap-
turable, and (4) numbers of unmarked fish are not
added to the population through recruitment or
immigration (Lagler 1956). Many variations on the
mark-recapture technique have been developed. For
details of these see Lagler (1956), Ricker (1958),
Otis et al. (1978), or White et al. (1982).

Length of Fish (Level 3). Rate of change in length
of individuals and length-frequency distributions
are key attributes of fish populations (Anderson and
Gutreuter 1983). Fish length distributions can be
developed from samples. For details of measurement
and analysis techniques, see Anderson and Gutreuter
(1983).

Weight of Fish (Level 4). Weight of individual fish
and fish populations are also key attributes of fish
populations (Anderson and Gutreuter 1983). The
relationship between weight and length can also be
used to develop condition indexes. For details of
these techniques see Anderson and Gutreuter
(1983).

Age Determination (Level 5). Age determination
requires a substantial increase in time and money
(Table 2) and should only be done where the
problem(s) being addressed require such
information. See Jearld (1983) for details on
techniques and analysis.

264

Fish

Radio or Sonar Tagging (Level 6). Radio or sonar
tagging is extremely costly (Table 2) and the
biologist should be sure that such an investment is
justified before beginning a study. Thome (1983)
describes the use of sonar or hydroacoustics. Winter
et ai. (1978) and Winter (1983) describe
underwater biotelemetry including sonar and
radiotelemetry.

HABITAT INVENTORY AND MONITORING
SYSTEMS

Inventory and monitoring fish habitat allow the
land manager to identify trends in habitat conditions
needed for fish production and to adjust land use
to stabilize or improve habitat conditions. No fish in-
ventory will provide all of the data needed to de-
scribe habitat conditions for specific fish species.

A number of systems have been developed to
inventory fish habitat conditions and to relate habitat
conditions to actual or potential fish populations.
Three of these are described below.

Habitat Quality Index (HQI)

Binns (1979) has developed a habitat quality
index (HQI) for predicting biomass of trout in
streams based on nine measurable habitat attributes.
The index is described in more detail in the Streams
chapter.

Habitat Evaluation Procedures (HEP) and
Habitat Suitability Index (HSI) Models

The U.S. Fish and Wildlife Service has developed
Habitat Evaluation Procedures (HEP) which provide
numerical ratings of habitat quality for pre- and post-
project conditions (U.S. Department of Interior, Fish
and Wildlife Service 1976). HEP uses one or more

Habitat Suitability Index (HSI) models for such anal-
yses. These models use one or more habitat variables
to generate an index of habitat quality between 0
and 1 for individual fish species. Aquatic applications
of HEP and HSI models are described in detail by
Terrell et al. (1982). HSI models have been devel-
oped and published for many fish species. The biolo-
gist may find these useful in developing an inventory
and monitoring system or program other than HEP.
Models and their use in HEP are described in more
detail in the chapter on Habitat Evaluation Systems.
More information on HEP and HSI models and on
availability of HSI models for fish can be obtained
from "the National Ecology Center,'' U.S. Fish and
Wildlife Service, Drake Creekside Building One,
2627 Redwing Road, Fort Collins, CO 80526-2899.

Riparian/Aquatic Information Data
Summary (RAIDS)

Riparian/ Aquatic Information Data Summary
(RAIDS) is a computer-based system for storing data
on aquatic habitat and the associated fish popula-
tions. It does not specify how such data must be col-
lected. Details on the system can be obtained from
the BLM's Denver Service Center, Division of
Resources.

SUMMARY

Inventory of fish and fish habitat is needed to
provide baseline information for the land manager.
After baseline conditions have been assessed, moni-
toring helps evaluate land management actions and
provides justification for leads toward adjusting land
use where necessary. This chapter provides guidance
for planning and organizing inventories and monitor-
ing studies of fish-producing waters. For detailed
descriptions of techniques, equipment, and analyses,
the biologist will need to consult the Literature
Cited section in this chapter.

Fish

265

LITERATURE CITED

ANDERSON, R.O. and S.J. GUTREUTER 1983. Length,

weight, and associated structural indices. Pages 283—
300 in Nielsen, L.A. and D.L. Johnson, eds., Fisheries
Techniques. Am. Fish. Soc. Bethesda, MD. 468pp.

ARMOUR, C.L. and W.S. PIATTS 1983. Field methods and
statistical analyses for monitoring small salmonid
streams. U.S. Dep. Inter., Fish and Wildl. Serv., FWS/
OBS-83/33. 200pp.

BINNS, N.A. 1979- A habitat quality index for Wyoming
trout streams. Fishery Research Report Monograph
Series 2. Wyoming Game and Fish Dep., Cheyenne,
82002. 75pp.

BOWEN, S.H. 1983. Quantitative description of the diet.
Pages 325—336 in Nielsen, LA. and D.L. Johnson, eds.,
Fisheries Techniques. Am. Fish. Soc. Bethesda, MD.
468pp.

CARLSON, CA. and C.L. GIFFORD 1983. Finding literature
and reports. Pages 23—42 in Nielsen, LA. and D.L.
Johnson, eds., Fisheries Techniques. Am. Fish. Soc. Be-
thesda, MD. 468pp.

CROUSE, MR and RR. KINDSCHY. 1981. A method for
predicting riparian vegetation potential of semiarid
rangelands. Pages 1 10-1 14 in Proc. of Symp., Acquisi-
tion and Utilization of Aquatic Habitat Inventory
Information, Western Div. Am. Fish. Soc. Portland,
OR. Oct. 1981.

DAVIES, WD. 1983- Sampling with toxicants. Pages 199-
214 in Nielsen, L.A. and D.L. Johnson, eds., Fisheries
Techniques. Am. Fish. Soc. Bethesda, MD. 468pp.

DAVIS, D.E. and R.L. WINSTEAD. 1980. Estimating the
numbers of wildlife populations. Pages 221-246 in
Schemnitz, S.D., ed., Wildlife Management Techniques
Manual. 4th ed. revised, The Wildl. Soc, Washington,
DC. 686pp.

EISERMAN, F.M. 1979. Quantification of fluvial trout habi-
tat in Wyoming. Trans. Am. Fish. Soc. 108:215-228.

FRITZ, E.S., P.J. RAGO, and LP. MURARKA. 1980. Strategy
for assessing impacts of power plants on fish and
shellfish populations. U.S. Dep. Inter., Fish and Wildl.
Serv., Biological Services Program, National Power
Plant Team, FWS/OBS-08/34. 68pp.

GREESON, P.E., T.A. EHLKE, GA. IRWIN, B.W. LIUM, and
K.V. SLACK. 1977. Techniques of water resources
investigations of the U.S. Geological Survey, Chapter
A4. 332pp.

HAMILTON, K. and E.P. BERGERSEN. 1984. Methods to
estimate aquatic habitat variables. Colorado Coop.
Fishery Res. Unit, Colorado State University, Ft. Col-
lins.

HAYES, ML. 1983. Active fish capture methods. Pages

123-146 in Nielsen, LA. and D.L. Johnson, eds., Fish-
eries Techniques. Am. Fish. Soc. Bethesda, MD.
468pp.

HUBERT, W.A. 1983- Passive capture techniques. Pages
95—122 in Nielsen, LA. and D.L. Johnson, eds., Fisher-
ies Techniques. Am. Fish. Soc. Bethesda, MD. 468pp.

JEARLD, A. Jr. 1983- Age determination. Pages 301-324 in
Nielsen, LA. and D.L. Johnson, eds., Fisheries Tech-
niques. Am. Fish. Soc. Bethesda, MD.

JOHNSON, D.L. and LA. NIELSEN. 1983. Sampling consid-
erations. Pages 1-21 in Nielsen, LA. and D.L Johnson,

eds., Fisheries Techniques. Am. Fish. Soc. Bethesda,
MD. 468pp.

JOHNSON, J.E. and J.N. RINNE. 1982. The Endangered

Species Act and Southwest fishes. Pages 2-7 in Fisher-
ies Bulletin, 7(4), Am. Fish. Soc.

LAGLER, K.R. 1956. Freshwater fishery biology, 2nd ed.
Wm. C. Brown Co., Dubuque, LA. 421pp.

NEILSON, J.D. and GH. GREEN. 1981. Enumeration of
spawning salmon from spawner residence time and
aerial counts. Trans. Am. Fish. Soc. 110(4):554-556.

NIELSEN, L.A. and D.L. JOHNSON, eds. 1983. Fisheries
techniques. Am. Fish. Soc. Bethesda, MD. 468pp.

ODUM, E.P. 1 978. Ecological importance of the riparian
zone, strategies for protection and management of
floodplain wetlands and other riparian ecosystems, in
Proc. of Symp., U.S. Dep. Agric, For. Serv., Washing-
ton, DC.

ORTH, D.J. 1983- Aquatic Habitat Measurements. Pages
61—84 in Nielsen, LA. and D.L. Johnson, eds., Fisher-
ies Techniques. Am. Fish. Soc. Bethesda, MD. 468pp.

OTIS, D.L., K.P. BURNHAM, G.C. WHITE, and DR. ANDER-
SON. 1978. Statistical inference from capture data
on closed animal populations. Wildl. Monogr. 62.
135pp.

REYNOLDS, J.B. 1983. Electronshing. Pages 147-164 in
Nielsen, LA. and D.L. Johnson, eds., Fisheries Tech-
niques. Am. Fish. Soc. Bethesda, MD. 468pp.

RICKER, W.E. 1958. Handbook of computations for bio-
logical statistics of fish populations. Fish. Res. Board,
Can. Bull. 119. 300pp.

SEBER, G.A.F. and ED. LECREN. 1967. Estimating popula-
tion parameters from large catches relative to a popu-
lation. J. Animal Ecol. 36(3):631-643.

SIMPSON, J.C. and R.L. WALLACE. 1978. Fishes of Idaho.
The University Press of Idaho. Idaho Research Foun-
dation, Inc. Moscow. 237pp.

SNYDER, D.E. 1983. Fish eggs and larvae. Pages 165-198
in Nielsen, L.A. and D.L. Johnson, eds., Fisheries Tech-
niques. Am. Fish. Soc. Bethesda, MD. 468pp.

TERRELL, J.W., T.E. MCMAHON, P.D. INSKIP, R.F. RA-
LEIGH, and K.L. WILLIAMSON. 1982. Habitat suitabil-
ity index models: Appendix A. Guidelines for riverine
and lacustrine applications of fish HSI models with
the habitat evaluation procedures. U.S. Dep. Inter.,
Fish and Wildl. Serv., FWS/OBS-82/10.A. 54pp.

THORNE, RE. 1983. Hydroacoustics. Pages 239-259 in
Nielsen, LA. and D.L. Johnson, eds.. Fisheries Tech-
niques. Am. Fish. Soc, Bethesda, MD. 468pp.

U.S. DEPARTMENT OF INTERIOR, FISH AND WILDLIFE
SERVICE. 1976. Habitat evaluation procedures. Divi-
sion of Ecological Services. Washington, DC. 30pp.

WHITE, G.C, DR. ANDERSON, K.P. BURNHAM, and D.L.
OTIS. 1982. Capture-recapture and removal methods
for sampling closed populations. Los Alamos National
Laboratory, Los Alamos, NM. LA-8787 NERP. 235pp.

WINTER, J.D. 1983- Underwater biotelemetry. Pages 371-
396 in Nielsen, LA. and D.L. Johnson, eds., Fisheries
Techniques. Am. Fish. Soc. Bethesda, MD. 468pp.

, V.B. BUECHLE, D.B. SINIFF, and JR. TESTER. 1978.

Equipment and methods for radio tracking freshwater
fish. Univ. Minnesota, Agric. Experiment Sta. Misc.
Rep. 152-1978. 18pp.

WYDOSKI, R. and L EMERY. 1983. Tagging and marking.
Pages 215-238 in Nielsen, LA. and D.L. Johnson, eds.,
Fisheries Techniques. Am. Fish. Soc, Bethesda, MD.
468pp.

266

Fish

,

14

AMPHIBIANS
AND REPTILES

K. Bruce Jones

U.S. Bureau of Land Management
Phoenix Training Center
Phoenix, AZ 85015

Editor's Note: Among the vertebrates, amphibians
and reptiles have often been ignored. "Traditional"
biologists trained in game management may not
be familiar with studies or techniques for these
species. Very little has been published, and what's
available tends to be located in hard-to-get publi-
cations— not the journals or manuals familiar
to most biologists.

Nonetheless, amphibians and reptiles are important
components of many ecosystems. This chapter
brings together the diffuse techniques for studying
these species and their habitats.

INTRODUCTION

Until recently, amphibians and reptiles have
been largely overlooked in land management consid-
erations. Federal and State agencies have spent most
of their time and money on commercially important
species, and the public is generally unaware of how
important these animals are in natural ecosystems.
Only within the past 10 years have we become con-
cerned about conditions of these animals and their
habitats.

Public awareness of the importance of the entire
wildlife community has led to legislation (e.g., the
Federal Land Policy and Management Act of 1976)
requiring study and management of nongame wildlife
such as amphibians and reptiles. Studies have demon-
strated the importance of these animals in natural
ecosystems. Terrestrial amphibians, lizards, and bur-
rowing snakes are excellent indicators of the relative
amounts of microhabitats in ecosystems. For exam-
ple, the Sacramento Mountain salamander (Aneides
hardii) is entirely dependent on moist, downed logs
and litter on north- and northeast-facing slopes
(Wiegman et al. 1980). In addition, information on
amphibian and reptile abundance and diversity helps
determine the relative health of ecosystems. For
example, lizard abundance and diversity fluctuate
directly with changes in the composition and
amount of microhabitats. These microhabitat changes
often result from land management practices (Jones
1981a; Ortega et al. 1982; Tinkle 1982; Luckenbach
and Bury 1983).

Aquatic amphibians and snakes are good indica-
tors of the health of aquatic systems. These animals
are especially sensitive to pollution and loss of
aquatic habitat (Hall 1980).

Amphibians and reptiles are also important in
food chains, and they make up large proportions of
vertebrates in certain ecosystems (Bury and Raphael

'Current address: Office of Endangered Species, U.S. Fish and
Wildlife Service, Washington, DC 20240.

Amphibians and Reptiles

267

1983). Many carnivorous mammals and raptorial
birds rely on these animals for food. For example, up
to 50% of the spring diet of common black hawks
(Buteogallus anthracinus) consists of aquatic herpe-
tofauna (Millsap and Harrison 1981). Recent studies
have also demonstrated the economic importance,
both consumptive and nonconsumptive, of these
animals (see Bury et al. 1980).

Because of recent concern for nongame wildlife,
biologists and land managers find themselves faced
with studies and management needs for a group
of animals they know little about. Only a few biolo-
gists and virtually no land managers have had any
formal education in herpetology.

There is a wealth of information on the ecology
of individual amphibians and reptiles in literature.
Unfortunately, few papers deal with amphibian and
reptile communities, and even fewer summarize
information needs, data collection, and management
on public lands.

This chapter is designed to help biologists who
have little or no education or experience with am-
phibians and reptiles organize and run inventories or
monitoring studies on selected species or major
groups (e.g., lizards).

HABITAT FEATURES

The most important factor affecting amphibian
and reptile distribution and habitat use is horizontal
and vertical habitat availability. In studying popula-
tion condition, trend, and cause and effect factors,
biologists must collect data on habitat.

Sixteen habitat components and attributes are
important determinants of amphibian and reptile
abundance (Table 1).

Microhabitat components are site-specific, physi-
cal entities that provide environmental conditions
necessary for a wide variety of ecological functions
such as reproduction, foraging, predator avoidance
or escape, thermoregulation, and resting. Amphibians
and reptiles are ectothermic (cold-blooded); body
temperatures are not derived from metabolic pro-
cesses but rather from the surrounding environment.
Therefore, behavioral adaptations and use of different
microhabitats by amphibians and reptiles are diverse.
These animals often demonstrate a high dependence
on certain microhabitats to thermoregulate. Removal
or reduction of microhabitats necessary for thermo-
regulation can detrimentally affect all other ecologi-
cal functions because internal temperature regulation
determines the intensity of activity (see Brattstrom
1965 for examples). For example, certain basking
turtles can be extirpated from a pond if floating logs
are eliminated.

Amphibians are even further restricted to moist
or wet habitats with varying degrees of free-standing
water. Many larval toads, frogs, and salamanders re-
quire water for development into adult stages (Steb-
bins 1966). Additionally, some adult frogs and
salamanders require water to avoid desiccation and
to respirate cutaneously. For example, many Pletho-
don (lungless) salamanders require moist, rotting
logs and litter for egg development and adult cuta-
neous respiration (Stebbins 1966). Without rotting
logs and associated litter, there is insufficient mois-
ture for egg development and adult survival.

Although most biologists and some land man-
agers understand the significance of vegetation struc-
ture and composition to wildlife, few realize how
important litter (fallen logs, leaves), soil, plant root
structure, and horizontal vegetation structure are to
most forms of wildlife, especially amphibians and
reptiles. Litter provides moisture regimes necessary
for development of terrestrial salamanders, but it
also provides cool surface and below-the-surface
temperature regimes. Often this is the only microhab-
itat with moderate environmental temperatures in
otherwise hot, dry regions. Because of low preferred
body temperature ranges, many species such as the
Arizona alligator lizard (Elgaria kingii) and Gilbert
skink (Eutneces gilberti) are entirely restricted to
deciduous riparian habitats within desert regions
(Stebbins 1966). In addition, litter provides feeding
substrate for several common lizards, especially
members of the genus Sceloporus (fence or spiny
lizards). Loss of fallen logs and trees can drastically
reduce populations of these lizards (see discussions
of Jones 1981a & b and Vitt et al. 1974, for depen-
dence on woody vegetation ).

Soil depth, texture, and diversity are extremely
important in determining amphibian and reptile dis-
tribution in an area. For example, soil depth and

Fallen logs provide important microhabitats for terrestrial
salamanders.

268

Amphibians and Reptiles

Table 1. Important components and attributes of amphibian and reptile habitat.

Habitat Component

Types

Variables/
Factors

Amphibian and
Reptile Associations

Relations of Species'
Ecology to
Components

CD
CO

c

3
"S3
Q

0
>

o

CJ

a>

CL
CO

o
en

LU

0)
CO

to

n

-D
CO

CD

c
■o

CD
CD
LL.

CD
Q.

O

r>
o
o

LL

s

t/5

n

CO

to
CD

Z

CO

o

CD

o
o

CO

>.

-C

0_

c
o

o

13

o

Q.
CD
DC

C

g
J5

Z3
CO
CD

6

E

CD

.C
1—

Microhabitat Component

Rivers and
streams

Riffle/run/pool
ratios, water
temperature, tur-
bidity, DO, or-
ganic content,
siltation, pollu-
tants

AQAL, SAAL

•

•

•

•

•

•

Water
(Lotic)

Water
(Permanent Lentic)

Ponds, marshes,
lakes, reservoirs,
natural catch-
ments, agricul-
tural run-off

Water tempera-
ture, DO, organic
content, siltation,
pollutants, emer-
gent vegetation,
logs and other
litter, substrate

AQAL, SAAL

•

•

•

•

•

•

Water
(Temporary Lentic)

Temporary rain
pools, irrigation
ditches

DO, water tem-
peratures, silta-
tion, pollutants,
duration, fre-
quency, emer-
gent vegetation,
substrate

SATD, SASA
SATU, SASN

•

•

•

•

•

•

Rock

Talus slopes,
cliffs, boulders,
substrate

Rock size, heter-
ogeneity, inter-
faces, origin,
vertical and
horizontal struc-
ture

AQAL, SAAL (Where
interfaced with aquatic
habitats), TAL

•

•

•

•

•

•

Vegetation —
Litter/Debris

Leaves, logs,
limbs

Litter size, depth,
heterogeneity,
horizontal struc-
ture, type, mois-
ture retention,
temperature

AQAL, SAAL (Where
interfaced with aquatic
habitats), TAL

•

•

•

•

•

Vegetation —
Live

All vegetation
including roots

Horizontal and
vertical structure,
interfaces with
abiotic compo-
nents, hetero-
geneity

AQAL, SAAL (Where
interfaced with aquatic
habitats) TAL

•

•

•

•

•

•

•

Vegetation —
Dead

Standing vegeta-
tion, roots

Size, interface
with other habi-
tats, heterogene-
ity, vertical and
horizontal struc-
ture, soils (roots
only)

AQAL, SAAL (Where
interfaced with aquatic
habitats) TAL

•

•

•

•

•

•

Vegetation —
Plant Species

Individual plant
species

Individual plant
species abun-
dance and oc-
currence

TL, TSN

•

•

•

Amphibians and Reptiles

269

Table 1. Important components and attributes of amphibian and reptile habitat (concluded).

Habitat Component

Types

Variables/
Factors

Amphibian and
Reptile Associations

Relations of Species'
Ecology to
Components

CD

CO

C

£

CD

Q

CD
>

O

o

CD
Q.
CO
O
CO
LU

CD

CO
JD

3
CO

a>

c

CD
CD

LL

>>

CD

Q.

O

"O

o
o

LL

CD
00

n

3
CO

to
CD

CO
CJ
O)

o
o

CO

>.

-C
Q.

c
o

o

3
"D
O

Q.
CD
DC

g
to

3
O)
CD

O

E

CD
-C

Microhabitat Component
(cont.)

Surface and

subsurface soil
types

Types, death,
heterogeneity,
horizontal and
vertical structure,
interfaces with
other habitat
components

AQAL, SAAL (Where
interfaced with aquatic
habitats), TAL

•

•

•

•

•

•

Soil

Macrohabitat
Components/Factors

% angle of area
from horizontal

% slope, mois-
ture availability,
thermal regimes,
vegetation struc-
ture

SAAL, TAL

•

•

•

•

•

Slope

Aspect

South, north, east,
and west facing

Direction, mois-
ture availability,
temperature
regimes, vegeta-
tion structure

SAAL, TAL

•

•

•

•

•

Elevation

Vertical, above or
below sea level

Vertical distance,
moisture avail-
ability, thermal
regimes, vegeta-
tion structure

AQAL, SAAL, TAL

•

•

•

•

•

•

Precipitation

All forms

Quantity, type,
duration, fre-
quency, moisture
availability, ther-
mal regimes,
vegetation struc-
ture

AQAL, SAAL TAL

Ecotones/Habitat
Juxtaposition

Habitat interfaces
and locations

Heterogenity,
interface size
and cumulative
numbers, posi-
tion of habitats

AQAL, SAAL TAL

•

•

•

•

•

•

•

Topography

General

Slope, aspect,
and physical
feature mixture
and position

SAAL, TAL

•

•

•

•

•

•

•

Geographic Location

Major geographic
boundaries and
barriers

Size, location
and frequency,
habitat size and
disjunction

AQAL, SAAL, TAL

•

Codes for Amphibian and Reptile Associations:
AQ — aquatic
SA — semiaquatic

T — terrestrial
TU— turtles
FR — frogs

TD— toads
SA — salamanders
LI — lizards
SN — snakes
AL — all amphibians and reptiles

270

Amphibians and Reptiles

texture determine the rate of percolation and soil
layer moisture.

In some areas with heavy clay soils, water accu-
mulates on the surface, especially during rainy sea-
sons. Because surface water may be available for up
to 3 weeks, many semiaquatic amphibians occur in
these regions. For example, six species of semi-
aquatic toads were verified within a small, hot region
of the Sonoran Desert that had deep clay soils (Jones
et al. 1983).

Although moist, clay soils enhance the distribu-
tion of toads in some regions, hard, shallow, or rocky
soils may prevent species colonization of an area.
For example, spadefoot toads {Scaphiopus sp.) are
well adapted to survive in dry, desert regions, but
without soils loose enough for burrowing, generally
will not occur in an area (unless rodent burrows are
present ).

The diversity of soils interfaced with rocks and
other structures often determines amphibian and
reptile species richness. Diversity provides more
niches for amphibian and reptile colonization and
existence.

Plant root structure also contributes greatly to
amphibian and reptile species richness by providing
avenues into subsurface space. These plant-created
avenues often account for greater species richness
than expected in what are thought to be structure-
less plant communities. For example, deserts domi-
nated by creosotebush (Larrea sp. ) appear very
homogeneous with little horizontal or vertical struc-
ture. However, the root systems of creosote are ex-
tensive, and they provide a wide variety of
subsurface opportunities to many of the desert's
nocturnal amphibians and reptiles. The result is a
diverse amphibian and reptile community.

Generally, horizontal vegetation structure deter-
mines amphibian and reptile composition more often
than vertical structure, especially in lizards. This
results primarily from foraging and thermoregulation
which are generally conducted on the surface (there
are exceptions — some amphibians and reptiles forage
in trees and below the surface). For example, zebra-
tailed lizards (Callisaurus draconoides) prefer
brushy, open habitats where they ambush prey while
sitting in the open and then move into shade to ther-
moregulate (Pianka 1966; Jones 1981a & b). Some
species such as desert iguanas (Dipsosaurus dor-
salts) are common in very open, sparsely vegetated
creosotebush habitats; this habitat allows them to
reach high preferred body temperatures and obtain
specific food items (creosotebush buds). Other spe-
cies that forage in or on litter and have moderate
internal temperature ranges prefer densely vegetated
habitats with large amounts of surface litter (e.g.,
alligator lizard).

Macrohabitat components are physical phenom-
ena that affect environmental conditions of microhab-
itats and in some instances entire regions. Different
types of microhabitat and macrohabitat components
are listed with corresponding variables or factors
that affect the ecological requirements of species
(Table 1 ). Amphibians and reptiles are listed by ma-
jor taxonomic group (e.g., frogs, lizards). Table 1
serves primarily as a checklist of habitat components
and variables that biologists should consider when
conducting habitat and amphibian and reptile sur-
veys or studies. Biologists should collect data on
soils, litter, water availability, and horizontal vegeta-
tion on each amphibian and reptile sample site. This
information will allow the biologist to accurately
assess the relation between microhabitat availability
and presence or absence of certain species. By being
aware of macrohabitat conditions, biologists will
better understand causes of microhabitat variables,
and how and where to set up samples (e.g., biolo-
gists can reduce sample variability due to slope by
sampling only north-facing slopes, or they may elect
to sample all different aspects to determine their
effect on species composition).

Zebra-tailed lizard.

POPULATION MEASUREMENT TECHNIQUES

Sources of Information

Although amphibian and reptile data collection
generally does not require the degree of expertise
required for surveys of animals such as birds, a biolo-
gist with little or no previous experience will need
to review certain literature and, when possible, con-
tact local or regional experts. Field guides and keys
that provide characteristics for identifying various
herpetofauna (Table 2) are available for most regions
of the U.S. The biologist should thoroughly review
species characteristics before starting field surveys.

Amphibians and Reptiles

271

Review of field guides, books, and amphibian and
reptile journals will also provide ecological back-
ground on various species that will help biologists
select the most appropriate sampling methods.

Museum and individual collection records and
publications can be used to verify amphibian and
reptile species occurrence within a geographic area.
Searches of these sources should be made before
starting censuses to avoid duplication.

Several large museums in the U.S. hold vast rec-
ords of amphibians and reptiles. Edwards ( 1975) lists
U.S. collections of amphibians and reptiles, their
locations, and major sources of records (e.g., Desert
Southwest). This publication serves as an excellent
starting point for record searches.

Expert herpetologists often possess personal
locality records that can add to species lists in spe-

cific geographic regions. Herpetologists with regional
expertise should be consulted before field work is
initiated.

Many journals contain valuable amphibian and
reptile occurrence and ecological data (Table 2).

Problems Affecting Sampling

Biologists should be aware of several factors that
may affect their results when sampling amphibians
and reptiles. Perhaps the largest problem in assessing
amphibian and reptile populations is that these ani-
mals' activities and reproduction vary with natural
environmental fluctuations, such as precipitation and
temperature (Whitford and Creusere 1977; Gibbons
and Semlitsch 1981; Vogt and Hine 1982). There-
fore, biologists should be cautious in interpreting
cause and effect data because observed differences
may result from natural fluctuations in weather

Table 2. Examples of major references providing preliminary and supplemental data for amphibian and
reptile surveys and studies.

Journal or Publication

Types of Information

Geographic Applicability

05

o
>

CO
-C
CD
CO

CD

cz
o

JD

CO

b

CO

O

en
o
o
o

LU

00
CD

O CO
CD Q.

S o
■£ ' «

c
o

00

o
c

CD
"O

C
CD

E

CD
CD
CO

c
CO

CO
O
D)
O
O
00

x:
CL

00
CD

Q.
CO

O)
O

in

u

E
£

00

>^

in

o

E
o

O

CO

h-

Herpetologica

•

•

•

•

WH

Copeia

•

•

WH

Journal of Herpetology

•

•

WW

Ecology

•

O

WW

Ecological Monographs

•

WW

Herpetological Review

•

•

•

•

Primarily US

American Midland Naturalist

•

o

•

Primarily CUS

Great Basin Naturalist

•

o

•

Primarily WUS

Southwestern Naturalist

•

•

O

•

Primarily SW and ME

Academy of Sciences publica-
tions (Arizona, Utah, Kansas,
Texas, etc.)

•

•

•

•

•

Limited generally to regions
around applicable states

Bulletin of the Maryland Herpeto-
logical Society

•

•

•

•

•

Primarily US and ME

American Museum Novitiates

•

WW

University of Kansas Publications
in Zoology

•

•

•

Primarily CUS and WUS

University of California Publica-
tions in Zoology

•

•

•

Primarily WUS

Smithsonian Herpetological Infor-
mation Service

•

•

•

•

WW

° Limited

272

Amphibians and Reptiles

.

Table 2. Examples of major references providing preliminary and supplemental data for amphibian and
reptile surveys and studies (concluded).

Journal or Publication

Types of Information

Geographic Applicability

2
o
>

CO

.c

03
CD

03

c
g

"3
n

00

b

CO

o
D)
O
O

o

LU

00

cu

CJ 00

cd o.

" "to
■*-> d

So
5 ' «
1*

c
g

to
o

"c

CD
"O

c

CD

E

CD
O)
CO
c
CO

2

CO
O
O)
O

o

00

Q.

00
CD

a

CO

a>
o

n

CD

o
to

E

0)

w
>.

o

'E
o

c
O
X
CO

1—

Examples of books with broad geographic coverage

Turtles of the United States by C.
H. Ernst and R. W. Barbour
(1973)

•

•

US

Handbook of Lizards by H. M.
Smith (1946)

•

•

US

A Field Guide to Western Reptiles
and Amphibians by R. C. Steb-
bins(1966)

•

cus, wus

Handbook of Frogs and Toads of
the United States and Canada
by AH. Wright and A. A. Wright

•

US and CA

Handbook of Snakes of the
United States and Canada by
A.H. Wright and A.A. Wright
(1957)

•

US and CA

Handbook of Salamanders by
S.C. Bishop (1947)

•

•

•

•

•

•

•

•

US and CA

A Field Guide to Reptiles and
Amphibians of Eastern and Cen-
tral North America by R. Conant
(1975)

•

•

•

•

•

•

•

•

CUS and EUS

Example of books with regional geographic coverage3

Amphibians and Reptiles of the
Pacific Northwest by R. Nuss-
baum et al. (1983)

Be sure to check thoroughly for these. There are many regional- and area-specific books.

Codes for Geographic Applicability:

CA —Canada
CUS— Central United States
EUS —Eastern United States
ME — Mexico

SW — Southwestern United States
US —United States
WH — Western hemisphere
WUS— Western United States
WW —Worldwide

rather than man-caused changes. Gibbons and Sem-
litsch (1981) recommend long-duration studies (> 5
years) in assessing amphibian and reptile popula-
tions. Shorter-duration studies in an area can yield
species occurrences, and they are also used to com-
pare composition of herpetological communities
provided sample areas are roughly adjacent to each
other and climatic factors are considered
(Jones 1981a).

Other major factors affecting amphibian and
reptile sampling are differences in species morphol-
ogy, physiology, and behavior such as activity pat-
terns and movement. For example, certain lizards
and snakes are too large (e.g., desert iguana) to cap-
ture in pitfall traps. Other methods should be used
in conjunction with pitfall trapping to adequately
census large lizards and snakes. Populations of rela-
tively sedentary, microhabitat-specinc animals are

Amphibians and Reptiles

273

also underestimated by pitfall and other trapping
methods because the probability of these animals
falling into traps is far less than those species that
move over larger areas. An active searching method
such as turning logs, rocks, and debris (Bury and
Raphael 1983) would yield more accurate estimates
of sedentary species occurrence and abundance than
traps.

The frequency and type of amphibian and rep-
tile movement will also affect sampling. For example,
there are many burrowing, nocturnal snakes in the
Sonoran Desert. Transects run during daylight hours
rarely verify nocturnal, burrowing species even
when rocks and debris are searched. Addition of
pitfall traps and drift fences to a diurnal transect
greatly increases the accuracy of the sample.

Yearly fluctuations in amphibian and reptile
activity also affect verification of species occurrence
in an area. Toads within the Desert Southwest are
active primarily during thunderstorms in July and
August. Samples taken other than in July and August
yield poor results.

Certain microhabitat-specific species tend to
move in nonrandom patterns between preferred hab-
itats (see Pianka 1966 for lizard examples), and this
affects accuracy of census methods (Gibbons and
Semlitsch 1981 ). Placement of traps and fences is ex-
tremely important when considering different move-
ment patterns in lizards. For example, zebra-tailed
lizards move between large bushes along drainages
(Pianka 1966). Pitfall traps and drift fences placed in
open areas with few bushes will catch few zebra-
tailed lizards, whereas fences and traps placed be-
tween bushes will catch more lizards.

Daily and weekly weather patterns also affect
amphibian and reptile activity. Species with narrow
preferred temperature ranges are more likely to be
active during a smaller range of environmental tem-
peratures. Therefore, samples can be variable during
weeks with large temperature differences. There may
also be considerable daily and seasonal differences
in movement between different age and size classes
and sexes.

Abilities of the biologist will also affect the accu-
racy of samples, especially transect and tape-record-
ing techniques where animals are not closely
viewed. For example, a high degree of expertise is
needed to identify lizards on a walk-through transect.
Similarly, a biologist must be familiar with toad and
frog calls if auditory data are to be used for species
identification. Conversely, pitfall and other trapping
methods do not require specific expertise because
amphibians and reptiles can be closely viewed.

Other life history limitations may also affect the
interpretation of sampling data. Whereas one lizard

may produce multiple clutches within a year, an-
other may reproduce only every other year. This
type of result could be easily misinterpreted if the
biologist does not consider differences in reproduc-
tive strategies. Similarly, size and age class ratios
can be misinterpreted if the surveyor is not familiar
with differences in species' life history. For example,
juveniles tend to move less and, therefore, would
be underestimated by techniques such as pitfall
trapping.

Relatively accurate samples can be obtained for
individual species and entire amphibian and reptile
communities provided that species' life histories and
sampling method limitations are clearly understood.

Sampling Methods

There are many methods used to sample am-
phibians and reptiles. The method chosen by a biolo-
gist will depend on the objectives. For example, if
biologists need to develop a species list by habitat
type for a planning document, then they should se-
lect some combination of verification techniques.
However, if biologists need to show differences in
species abundance between habitat types, then they
must select a method that yields relative abundance.
Generally, data on species presence are adequate
for assessing species richness of habitat types. Abun-
dance and density techniques provide additional data
for comparing habitat types and also provide infor-
mation about individual species fitness. For example,
density estimates provide biologists with data on
the relative fitness of special status reptiles such as
the desert tortoise (Gopherus agassizii).

Depending on the biologist's objectives, there
are generally two ways to collect species informa-
tion: direct and indirect. Direct sampling of amphibi-
ans and reptiles involves observation of animals
occurring on a sample site. Indirect sampling in-
volves obtaining species information on a sample site
without observing the animal. Before initiating any
sampling, the biologist should contact the State game
and fish agency to obtain the necessary regulations
and collecting permits.

Direct Search Methods

One of the simplest ways of verifying species
occurrence in an area is to walk or drive through
the area, recording all amphibians and reptiles seen.
Because species are separated in time as well as
space (see Creusere and Whitford 1982 for an exam-
ple in a lizard community), searches run at different
times of the day will yield different species in vary-
ing numbers (searches run during the daylight hours
verify some diurnal species and those run at night
some nocturnal species).

274

Amphibians and Reptiles

Searches of areas are either random and oppor-
tunistic or systematic within a defined time and area.
Many records of amphibians and reptiles result from
casual, opportunistic observations made during field
work. Cumulatively, these observations have contrib-
uted more to known occurrence of amphibians and
reptiles than any other method. Although this
method often yields valuable locality records, it gen-
erally involves a great deal of search time. It also
verifies only those species that spend a great deal of
time on the surface or under rocks and debris such
as logs. Generally, only a small percentage of species
that occupy an area are verified during searches.
Only through months and often years are all species
in an area verified by search methods.

Road riding (road cruising) is one of the most
popular search methods used to verify and collect
amphibians and reptiles, especially nocturnal species
(see discussion by Campbell and Christman 1982).
Generally road riding consists of cruising secondary
roads at speeds of 35-55 km/h (22-34 mph), using
low headlight beams. Many small nocturnal snakes,
lizards, and frogs not normally found during daytime
field searches are obtained by this method when
conducted between dusk and 2 to 3 hours after
dusk. Use a spotlight or strong flashlight while road
riding to spot animals moving off of roads. Flashlights
are useful for locating amphibians and reptiles on
sides of roads (see Vitt and Ohmart 1978 for an ex-
ample of road-riding results). Road riding is most
productive when run without moonlight. Road-riding
routes can be systematically established, and relative
abundance of amphibians and reptiles can be ob-
tained by standardizing sampling efforts by units of
time in each habitat. For example, biologists can
measure the distance of each habitat along a road,
and measure the amount of time spent in each seg-
ment of the habitat. They can then express individ-
ual species abundance data as the number of animals
seen per mile of habitat per hour. This allows for
comparison of habitats using road-riding procedures.
Bury and Raphael (1983) refer to searches con-
ducted per unit effort of time as time-constraint
procedures.

Although providing occurrence and abundance
data on some nocturnal and secretive herpetofauna,
road riding:

( 1 ) is time-consuming;

(2) yields relatively few records;

(3) verifies only nocturnal species that migrate
across roads;

(4) biases samples because it is limited to areas
with roads; and

(5) is sometimes dangerous to observers, espe-
cially on well-traveled routes.

Tortoises (especially desert tortoises) and other
turtles are often verified by systematic searches.
Burge (1979) used a series of 1.6-km and 4.8-km (1-
mi. and 3-mi. ) search transects in the Arizona Sono-
ran Desert to verify occurrence of desert tortoises.
The 4.8-km ( 3-mi. ) transect consisted of a triangle
(1.6 km [1 mi.] each side) set on a random compass
line. Along each line, observers look for live tor-
toises, scat, burrows, and tracks. Data are then re-
corded on standard forms.

Systematic searches of defined areas yield spe-
cies occurrence data and allow for further assess-
ment of populations of amphibians and reptiles. This
method is especially valuable when the biologist
needs population estimates of special status amphibi-
ans and reptiles. Bury (1982) used 2-ha (4.9-a.)
quadrats to verify and study diurnal reptile popula-
tions in the Mojave Desert. He used a removal
method to determine densities of reptiles on his
quadrats (discussed in more detail later in this chap-
ter). Bury and Luckenbach ( 1977) successfully cen-
sused tortoise populations with a quadrat and grid
location system. They established 25- to 100-ha (62-
to 247-a. ) quadrats, subdividing each into 1-ha (2.47-
a. ) sections. Systematic searches of each section ac-
curately located tortoises and served as permanent
sites for monitoring tortoise populations. Schneider

(1981) used 2.59-km (1-mi. ) plots to collect popu-
lation data on desert tortoises in Arizona. Both Bury

( 1982) and Schneider ( 1981 ) used systematic
searches of these grids to determine occurrence and
population structure.

Systematic search procedures in defined areas
are very time-consuming and should only be used if
the biologist needs an accurate estimate of popula-
tion density. Generally this involves only special
status species such as federally threatened and en-
dangered amphibians and reptiles where density data
are needed to assess populations.

Another systematic search method commonly
used involves scouring known habitats of certain
amphibians and reptiles. For example, canyon tree
frogs (Hyla sp.) are most easily observed in rocky,
boulder-strewn canyons with permanent water, pri-
marily pools (Jones 1981b). Secretive diurnal and
nocturnal species can be verified along systematic
search routes by lifting rocks, vegetative debris, and
uprooting animal shelters such as pack rat dens.
Zweifel and Lowe ( 1966) obtained desert night liz-
ard specimens (Xantusia vigilis) by lifting downed
Joshua tree (Yucca brevifolia) limbs. Without search
of their preferred habitat (surface debris), these liz-
ards cannot be verified. Amphibians and reptiles also
use a variety of other animal cover sites; therefore,
a large number of amphibians and reptiles can be
verified by searching mammal dens and burrows
(Lee 1968).

Amphibians and Reptiles

275

These microhabitat specific searches are gener-
ally used to verify and collect data on the abundance
of a few species. If used for verification only, this
method can be quick and easy. When relative abun-
dance or density is needed, this method is consider-
ably more time-consuming.

There are several small tools that increase effec-
tiveness of searches. Mirrors, in particular, locate
and identify animals in burrows, especially turtles
and snakes. Similarly, flashlights increase light in and
around burrows and are extremely helpful in verify-
ing species such as tortoises.

Several types of poles have been used to locate
inactive, partially, or completely concealed reptiles,
especially turtles. Noodling is an effective way to
locate aquatic turtles concealed in muddy creek
bottoms (Lagler 1943). The technique involves
cruising shallow lakes and ponds using a blunt steel
rod to protrude through mud and emergent vegeta-
tion. A hollow sound is given off when a turtle is hit.
Noodling can also be used to protrude through ter-
restrial litter and soil to verify hidden turtles
(Carpenter 1955).

Another extremely effective method of deter-
mining species occurrence in an area is to search
nests of predators for amphibian and reptile remains.
Large proportions of certain raptor diets consist of
reptiles (Millsap 1981). A total of 13 species of rep-
tiles were verified in and below red-tailed hawks
(Buteo jamaicensis) nests by Millsap (1981). Red-
tailed hawks forage over relatively large areas and,
therefore, reptile occurrence in nests cannot be cor-
related with specific areas or habitats. However,
some raptors forage over extremely limited areas or
in specific habitats (e.g., riparian and aquatic habi-
tats). For example, black hawks are highly dependent
on aquatic prey within western Arizona (over 90%
of their diets are aquatic organisms). Millsap and
Harrison ( 1981 ) identified six species of reptiles at
black hawk nests, including a mud turtle {Kinoster-
non sp. ). Generally raptor nests cannot be used to
identify amphibian species because entire animals
are usually consumed. However, prey items can be
identified as they come into raptor nests.

Owl pellet dissections also verify amphibian and
reptile occurrence in an area (B.A. Millsap, pers.
commun. ). Positive identifications are obtained by
comparing samples with skeletal references.

Although examination of predator nests and
feces provides species occurrence for a region, it
does not consistently indicate relative abundance of
amphibians and reptiles because predators eat what
they prefer and not necessarily what is most
abundant.

Black hawk feeding nestling.

Trapping and Collecting Methods

Trapping methods have been designed around
specific life histories of certain amphibians and rep-
tiles. For example, a trout fly was extended over a
pitfall trap to lure zebra-tailed and fringe-toed lizards
(Uma sp.) into traps (Lannon 1962). Both of these
lizards are sit-and-wait species that forage out in
open spaces (see Pianka 1966). They remain rela-
tively stationary until visual contact is made with an
invertebrate and then attempt to capture and swal-
low the prey. Although Lannon's method may attract
lizards that forage in similar ways on similar prey, it
will not attract other amphibians and reptiles that
forage in different ways.

Two basic procedures exist for collecting or
capturing amphibians and reptiles: direct and indi-
rect capture. Direct capture consists of actively seek-
ing animals and using various equipment to increase
capture success. Indirect capture involves use of
equipment capable of securing animals without ac-
tive searching by the biologist.

Several types of equipment are used to capture
animals during active pursuit. I will highlight some
of the most commonly used equipment.

Large samples of frogs, toads, salamanders, tad-
poles, and turtles can be obtained because their hab-
itats (water and water/land ecotones) are limited to
relatively small areas (as compared with terrestrial
habitats).

Dip nets increase effectiveness of capturing tur-
tles, frogs, toads, and salamanders during active pur-
suit or while turning stones and logs in streams.
Svihla (1959) greatly increased capture of tailed frog
tadpoles (Ascaphns truei) by simultaneously placing
a dip net directly downstream while turning rocks.
Without the net, many tadpoles escaped unnoticed.

276

Amphibians and Reptiles

Lagler (1943) also described dip nets of various sizes
used to collect small tadpoles, salamander larvae,
and small turtles.

Gunning and Lewis (1959) reported low mortality
rates when using shockers that generated direct cur-
rent fields.

Seines, placed across a stream or creek, greatly
increase capture success of salamander and salaman-
der larvae, and frog and toad tadpoles (Balgooyen
1977). With a seine placed downstream, biologists
move toward the net, turning rocks, debris, and veg-
etation. Disturbed animals are then swept down-
stream into the net by the stream's current.

Stationary, polyethylene plastic dams can be
used to corral stream-dwelling organisms including
shovel-nosed salamander larvae ( Leurognathus sp. )
(Martof 1963)- The damming effect causes larval
salamanders to congregate at the bottom of the plas-
tic sheet.

Electroshocking is an effective means of collect-
ing and censusing aquatic habitats. Although more
traditionally used for collecting fish, direct current
electroshockers have been used to census aquatic
amphibians and reptiles (Gunning and Lewis 1959;
Williams et al. 1981 ). Electroshocking stuns adult
and larval salamanders, adult and larval frogs, and to
a lesser degree, small turtles, making capture easier.
Several combinations of aquatic collecting tech-
niques were tested on hellbenders (Cryptobranchus
alleganiensis). Combinations of electroshocking
and dip nets were most effective although results
varied in different aquatic habitats. Electroshocking
and dip nets were most successful in pools and slow-
moving stream sections because dip nets capture
hellbenders before they sink. Conversely, seines cap-
tured hellbenders more effectively in fast-moving,
riffle-dominated stream sections because animals are
more easily swept into a seine than into a dip net.
Williams et al. ( 1981 ) found conventional dip-net-
ting and seining to be inferior to netting with elec-
troshocking. Both Williams et al. ( 1981 ) and

Abundance and density of aquatic amphibians
can be determined for a reach of stream by using the
previously described equipment. First, an area of the
stream is selected and measured in square meters.
Second, a seine is stretched across the stream at the
station's farthest point downstream. Finally, the biol-
ogist systematically walks downstream, swooping
the electrical field across the stream. Shocked am-
phibians are then placed in the net and counted. A
second swoop is conducted to shock the remaining
amphibians. The biologist then obtains a number
of each species for the area shocked.

Seining by itself provides verification of species,
but is too slow to provide an absolute count (unless
sampling a small isolated pool). Electroshocking,
although more effective, is generally not effective on
large species with small surface to volume ratios.
As a quantitative method, electroshocking is limited
to relatively narrow, shallow, non-turbid reaches.

Anderson and Smith (1950) designed and tested
an electrical apparatus that effectively stunned both
aquatic and terrestrial amphibians and reptiles (Fig-
ure 1 ). Mortality rates of shocked animals varied,
usually with size. Large snakes such as water mocca-
sins (Agkistrodon sp. ) must be shocked on the up-
per third of their body to obtain paralysis, whereas
single shocks to small amphibians (e.g., tree frogs)
are often fatal.

In water with thick vegetation and organic de-
bris, dredges can increase collecting success. Goin
( 1942) used a double-trough dredge held together
by hinges (Figure 2). The upper trough is con-
structed of larger mesh hardware cloth than the

ELECTRODES

kmd

mm

BATTERIES

Figure 1. Electrical stunning apparatus.

SWITCH

S^MI —

MMm

m

aSwstsS

lli

COIL

PLUG

Amphibians and Reptiles

277

Figure 2. Double-trough dredge.

lower trough. Vegetation-dwelling larvae and adult
salamanders, tadpoles, frogs, and turtles are captured
when the dredge is lifted up through vegetation
and debris. Different sized animals can be collected
by varying the size of the screen mesh.

Other methods, such as gaffing and setting lines
of baited fish hooks, have been used to verify aquatic
turtles (Lagler 1943). A gaff consists of a sharp hook
placed on the end of a pole. Turtles are caught by
spearing the hook through the shell. This method, al-
though not used as a quantitative procedure, pro-
duces records of large turtles.

A snare has been designed and successfully
tested to capture water snakes (Franklin 1947). The
snare consists of a three -prong sterilizing forcep
attached by a scissor grip to the end of a 3.7-m (12-
ft) lightweight cane pole. The snare is particularly
effective in capturing large, swimming water snakes
(e.g. water moccasin) and provides a means of veri-
fying certain species.

An alternative to collecting amphibians and rep-
tiles with a gun is to use a rubberband mechanism.
This type of collecting method is considerably safer
than a firearm and, with practice, yields live speci-
mens. Neill (1956) and Dundee (1950) designed
effective rubberband-launching devices that they re-
ported using successfully to immobilize frogs and
lizards (see Figure 3).

Hand snares are used to secure reptiles, espe-
cially lizards. Eakin ( 1957) designed a copper wire

Figure 3. Rubberband-stunning devices.

slip noose (Am. Stand. Wire, gauge 34, [0.15 mm])
attached to a light pole. Lizards are captured by
slowly placing the noose over their heads and jerk-
ing the wire tight. Stickel (1944) designed and suc-
cessfully tested a snare that has a trigger mechanism
and a thread (Figure 4). He stated that his mecha-
nism is superior to hand-pulled snares because of the
speed of the snare mechanism. Fishing line (0.91-
1.36 kg [2-3 lb.] test) can also be used to form the
noose (R. Bowker, pers. commun.).

All of the previously described equipment such
as snares, nooses, and dredges help biologists verify
occurrence of species on a site or in a habitat. Rarely
is this equipment used to determine relative abun-
dances or densities unless they are part of a system-
atic search or procedure.

Perhaps the most widely and successfully used
amphibian and reptile collecting and censusing tech-
niques are those that involve indirect capture, specif-
ically traps and funnels. Traps and funnels have the
advantage of capturing animals while the biologist is
not present. Generally they provide a greater num-
ber of species per sampling effort than active, direct
capture methods. Funnels and traps are used widely
for both terrestrial and aquatic sampling. Most trap
and funnel procedures are designed to maximize
capture of certain types of herpetofauna. Their place-
ment in certain microhabitats and the use of differ-
ent baits also bias sampling toward certain species.

Several investigators successfully designed and
tested hoop funnel traps for capturing turtles (Lagler
1943, I960; Pirnie 1935; Iverson 1979; Feuer 1980).

278

Amphibians and Reptiles

TRIGGER STICK

RUBBER BAND
.TRIGGER NAIL

Figure 4. Trigger snare.

Generally, traps consist of mesh netting connected
to four aluminum hoops or guide lines (Figure 5). A
purse-string regulated opening provides turtles with
access through the throat to a piece of suspended
bait. Funnels can have single or double throats. Iver-
son's ( 1979) funnel trap is constructed of chicken
wire rather than netting. These funnel traps are most
effective in shallow water and should be checked
frequently to avoid drowning the specimen.

Funnel traps have also been designed to capture
larvae and adult salamanders and frogs (Carpenter
1953; Moulton 1954). Carpenter's trap consists of a

6.35-mm (0.25-in.) mesh hardware cloth formed into
a funnel that is lowered into deep pools by a 1 .8- to
3.0-m (6- to 10-ft) rope. This method collects larval
and adult salamanders in pools too deep for seines.
Funnel openings should face toward shores of ponds
to take advantage of animal movements.

Moulton's trap is slightly different than Carpen-
ter's, consisting of a rectangular wire screen box
placed on a wooden frame with an access funnel
placed 63.5 cm (25 in.) above the box floor. The
trap is placed in a shallow depression along the
shore of a pond or lake with the funnel facing to-
ward land. Salamander and anuran prey accumulate
in the trap and lure animals toward the trap during
movements between land and water.

Funnel traps have been used extensively in cap-
turing terrestrial and semiaquatic frogs, salamanders,
lizards, and snakes. Generally these traps function
similarly to aquatic funnel traps; they encourage
entrance into traps and restrict access out. These
traps are usually constructed of hardware cloth,
shaped into cylindrical form, and accessed at one or
both ends by funnels (Vogt 1941; Lagler 1943).
Many minor variations of this general construction
have been used. Fitch (1951) used a transparent,
cellulose pivotal door at the entrance to the trap,
preventing lizards and snakes from crawling out (Fig-
ure 6). Clark (1966) and Hall (1967) further modi-
fied the trap by using mason jars and 0.45-kg (1-lb)
coffee cans, respectively, as the collecting area. Both
used hardware cloth to funnel animals into the trap
(Figure 7). Many of these investigators used bait
to attract animals.

Many investigators have also combined funnel
trapping with other techniques, especially use of
drift fence. Dargan and Stickel (1949) used a 0.1 0-m
(4-in.) high, 7.6-m (25-ft) long hardware cloth drift
fence to guide snakes into hardware cloth funnel

*' ***•

Figure 5. Standard funnel trap for capturing turtles.

Drift fence reptile trap.

Amphibians and Reptiles

279

traps. Clark ( 1966) found that rolled aluminum
(0.019 gauge) and board drift fences (2.4 to 3.0 m
[8 to 10 ft]) used with funnel traps increased trap-
ping success of small snakes. Migrating salamanders
and frogs can be surveyed by completely enclosing
ponds with 6.35 mm (1/4 in.), 0.25-m (104n.) high
hardware cloth, placing hardware cloth and bronze
screen funnel traps inside and outside the enclosure
along the fence (Storm and Pimentel 1954).

Pitfall trapping is one of the most widely used
methods for collecting data on amphibians and rep-
tiles. Generally this method involves placing a square
or round container below water or ground with the
container top flush with the surface. Container sizes
and shapes vary greatly, depending on which species
is trapped.

Floating pitfall traps can be used in water for
capturing turtles. Lagler (1943), Breen (1949), and
Petokas and Alexander (1979) successfully used
these traps for turtles in ponds and lakes. Lagler
( 1943) and Petokas and Alexander ( 1979) used a
floating wood frame with a submerged trap to cap-
ture basking turtles (Figure 8). Breen (1949) used a
submerged wooden barrel with the top above the
water surface. Wooden planks were extended from
the barrel top to the water with bait placed at the
top of each plank. Turtles fell into the barrel while
attempting to swallow bait. Pitfall traps capture
mostly basking turtles that spend time on or above
the water's surface.

Pitfall traps have been used extensively in ter-
restrial amphibian and reptile surveys and studies.
Generally pitfall traps are various sized containers
placed in the ground with tops flush to the surface.
There are a number of variations in pitfall traps rang-
ing from wooden box traps to 18.9-L (5-gal.) plastic
containers.

Rogers (1939) used a trapdoor box-trap buried
with the top flush to the surface to capture lizards.
Banta (1957) used 4.7-L (5-qt) tin cans and covered
tops (cardboard) leaving a few inches between tops
and pitfall openings (Figure 9). Lizards fell into traps
while attempting to seek shade. Whitaker (1967)
increased pitfall trapping success of some lizards by
using fruit and honey bait. Besides providing occur-
rence data, pitfall traps obtain specimens for popula-
tion, reproductive, and taxonomic studies. Porzer

(1982) used double-depth 18.9-L (5-gal.) plastic
containers to capture Gila monsters (Heloderma sus-
pectum) for radio telemetry studies.

When arranged systematically and standardized
per unit effort (e.g., animals caught in a 24-hour
trapping period), pitfall and other trapping data can
be quantified and compared. For example, biologists
can compare relative abundance of certain species
between two or more habitat types if they set out
traps in a standardized way (similar arrangement and
size) during the same season. Bury and Raphael

(1983) described a systematic pitfall trapping system

Figure 6. Funnel trap for capturing terrestrial amphibians and reptiles.
280 Amphibians and Reptiles

Figure 7. Jar funnel trap.

consisting of 7.6-L (2 gal.) sized pitfall traps, ar-
ranged in a 5- x 13-cm (2- x 5-in. ) grid, 20 m (66 ft)
apart. This procedure captured seven times the num-
ber of lizards than a technique of searching a prede-
fined area, lifting logs and rocks (previously
described as time-constraint collecting), but pro-
duced only one-half the number of salamanders. K.B.
Jones (unpubl. data) and Price (pers. commun. )
used a large (26- x 30-cm [10- x 12-in.] trap grid)
2-ha (5-a.) pitfall grid to determine relative abun-
dance of lizards in riparian and Chihuahuan Desert
habitats, respectively (Figure 10). Both used double-
deep, 1.35-kg (3-lb) coffee cans spaced 15 m (49.5
ft) apart with wooden tops placed 15.2-20.3 cm (6-8
in.) over the traps for shade. Jones and Price
(unnubl. data) accumulated large samples of most
lizards in a period of less than a month during the
spring.

By marking amphibians and reptiles for individ-
ual identification while using a pitfall grid, biologists
can determine species home ranges and population
sizes (mark and recapture).

Drift fences have been widely used in combina-
tion with pitfall traps, especially in recent years.

Figure 8. Floating pitfall trap for capturing turtles.

Gibbons and Semlitsch (1981 ) used several configu-
rations of pitfall traps and drift fences for capturing
amphibians, reptiles, and small mammals. They found
these methods yielded a large number of species
and accurately estimated the abundances of readily
trapped species. They also provide time and cost
estimates for trapping. Campbell and Christman
(1982) described an array-trapping method consist-
ing of eight 18.9-L (5-gal.) plastic containers, 7.6-
m (25 -ft ) drift fences, and hardware cloth funnel
traps ( Figure 11). The array technique yielded a
large number of amphibian and reptile species, espe-
cially secretive species not normally verified by
other techniques. Jones (1981a & b) modified Camp-
bell's and Christman's (1982) array system by using
only three drift fences and four 18.9-L (5-gal.) plastic
containers (Figure 12). As in Campbell and Christ-
man's ( 1 982 ) studies, Jones verified a large number
of lizards and snakes, especially secretive forms, but
with about one-half the equipment at each site.

Amphibians and Reptiles

281

Figure 9. Simple pitfall trap for terrestrial
amphibians and reptiles.

Vogt and Hine (1982) also used a wide variety
of pitfall/drift fence configurations (Figure 13) to
sample amphibians and reptiles. These configurations
were designed to maximize trapping success in each
habitat. Although Vogt and Hine (1982) reported
drift fence lengths of less than 15 m (49. 5 ft) as
generally ineffective, Jones (1981a & b) demon-
strated that a 7.6-m (25-ft) fence length was very
effective. These differences may reflect different
types of habitat and animals sampled by the two in-
vestigators (wet woodland versus desert).

The sampling duration and seasons when array
or pitfall/drift fence trapping is most successful de-
pends on general climatic features of an area, and
the species and habitats to be sampled. For example,
Vogt and Hine (1982) suggested that several short
sampling periods, staggered over an activity season
(especially after rains), yield the most accurate esti-
mate of species composition and species abundance.
They believed this procedure takes advantage of
increased herpetofauna activity resulting from rain.
Unlike Wisconsin, the Desert Southwest receives
infrequent precipitation and has a far greater number
of habitats where temperature and moisture availabil-
ity vary. In Arizona, desert habitats are most effec-
tively trapped in April and May, and forest and
woodland habitats between May and July. I have also
found that broken, short-duration sampling suggested
by Vogt and Hine (1982) does not verify all species,

v

/ II buckets

o

o

o

o o o o

o

o

o

o o o

o

o

o

Grid area = 1 ha

\1/

110 buckets

10 CK

o

> 1 5m

o

oJ

Double-deep 31b.
coffee can

tape

Figure 10. Grid pitfall trapping system.

especially many secretive desert snakes, present in
a habitat. I attribute this finding to infrequent, erratic
rainfall and suggest sample periods of not less than
30 days.

To obtain differences in species composition
and abundance, biologists should place arrays in
different habitats, provided that a standard configura-
tion is used. Fences located so that amphibian and
reptile migration routes are dissected will increase
capture success. Jones (1981a & b) used arrays for
assessing lizard and small snake relative abundances
in a variety of habitats in western Arizona. Array
abundance data can be expressed as the number of
animals caught in 24 hours (Jones 1981a & b).

Pitfall/drift fence or array trapping most accu-
rately assesses lizard and small snake composition

282

Amphibians and Reptiles

and abundance, but is less effective in estimating
salamander, frog, and toad abundances, although it
does provide some excellent records for these her-
petofauna. This method is not recommended for
assessment of most terrestrial turtles and large
snakes. There is some amphibian and reptile mortal-
ity, especially salamanders, frogs, and toads, associ-
ated with most trapping methods previously listed.
To reduce mortality, biologists should check traps as
often as possible, generally every other day. Mortal-
ity due to exposure can be reduced by placing traps
in shaded areas, by covering trap openings with
shades, or by placing about 2 to 4 cm (1 to 2 in. ) of
soil or litter at the bottom of the trap. These precau-
tions will also reduce predation on trapped animals.

Water accumulation in pitfall traps, especially in
wet regions, can cause significant mortality. Whereas
punching holes in the bottom of traps provides little
drainage, floatable objects such as styrofoam reduce
mortality due to drowning.

Q

FENCE

FUNNEL TRAP

6—

CAN TRAP

76M

O

-O I5M O

-o

Q

DO

6

Baited snap traps, such as those commonly used
to census small mammals, can be used to collect
large diurnal lizards. Heatwole et al. (1964) trapped
several large, diurnal lizards by baiting museum snap
traps with live insects, meat, fruit, and peanut butter.

Figure 11. Array pitfall trapping technique, with
funnels.

Table 3. Comparison of mean relative abundance and species diversity for lizards of five heavily grazed
vegetative communities. Tests are based on Student's t-test values at the 95% confidence interval. Each
heavily and lightly grazed site is based on seven sample sites.

&

it

K

H = Heavily grazed L = Lightly grazed

Vegetative Community

Mean Relative
Abundance ± SD

Species
Diversity (H')

Chaparral

Grazed

Array-
Nights

Lizards
Trapped

1.18 ± 32
1.69 ± .41a

1.02
1.09

H

L

672
623

782
1050

Desert grassland

H
L

294

224

179
176

0.56 ± .08
0.77 ± .13

0.88
1.01

Mixed riparian
scrub

H
L

392

252

295

310

0.73 ± 0.5
1.20 ± .06d

0.69
0.90

Cottonwood-willow
riparian

H
L

658
175

419
201

0.64 ± .23
1.13 ± .15a

0.59
0.86

Sonoran desert
scrub

H
L

714
238

757
244

1.06 ± .10
1.03 ± .12

0.93
0.93

aThe comparison of heavily grazed with lightly grazed is significantly different at the 95% confidence interval in three vegetative
communities.

Amphibians and Reptiles

285

Large, man-made developments such as canals
act as large traps and can provide valuable distribu-
tional records of amphibians and reptiles. Hawken
( 1951 ) captured large numbers of amphibians and
reptiles by a flume located downstream from a lake.

A total of 582 specimens (15 species) of amphibians
and reptiles were trapped by this man-made system.

Because some records obtained during trapping
represent significant range extensions, biologists
should collect voucher specimens of amphibians and
reptiles taken out of their published geographic
range. Pisani (1973) discussed field notes and preser-
vation techniques for amphibians and reptiles. After
preserving range extension specimens, biologists
should contact regional experts to verify records.
These animals should then be deposited in a univer-
sity, regional, or national museum. For example,
range extension specimens collected by the BLM's
Phoenix District were placed in the National Mu-
seum of Natural History at the Smithsonian Institute
in Washington, DC.

Certain field procedures have been developed to
identify amphibians and reptiles where direct obser-
vations are not possible. Many frogs and toads pro-
duce specific calls that can be recorded and used to
identify individual species (see Michaud 1964;
Brown and Littlejohn 1972; Fouquette 1980 for
examples).

Certain species of reptiles, especially lizards and
snakes, leave characteristic tracks and when com-
pared to standards developed for each species, can
be used to identify species occurrence in an area.
Lillywhite (1982) used tracks left in fine-textured
soils and on roads to identify common snakes such
as sidewinders (Crotalus cerastes) and whipsnakes
{Masticophis sp. ). Although this procedure provides
some indication of species and habitat use, track
identification does not provide accurate estimates of
abundance, and it may not be practical to use be-
cause of precipitation and wind. A biologist may
want to use this method when movement and habi-
tat use data are needed for priority species that leave
easily-recognized tracks.

bushes

7.6 M

drift fence placed to
cut off migrational
routes between vege-
tation

Figure 12. Modified array pitfall trapping technique.
284 Amphibians and Reptiles

Species and Population Measurements

Depending on the specific objectives of a survey
or study, a number of morphological, behavioral,
and ecological measurements can be taken. Only a
few measurements such as general age class and size
are taken from opportunistic and systematic observa-
tions when animals are not captured.

Capture methods allow biologists to obtain a
number of measurements such as total length (e.g.,
snakes), snout-vent length (lizards), weight, sex
(from coloration or sex probes), reproductive condi-
tion (swollen testes or coloration that indicates
breeding), and limb and other morphological traits
(used to tie species to habitats) such as jaw width/
length ratios. All of these measurements provide
biologists with indicators of species' conditions, es-
pecially when animals are released and recaptured.

Stomach analysis of captured animals can also be
obtained, especially on lizards and snakes. Pietruszka
( 1981 ) found that stomach-flushing devices provided
data on lizard prey. Kephart and Arnold ( 1982) de-
termined garter snake (Thamnophis sp. ) diets by
running their hands tightly up snakes, forcing
regurgitation.

Diet information of lizards and snakes can be
used to determine these animals' food needs, and it
also provides information on the quantity and trends
of certain prey species (see Kephart and Arnold
1982).

Fecal analyses of certain herbivorous reptiles
can be conducted to determine dietary preferences.

■ /

^ /////.
///_///,

////ft//.
^-///^■//.

nil

//////'?".

15m ALUMINUM VALLEY
2) — 3.8m— (in — 7.4m -(IF)— 3.8m— (2'

/////////////////
/////////////////
/////////////////

'/////////
'///////

'///?

22.5m ALUMINUM SCREEN
5j— 3m — (5) — 6m (2) — 9m — ( I) — 4.5m — (5'

15m ALUMINUM VALLEY
5) 7.5m (IT)

15m-

FUNNEL TRAP

FUNNELTRAP

30m ALUMINUM VALLEY

■7.5m-

ShUNNtL
TRAP
FUNNEL
TRAP

15m ALUMINUM VALLEY

Figure 13. Pitfall/drift fence trapping schemes in different habitats.

Hansen et al. (1976) used a microhistological analy-
sis of fecal droppings to determine dietary prefer-
ences of the desert tortoise.

When individual amphibians and reptiles are
marked, released, and recaptured at a later date, a
number of statistical treatments can be used to de-
termine species' density, movement, and activity
patterns. Capture of amphibians and reptiles allows
biologists to mark individual animals for future iden-
tification. A wide variety of marking techniques,
ranging from tagging (including radioactive) to se-
quential toe-clipping, have been used. Pough (1970)
attached plastic plug tags to identale clipped just
anterior to the cloaca (Figure 14). Shell notching is
perhaps the most common method of marking tur-
tles, although tagging, branding, and painting are also
used for individual identification (Ferner 1979;
Figure 14).

Heckel and Roughgarden (1979) used a paint
spray gun to mark lizards. Although the marking
technique did not provide individual lizard recogni-
tion necessary to determine movements, it provided
quick and accurate data necessary to determine
home range and population size.

Although more commonly used for fish, removal
methods yield data that can be submitted to certain
statistical treatments to estimate amphibian and rep-
tile population size. Bury (1982) used a removal
method to estimate population size of diurnal lizards.
He verified all resident diurnal lizards after 2 days of
sampling and used the total number of lizards re-
moved over the 2-day period as the population size.

Trailing devices and radiotelemetry provide data
on individual reptile movement and activity. Scott
and Dobie (1980) used an aluminum-canister,
thread-trailing device to determine movements of
turtles (Figure 15).

All of the previously described procedures and
measurements can be used by biologists to deter-
mine the relative health of amphibians and reptile
species and communities. Most methods involved
with marking, tracking, and recapture are time-con-
suming and, in some instances, expensive (e.g., te-
lemetry). These methods should be used when
detailed information on the condition of a species
(e.g., threatened or endangered) or a group of spe-
cies in a specific habitat (e.g., riparian) is needed.

Amphibians and Reptiles

285

* Js

Gila monster leaving tracks in the sand.

Community Surveys

Broader, more extensive surveys of entire am-
phibian and reptile faunas in a variety of habitat
types do not allow for intensive study of individual
species. In a community survey, biologists attempt to
determine species composition of each major habitat
type; some rough estimate of abundance; and, if pos-
sible, species' uses of microhabitats.

Because individual amphibian and reptile mor-
phology, behavior, and ecology vary, biologists
should use several censusing methods for determin-
ing herpetofauna community composition. Campbell
and Christman (1982) and Gibbons and Semlitsch
(1981) emphasized the need to use combinations of
opportunistic observations, transects, pitfall and fun-
nel traps, drift fence, and road riding for deriving
complete species lists (occurrences) within specific
areas or habitats. Jones (1981b) used several meth-
ods to determine amphibian and reptile composition
and abundance in 14 major habitat types. Conducted
during spring, summer, and fall of a year, the survey
generated 27 significant range extensions for the
region. Information gained from the survey also con-
tributed heavily to management decisions made for
the area. Bury and Raphael ( 1983) recommend a
combination of time-constraint and pitfall array tech-
niques for generating data useful in making most
types of management decisions.

Herpetofauna richness and abundance data can
help determine the effects of land use on these ani-
mals. Busack and Bury (1974) and Jones (1981a)
provided examples of studies describing land-use im-
pacts, including grazing effects on lizards. Data such
as in Table 3 can be used to assess impacts of land
use.

There are other statistical analyses that assess
amphibian and reptile community structure. These
analyses, which include Horn's Overlap Index (Horn
1966) and several similarity coefficients, are treated
in a later chapter.

DISCUSSION

The primary difference between inventory and
monitoring of amphibians and reptiles and their habi-
tat is the objective established by the biologist. An
inventory and monitoring study may involve similar
data collection methods. However, inventories usu-
ally verify what is there and how habitat resources
are being used, whereas monitoring determines how
individual species or communities change as a result
of specific types of land use. An inventory can suc-
cessfully verify most amphibians and reptiles in rep-
resentative habitat types within a year, provided
that —

( 1 ) there are multiple samples in each habitat;

( 2 ) samples are taken during all peak activity
periods; and

(3) a variety of methods are used.

If funds prohibit complete community samples, the
biologist must then decide which species or habitats
will produce the most useful information for the
money.

Monitoring generally requires sampling over
several years so that species and community health
can be accurately estimated. This is especially
needed in sampling amphibians and reptiles because
populations fluctuate greatly from year to year with
environmental changes, particularly precipitation.
Multiyear data collection allows the biologist to de-
termine which population trends are due to naturally
fluctuating environmental conditions and which ones
are due to land-use practices.

Most agency budgets will not permit long-term,
intensive, multiyear samples of individual species
or entire communities. To offset budget limitations,
biologists should concentrate on long-term changes
in species richness and important microhabitats,
especially when losses are involved. To determine
these changes, the biologist does not need to use
expensive, time-consuming species trend techniques,

286

Amphibians and Reptiles

1000
900

900

t==l

No. 718

Figure 14. Marking techniques for amphibians and reptiles. A & B. Turtles C. Snakes D. Lizards

Anterior anchor wire

Film cam ster

Canister lid (inside
view)

Eye ring

Washer
Machine screw

Machine screw nut

Film canister

ter lid

Machine Posterior Spool and thread

screw anchor (inside canister)

wires

^ Washe r

Cani ster lid //'

fS^

JX\ Machine screw

Posterior /\\

Jl\ Posterior

anchor — _/ x;

;// \-—' anchor

wire /

^\ \ wire

Anterior

Film canister

anchor '

wire

Figure 15. Trailing device to determine turtle
movement.

Short-term method of marking desert tortoises.

Amphibians and Reptiles

287

but rather species verification and habitat measure-
ment techniques. Data can also be obtained intermit-
tently over 3- to 5-year intervals to reduce costs.
Generally these types of data will provide biologists
with adequate information for management deci-
sions, although some management situations may re-
quire intensive, multiyear monitoring.

Another way of reducing costs, while obtaining
community data on amphibians and reptiles, is to
sample indicator species. There are two general
types of indicator species:

• amphibians and reptiles that represent species
assemblages that use habitats in similar ways
(species guilds), or

• species that use specific habitat components.

For example, population trend data can be collected
on one of several species that require downed litter.
The trend of the species sampled should reflect
trends of the entire species assemblage because they
require the same habitat component. However, some
tests have revealed that species within delineated
guilds respond differently to land-use practices. Man-
nan et al. ( 1984) found that 67% of birds within
guilds responded differently to timber harvest. The
problem with guilds is that all species occupy sepa-
rate niches, some more specific than others. There-
fore, many amphibians and reptiles within the same
guild as the indicator may respond differently to
habitat changes. Mannan et al. (1984) pointed out
that guilding works either when the investigator
defines very specific guilds or the land-use impact is
so severe that several habitat components are lost.

The other use of indicator species is to deter-
mine conditions of a habitat component based on
condition and trend of an amphibian or reptile. Al-
though certain microhabitat specific species could be
used in this manner, greater accuracy and less sam-
ple cost may be obtained by sampling the habitat
components directly.

In selecting indicator amphibians and reptiles, a
biologist must again be fully aware of sampling
method and species life history limitations. Amphibi-
ans and reptiles whose populations respond to cli-
matic conditions should not be used as indicator
species, especially in short-duration studies. How-
ever, it may be possible to use these species in long-
term studies where climatic effects can be separated.

Presence and absence can be used to assess
gross species assemblage and habitat component
trends provided sampling techniques are adequate.
For example, absence of a salamander reflects the
lack of required habitat components on one site, and
presence of the species on a similar site reflects the
presence of required habitat components. The key
phrase in the previous sentence is "on a similar site."
To be compared, these sites must have similar topog-
raphy, rainfall, temperature, and vegetation. If these
variables are not similar on each site, the biologist
cannot determine whether lack of habitat compo-
nents on one site limits the salamander's occurrence.
Biologists should select species that rely heavily on
a specific microhabitat. This will increase the accu-
racy of predictions made about microhabitat
conditions.

288

Amphibians and Reptiles

LITERATURE CITED

ANDERSON, K and C.L. SMITH. 1950. An electrical appa-
ratus for herpetological collecting. Copeia 1950:322.

BALGOOYEN, T.G. 1977. Collecting methods for amphibi-
ans and reptiles. U.S. Dep. Inter., Bur. Land Manage.
Tech. Note 299. Denver, CO.

BANTA, B.H. 1957. A simple trap for collecting desert
reptiles. Herpetologica 13:174-176.

BRATTSTROM, B.H. 1965. Body temperatures of reptiles.
Am. Midi. Nat. 73:376-422.

BREEN, J.F. 1949. Reptiles: their habits and care. All-Pets
Mag.

BROWN, L.E. and M.J. LITTLEJOHN. 1972. Male release
call in the Bufo americanns group. Pages 310-323 in
W.F. Blair, ed. Evolution in the Genus Bufo. Univ.
Texas Press, Austin.

BURGE, B.L. 1979. Survey of the present distribution of
the desert tortoise, Gopherus agassizii, in Arizona.
U.S. Dep. Inter., Bur. Land Manage. Contract YA-512-
CT8-108. Denver, CO.

BURY, R.B. 1982. Structure and composition of Mojave
Desert reptile communities determined with a re-
moval method. Herpetological Communities, U.S. Dep.
Inter., Fish and Wildl. Serv. Wildl. Res. Rep. 13.

, H.W. CAMPBELL, and N.J. SCOTT, Jr. 1980. Role

and importance of nongame wildlife. Pages 197-207 in
Trans. 45th North Am. Wildl. Nat. Res. Conf. Washing-
ton, DC.

and R.A. LUCKENBACH. 1977. Censusing desert

tortoise populations using a quadrat and grid location
system. Desert Tortoise Council Symp. Proc.
1977:169-178.

and M.G. RAPHAEL. 1983. Inventory methods for

amphibians and reptiles. Proc. Int. Conf. Renewable
Resour. Inventories for Monitoring Changes and
Trends. Oregon State Univ., Corvallis.

BUSACK, S.D. and R.B. BURY. 1974. Some effects of off-
road vehicles and sheep grazing on lizard populations
in the Mojave Desert. Biol. Consev. 6:179-183.

CAMPBELL, H.W. and S.P. CHRISTMAN. 1982. Field tech-
niques for herpetofaunal community analysis. Herpe-
tological Communities, U.S. Dep. Inter., Fish and
Wildl. Serv. Wildl. Res. Rep. 13.

CARPENTER, C.C. 1953- Trapping technique for aquatic
salamanders. Herpetologica 8(4): 183-

. 1955. Sounding turtles: a field location technique.

Herpetologica 1 1:120.

CLARK, DR., Jr. 1966. A funnel trap for small snakes.
Trans. Kansas Acad. Sci. 69( 1 ):91-95.

CREUSERE, F.M. and W.G. WHITFORD 1982. Use of time
and space by lizards. Herpetological Communities,
U.S. Dep. Inter., Fish and Wildl. Serv. Wildl. Res. Rep.
13.

DARGAN, L.M. and W.H. STICKEL. 1949. An experiment
with snake trapping. Copeia 1949:264-268.

DUNDEE, HA. 1950. An improved method for collecting
living lizards and frogs. Herpetologica 6(3):78-79.

EAKIN, R.M. 1957. Use of copper wire in noosing lizards.
Copeia 1957:148.

EDWARDS, S.R. 1975. Collections of preserved amphibians
and reptiles in the United States. SSAR Misc. Publ.
Herpetological Circular 3-

FERNER, J.W. 1979. A review of marking techniques for
amphibians and reptiles. SSAR. Herpetological Circular
9.

FEUER, R.C. 1980. Underwater traps for aquatic turtles.
Herp. Review 11(4):107-108

FITCH, H.S. 1951. A simplified type of funnel trap for
reptiles. Herpetologica 7:77-80.

FOUQUETTE, M.J. , Jr. 1980. Effect of environmental tem-
peratures on body temperature of aquatic-calling
anurans.J. Herpetol. l4(4):347-352.

FRANKLIN, M.A. 1947. An inexpensive snare for water
snakes. Copeia 1947:143-

GIBBONS, J.W. and R.D. SEMLITSCH. 1981. Terrestrial
drift fences with pitfall traps: an effective technique
for quantitative sampling of animal populations. Brim-
ley ana 7:1-16.

GO IN, C.J. 1942. A method of collecting the vertebrates
associated with water hyacinths. Copeia (3): 183- 184.

GUNNING, G.E. and W.M. LEWIS. 1957. An electrical

shocker for the collection of amphibians and reptiles
in the aquatic environment. Copeia 1957:52.

HALL, R.J. 1967. A simplified live-trap for reptiles. Trans.
Kansas Acad. Sci. 70(3):402-404.

. 1 980. Effects of environmental contaminants on

reptiles: a review. U.S. Dep. Inter., Fish and Wildl.
Serv. Spec. Sci. Rep. 228. Washington, DC.

HANSEN, R.M., M.K. JOHNSON, and T.R. VAN DEVENDER.
1976. Foods of the desert tortoise, Gopherus agassi-
zii, in Arizona and Utah. Herpetologica 32(3):247-
251.

HAWKEN, J.L. 1951. Water system acts as reptile and
amphibian trap. Herpetologica 7:81-83-

HEATWOLE, H„ A. MALDONADO, and J. OJASTI. 1964. A
trapping method for capturing lizards. Herpetologica
20(3):212-213.

HECKEL, D.G. and J. ROUGHGARDEN. 1979. A technique
for estimating the size of lizard populations. Ecology
60(5):966-975.

HORN, H.S. 1966. Measurement of overlap in comparative
ecological studies. Am. Nat. 100:419-424.

IVERSON, J.B. 1979. Another inexpensive turtle trap.
Harp. Review 10(2):55.

JONES, KB. 1981a. Effects of grazing on lizard abundance
and diversity in western Arizona. Southwest Nat.
26(2):107-115.

. 1981b. Distribution, ecology, and habitat manage-
ment of the reptiles and amphibians of the Hualapai-
Aquarius planning area, Mohave and Yavapai Counties,
Arizona. U.S. Dep. Inter., Bur. Land Manage. Tech.
Note 353- Denver, CO.

, LP. KEPNER, and W.G. KEPNER. 1983. Anurans of

Vekol Valley, Central Arizona. Southwest Nat.

28(4):469-470.
KEPHART, D.G. and S.J. ARNOLD. 1982. Garter snake

diets in a fluctuating environment: a seven-year study.

Ecology 63(5): 1232- 1236.
LAGLER, KF. 1943. Methods of collecting freshwater

turtles. Copeia 1943:21-25.
LANNON, JR., Jr. 1962. A different method of catching the

desert lizards, Callisaurus and Uma. Copeia

1962:437-438.
LEE, D.S. 1968. Herpetofauna associated with central Flor-
ida mammals. Herpetologica 24( 1 ):83-84.
LEGLER, J.M. I960. A simple and inexpensive device for

trapping aquatic turtles. Utah Acad. Sci. Proc. 37:63-

66.
LILLYWHITE, H.B. 1982. Tracking as an aid in ecological

studies of snakes. Herpetological Communities, U.S.

Dep. Inter., Fish and Wildl. Serv. Wildl. Res. Rep. 1 3-
LUCKENBACH, R.A. and R.B. BURY. 1983- Effects of off-

Amphibians and Reptiles

289

road vehicles on the biota of the Algodones Dunes,
Imperial County, California. J. Appl. Ecol. 20:265-286.

MANNAN, R.W., ML. MORRISON, and EC. MESLOW.

1984. Comment: The use of guilds in forest bird man-
agement. The Wildl. Soc. Bull. 12(4):426-430.

MARTOF, B.S. 1963. An effective technique for capturing
stream -dwelling organisms. Copeia 1963:439-440.

MICHAUD, T.C. 1964. Vocal variation in two species of
chorus frogs, Pseudacris nigrita and Pseudacris
clarki, in Texas. Evolution 18:498-506.

MILLSAP, B.A. 1981. Distributional status of Falconiformes
in west-central Arizona: with notes on ecology, repro-
ductive success, and management. U.S. Dep. Inter.,
Bur. Land Manage. Tech. Note 355. Denver, CO.

MILLSAP, B.A. and W. HARRISON. 1981. Food and forag-
ing habitats of common black hawks (Buteogallus
anthracinus) in western Arizona. U.S. Dep. Inter.,
Bur. Land Manage. Phoenix District, AZ. Unpubl.

MOULTON, J.M. 1954. Notes on the natural history, col-
lection, and maintenance of the salamander Ambys-
toma maculatum Copeia 1954:64-65.

NEILL, W.T. 1956. Another device for collecting lizards.
Copeia 1956:124-125.

ORTEGA, A., ME. MAURY, and R. BARBAULT. 1982. Spa
tial organization and habitat partitioning in a moun-
tain lizard community of Mexico. Ecol. Gen.
3(3):323-330.

PETOKAS, P.J. and MM. ALEXANDER. 1979. A new trap
for basking turtles. Herp. Review 10(3):90.

PIANKA, E.R. 1966. Convexity, desert lizards, and spatial
heterogeneity. Ecology 47(6):1055-1059.

PEETRUSZKA, R.D. 1981. An evaluation of stomach flush-
ing for desert lizard diet analysis. Southwest. Nat.
26(2):101-105.

PIRNIE, M.D. 1935. Michigan waterfowl management.
Michigan Dep. Conserv., Lansing, ML

PISANI, G.R. 1973. A guide to preservation techniques for
amphibians and reptiles. SSAR Misc. Publ. Herpetolog-
ical Circular 1.

PORZER, L.M. 1982. Movements, behavior, and body

temperatures of the Gila monster {Heloderma suspec-
turn Cope ) in the vicinity of Queen Creek, Arizona.
M.S. Thesis, Arizona State Univ., Tempe.

POUGH, F.H. 1970. A quick method for permanently

marking snakes and turtles. Herpetologica 26:428-430.

ROGERS, T.L. 1939. A lizard live-trap. Copeia 1939:51.

SCHNEIDER, P.B. 1981. A population analysis of the de-
sert tortoise, Gophenis agassizii, in Arizona. U.S. Dep.

Inter., Bur. Land Manage. Phoenix District Office.
Contract AZ-950-CT9-0014.

SCOTT, A.F. and J.L. DOBIE. 1980. An improved design for
a thread-trailing device used to study terrestrial
movements of turtles. Herp. Review 1 1(4): 106- 107.

STEBBINS, R.C. 1966. A field guide to western reptiles
and amphibians. Houghton Mifflin Co. Boston, MA.
27pp.

STICKEL, W.H. 1944. A simple and effective lizard snare.
Copeia 1944:251-252.

STORM, R.M. and R.A. PIMENTEL. 1954. A method of

studying amphibian breeding populations. Herpetolo-
gica 10:161-166.

SVIHLA, A. 1959- A simple method of collecting Ascaphus
truei tadpoles. Copeia 1959:72.

TINKLE, D.W. 1982. Results of experimental density ma-
nipulation in an Arizona lizard community. Ecology
63( 1 ):57-65.

VITT, L.J., AC. HULSE, and R.D. OHMART. 1974. Repro-
duction and ecology of a Colorado River population
of See iopous magister. Herpetologica 30:410-417.

VITT, L.J. and R.D. OHMART. 1978. Herpetofauna of the
lower Colorado River: Davis Dam to the Mexican
border. Proc. West. Found. Vert. Zool. 2(2>35-72.

VOGT, R.C. and R.L HINE. 1982. Evaluation of techniques
for assessment of amphibian and reptile populations
in Wisconsin. Pages 201-217 in Scott, N.J., Jr. ed.
Herpetological Communities. U.S. Dep. Inter., Fish and
Wildl. Serv. Wildl. Res. Rep. 13-

VOGT, W. 1941. A practical lizard trap. Copeia 1941:115.

WHITAKER, AH. 1967. Baiting pitfall traps for small liz-
ards. Herpetologica 23(4):309-310.

WHITFORD, W.G. and F.M. CREUSERE. 1977. Seasonal and
yearly fluctuations in Chihuahuan Desert lizard com-
munities. Herpetologica 33:54-65.

WDEGMAN, D.L., M. HAKKILA, K. WHITMORE, and R.A.
COLE. 1980. Survey of the Sacramento Mountain
salamander (Aneides hardyi) habitat on the Cloud-
croft and Mayhill Districts in the Lincoln National
Forest. U.S. Dep. Agric, For. Serv. Contract OM-40-
7512-0-632.

WILLIAMS, R.D., J.E. GATES, and C.H. HOCUTT. 1981. An
evaluation of known and potential sampling tech-
niques for hellbender, Cryptobranchus alleganiensis.
J. Herpetol. 15(l):23-27.

ZWEIFEL, R.G. and C.H. LOWE. 1966. The ecology of a
population of Xantusia vigilis, the desert night lizard.
Am. Mus. Novit. 2247:1-57.

290

Amphibians and Reptiles

15

SONGBIRDS

Ronald A. Ryder

Department of Fishery & Wildlife Biology
Colorado State University
Fort Collins, CO 80523

Editor's Note: This chapter on Songbirds is the first
of six chapters on birds. The chapter covers the true
songbirds (Order Passeriformes) and also other
orders that require similar techniques for habitat
and population surveys. Techniques for surveying
songbirds are relatively well-described and formal-
ized, and this chapter provides an overview of these
along with references to the appropriate publica-
tion for step by step procedures.

In addition, the chapter covers general sources of
information on birds. In North America, birds are
actively observed, counted, and recorded by mil-
lions of avid, amateur birdwatchers as well as
more limited numbers of professional ornitholo-
gists. The observations of these people, as summa-
rized in publications such as local bird lists and
periodicals such as American Birds, are valuable
sources of background information on distribution
and habitat use, for not only songbirds but for all
bird species. To avoid repetition, these sources of
information are covered in this chapter and not in
the others.

INTRODUCTION

Songbirds will be considered to include not
only birds of the order perching birds (Passeri-
formes), but also cuckoos (Cuculiformes), night-
hawks (Caprimulgiformes), swifts and hummingbirds
(Apodiformes), trogons (Trogoniformes), kingfishers
(Coraciiformes), and woodpeckers (Piciformes),
constituting over 350 species in North America,
north of Mexico. With few exceptions, most are ex-
tremely mobile, and many are migratory, moving
latitudinally or altitudinally between breeding
grounds and wintering areas each year. Most have
relatively high breeding potentials, and their num-
bers and densities may fluctuate greatly from season
to season due to recruitment and climatic changes.
Some even shift their centers of abundance based on
vegetation and prey densities — especially in the
West — depending on seasonal and regional variations
in precipitation.

Although it is possible to obtain comparatively
accurate estimates of absolute densities for a few
endangered forms (such as Kirtland's warbler [Den-
droica kirtlandii] and dusky seaside sparrow [Am-
modramus maritimus nigrescens]), for most species
it is neither physically possible nor economically
reasonable to obtain absolute densities. For most
management purposes, estimates of relative density
are adequate.

The literature concerning inventorying and
monitoring songbirds is vast, if not overwhelming.
Ralph and Scott (1981) edited symposium proceed-
ings which are probably the best available references
on methods for estimating numbers of terrestrial
birds. Franzeb (1977) and Call (1981 ) discussed

Songbirds

291

songbirds in their U.S. Bureau of Land Management
(BLM) Technical Notes on terrestrial wildlife inven-
tories. Davis (1982) and Davis and Winstead (1980)
also provided useful references; they gave specific
inventory methods for many species of songbirds.
The U.S. Forest Service has published proceedings of
several symposiums and workshops relating songbird
numbers to habitats and their management (Smith
1975; DeGraff 1978a, b; DeGraff and Evans 1979;
DeGraff and Tilghman 1980). Verner (1985) criti-
cally evaluated many counting techniques suitable
for songbirds. American Birds (and earlier Audubon
Field Notes and Bird Lore) regularly summarizes
Christmas Bird Counts (CBCs) in its July -August is-
sue, and winter bird-population studies and breeding
bird censuses in its January-February issue. Not only
do the main ornithological journals such as Auk,
Condor, Wilson Bulletin, and the Journal for Field
Ornithology regularly contain songbird-habitat pa-
pers, but a multitude of state and regional bird peri-
odicals are also useful sources of such information.
Many of these periodicals are indexed in Wildlife
Review and Wildlife Abstracts.

This chapter describes habitat characteristics,
such as cliffs and other physical features, as well as
vegetative structure and composition as they relate
to the over 20 families of songbirds found in the
West. Throughout, key references to major avian
groups are given, which can be consulted for specific
habitat requirements (See Table 1).

The major methods or techniques used to moni-
tor or inventory songbirds are described, with criti-
cal evaluations made of their applicability, especially
in the western U.S. (See Table 2).

Ornithologists and land-use managers believed
that numbers of bird species and individuals might
serve as a "litmus test of the environment." It was
thought that a decline in species richness and diver-
sity, as determined by routine monitoring, would
serve as an early warning of environmental degrada-
tion. More specifically, overgrazing, overcutting, or
too heavy recreational use of forest and range eco-
systems would be revealed by changes in bird
populations.

Efforts to use songbird trends as indicators of
habitat quality and habitat changes, however, have
not been as successful as many had hoped. Songbirds
are highly mobile and many individuals may, at most,
spend only 3 to 4 months per year on a given area.
Migratory species are probably more affected by
habitat destruction and pesticide use on their winter-
ing grounds in the Neotropics, where they spend
more of their annual life cycle, than they are by
changes on their breeding grounds in the western
U.S., where they may spend one third or less of the
year.

Lark bunting.

The skill, time, and dedication required to thor-
oughly inventory and monitor bird populations is
probably more than the average land-use manager
can provide. Identification of the many species, by
sight and sound, and determination of sex and age
classes of songbirds to be found in a BLM District re-
quire more knowledge and skill than most typical
"wildlifers" possess or can acquire without consider-
able training and practice.

Also, most songbirds are not primary consumers
but insectivores, at least during the breeding season.
Thus, cause-and-effect relationships with vegetative
changes are not as direct as with resident primary
consumers, such as most rodents.

With these reservations in mind, some monitor-
ing approaches that might be used to ascertain
trends in bird numbers follow.

HABITAT FEATURES CORRELATED WITH
SPECIES GROUPS

Physical Features

Most songbirds are associated with vegetation of
a variety of types, but a few species, especially for
nesting, are largely dependent on cliffs (cliff swal-
lows [Hirundo pyrrhonota] and white-throated
swifts [Aeronautes saxatalis], for example). Others
use earthen banks (bank swallows [Riparia riparia]
and belted kingfishers [Ceryle alcyon]). These special
features are described in Chapters 27 and 31 and
listed in Table 1.

An increasing number of songbirds are adapting
to nesting and roosting in or on structures, such as
buildings and bridges, as well as utilizing artificial
nest boxes and platforms.

292

Songbirds

Table 1. Selected taxa of birds of western North America, structural features used, and some key references
to avian ecology.

Taxa

Common Names

Structural Features Used
(mainly for nesting)

Key References

03
T3
03

E

C

CO

E

■o

c

O

i_
O)

d)

I—
CO
JD

CO
CD

c
>

CO

c
co

to

CO

CO
CO

.Q
o
w

CO
CO

h_

J3

\—
CO

o

CD
CD

i_

0
0

CO

PASSERI-
FORMES

Perching birds

Anderson (1979), Balda
(1975), Call (1981), Ralph &
Scott (1981), Trimble (1975)

Tyrannidae

Tyrant flycatchers

•

•

•

•

•

•

Bent (1942), Forest Service
(1982), Graberetal. (1974),
Maser et al. (1984)

Alaudidae

Larks

•

•

•

•

Beason & Franks (1973,
1974), Boyd (1976), Pickwell
(1931)

Hirundinidae

Swallows

Bent (1942), Graberetal.
(1972), Harrison (1975)

Corvidae

Jays, magpies, & crows

•

•

•

•

Balda & Bateman (1972),
Bent (1946), Goodwin
(1976)

Paridae

Chickadees & titmice

•

•

•

Bailey & Niedrach (1965),
Bent (1946), Dixon (1961)

Sittidae

Nuthatches

•

Bent (1948), Bock & Lep-
thien (1972), Norris (1958)

Certhiidae

Creepers

•

Bent (1948), McClelland &
Frissell (1975)

Troglodytidae

Wrens

•

•

•

•

•

•

Anderson & Anderson
(1973), Armstrong (1955),
Bent (1948)

Cinclidae

Dippers

•

Bakus (1959a, 1959b), Price
&Bock (1983)

Sylviinae

Kinglets & gnatcatchers

•

•

•

Bent (1949), Root (1967),
Trimble (1975)

Turdinae

Solitaires, thrushes, &
allies

Bent (1949), Graber et al.
(1971), Jackman & Scott
(1975), Young (1955), Ze-
leny (1976)

Mimidae

Mockingbirds &
thrushes

Bent (1948), Brazier (1964),
Graber et al. (1970)

Motacillidae

Wagtails & pipits

•

•

Bent (1950), Verbeek (1970)

Bombycillidae

Waxwings

•

•

•

Bent (1950), Lea (1942),
Putnam (1949)

Ptilogonatidae

Silky flycatchers

•

•

•

Bent (1950), Verner & Boss
(1980)

Laniidae

Shrikes

•

•

•

•

Bystrak (1983), Graber et al.
(1973), Miller (1931), Porter
et al. (1975)

Sturnidae

Starlings

DeHaven & DeHaven
(1973), Feare (1984), Kessel
(1957)

Songbirds

293

Table 1. Selected taxa of birds of western North America, structural features used, and some key references
to avian ecology (concluded).

Taxa

Common Names

Structural Features Used
(mainly for nesting)

Key References

0
-o

03

E

c

03

E

■o

c

o

I—

en
0
co

CO

CD

c
>

CO

k_

"D

c

CO
C/3

CO
CO

co

k_

o

CO

co

O)

X3

k_
JZ
CO

o

CD

0

I—
CO

0
0

sz

CO
03

Vireonidae

Vireos

•

•

•

Barlow (1962), Bent (1950),
Verner & Boss (1980), Win-
ternitz(1976)

Parulinae

Wood-warblers

•

•

•

Bent (1953), Graber et al.
(1983), Griscom & Sprunt
(1957), Harrison (1984)

Thraupinae

Tanagers

•

•

Bent (1958), Verner & Boss
(1980)

Cardinalinae

Cardinals, grosbeaks,
& allies

Austin (1968), Verner &
Boss (1980)

Emberizinae

Emberizine finches

Austin (1968), Forest Ser-
vice (1982)

Icterinae

Blackbirds & allies

Bent (1958), Nero (1984),
Orians (1980)

Fringillinae

Fringilline finches

Austin (1968), Maser et al.
(1984), Trimble (1975)

Carduelinae

Cardueline finches

Austin (1968), Trimble
(1975), Verner & Boss
(1980)

Passeridae

Old world sparrows

•

•

Packard (1966), Summers-
Smith (1963, 1967)

CUCULIFORMES

Cuckoos, roadrunners,
& anis

•

•

•

•

Bent (1940), Ohmart (1973),
Preble (1957)

Cuculidae

CAPRIMULGI-
FORMES

Goatsuckers

•

Bent (1940), Caccamise
(1974), Verner & Boss
(1980)

Caprimulgidae

APODIFORMES

Swifts

•

•

Bent (1940), Knorr(1961)

Apodidae

Trochilidae

Hummingbirds

•

•

•

•

•

Calder (1973), Johnsgard
(1983)

TROGONI-
FORMES

Trogons

•

•

•

Bent (1940), Forest Service
(1975), Phillips et al. (1964)

Trogonidae

CORACII-
FORMES

Kingfishers

•

Bent (1940), Cornwell
(1963), Eipper(1956)

Alacedinidae

PICIFORMES

Woodpeckers and
allies

•

•

•

Bent (1939), Bock (1970),
Jackson & Scott (1975)

Picidae

294

Songbirds

Table 2. Some methods for monitoring songbirds — advantages and disadvantages. (See text for more details.)

Method

Season of Use

Advantages

Disadvantages

Checklist

All

Quick, relatively cheap to
compile. Use literature and
volunteers. Can visually
portray arrival and depar-
ture dates and relative
abundance.

Mainly qualitative, only
relative abundance and
gross habitat preferences.

Atlas

Mainly stress-breeding
season (can show other
seasons also)

Gives "big picture." Fairly
understandable to the pub-
lic.

Usually rather large scale
(1° blocks of latitude and
longitude). Requires con-
siderable organization and
cooperation of many ob-
servers.

Christmas Bird
Count (CBC)

Winter

Many areas covered in
past. Large sample size
nationally.

Only one coverage/area/
year. Area of special inter-
est may not have been
covered. Urban areas
stressed. Only indexes, no
absolute densities.

Winter bird-popu-
lation studies

Winter

More coverages/season
than CBC. Replications,
hence greater possibility of
statistical analysis.

Smaller area and number
of areas covered than
CBCs. More time and effort
required.

Spot-mapping

Mainly breeding season
(some winter territorial
mapping possible)

Usually considered to give
good estimates of breeding
densities.

Requires repeated cover-
ages by skilled observer.
Need good vegetative map
and vegetative measure-
ments to be most useful.

Transects (mainly
walked)

Any season

Quicker, cheaper to con-
duct than spot-mapping.
More random samplings
possible.

Accuracy highly variable
species to species, habitat
to habitat. Less accurate
in nonbreeding season.

North American
Breeding Bird
Survey (BBS)

Breeding season (has been
used for other seasons)

Large sample size nation-
ally. Relatively cheap and
quick to conduct.

Only single coverage of a
given route each year.
Habitat data not regularly
gathered. Only indexes
of abundance. Area must
be accessible by road.

Point count (IPA)

Breeding season

Good statistical reliance
reported for Europe (not
well tested in U.S.).

Only two coverages/plot/
season, normally. Consid-
erable time (20 min) per
stop. Only frequency data.

Density-fre-
quency relation-
ship (EFP)

Breeding

As above. Said to give
density correlations.

More time-consuming than
point count.

Maryland winter
bird survey

Winter

Can be used where roads
are lacking. Comparatively
quick and cheap.

Walked (unlike roadside
count), requires more time
and effort than BBS. Not
well tested in western U.S.

Variable-circular
plots

All seasons (but mainly
used in breeding season)

Can cover large geographi-
cal areas, compare differ-
ent habitats, and work in
rugged and remote terrain.

Requires considerable
training in estimating or
time in measuring dis-
tances (aural as well as
visual distances).

Songbirds

295

Table 2. Some methods for monitoring songbirds
(concluded).

-advantages and disadvantages. (See text for more details.)

Method

Season of Use

Advantages

Disadvantages

Playback of tape
recordings

Mainly breeding season

Quick way to get male
responses. Speeds up
spot-mapping.

Only certain species re-
spond. Can alter behavior
(interfere with nesting suc-
cess?).

Mark and recap-
ture

Any season

Can use similar population
estimates as used with
game species.

Not highly efficient. Consid-
erable time and effort re-
quired. Special permits
required. Many biases in-
volved (trap-happy vs. trap-
shy individuals).

Nest monitoring

Breeding season

Relates to actual nesting
efforts.

Very time-consuming.
Highly variable with species
and habitats involved. Can
increase predation and
desertion.

Vegetation

Structure. Vegetative structure and habitat
configuration (physiognomy) are generally
considered more important in distribution and
abundance of birds than plant species composition
(floristics — Anderson and Shugart 1974; Hilden 1965;
James 1971; Wiens 1969). However, some believe
floristics may be more important within gross habitat
types (Rotenberry, pers. commun.). Rotenberry's
analysis of International Biological Program data
from grasslands disclosed that 55% of the variation
in bird community composition was associated with
floristic variation and only 35% was associated with
physiognomy.

Others have shown patterns of avian distribution
strongly correlated with vegetational structure
(MacArthur and MacArthur 1961; Willson 1974 et
al. ). Structural features commonly measured include
density, mean distance to nearest plant neighbors,
height, diameter at breast height (DBH) for trees,
relative dominance, and canopy cover. Readers of
American Birds are urged to follow "A Quantitative
Method of Habitat Description" by James and Shu-
gart (1970). Numerous review papers on bird-habitat
structure studies are included in Abbott ( 1976),
Anderson ( 1979), Anderson and Shugart (1974),
Balda (1975), Cody (1981), Hilden ( 1965), Holmes
et al. ( 1979), Holmes and Robinson (1981 ), James
(1971), Karr (1980), Meents et al. (1983), Rice et
al. (1983), Rotenberry and Wiens (1980a, b), Roth
(1981), Taylor and Littlefield (1984), Wiens (1973,
1983), and Willson (1974). Capen (1981 ) discusses
the use of multivariate statistics in describing avian
habitats. Principal component analysis and discrimi-
nate function analysis are frequently used to examine
avian habitat selection (Sedgwick 1981).

DENSITY

HEIGHT

3 °®

Mr

9 wL \,

n

DBH

CANOPY COVER

Species Composition. Although some researchers
have recently shown floristics, or species of plants
making up songbird habitat, to be important
(Rotenberry, pers. commun.), most past studies have
placed more importance on vegetative structure.
Notable exceptions are the close dependence of
Kirtland's warbler on young-growth jack pine (Pinus
banksiana ) and the red-cockaded woodpecker
(Picoides borealis) on southern pines (P. palustris,
P. elliottii, P. taeda, and P. echinata ) that are
infected with red heart (Formes pint).

296

Songbirds

STRUCTURE

■

FLORISTICS

Canopy Cover by Type

38% Trees
32% Shrubs
20% Grass

5% Forbs

5% Bare ground

Species Composition

5% Ponderosa pine

10% True mountain
mahogany

30% Arizona fescue

30% Bluebunch wheatgrass

25% Other

POPULATION MEASUREMENT TECHNIQUES

Presence

Sometimes mere lists of species or maps show-
ing species distribution are adequate for making
management decisions (for example, the location of
red-cockaded woodpecker colonies and decisions
to log or not log). Most state ornithological societies
have official state lists of birds and often have "offi-
cial records" committees and procedures for decid-
ing the validity of records. Many local chapters of
the National Audubon Society have checklists of
birds for local areas, as do various land use agencies
(e.g., BLM, U.S. Forest Service, U.S. National Park
Service, U.S. Fish and Wildlife Service, and state con-
servation agencies).

Breeding Bird Atlases

Breeding bird "atlasing" has become a popular
pastime among bird observers in Europe during the
past decade. Atlases showing breeding distribution
maps for each species have recently been published
for the British Isles, France, and Denmark. Similar
atlases are in progress in 16 other European nations
(Sharrock 1975), Australia (Serventy 1980), and
New Zealand (Gibb 1980).

The original purpose of preparing a breeding
bird atlas was to correlate bird distribution with that
of plants as shown in the Atlas of British Flora (Per-
ring and Walters 1962). In a government-sponsored
program carried out through the British Trust for

Western tanager.

Ornithology and the Irish Wildbird Conservancy, ob-
servers visited every one of the 3,862 10-km squares
( 100 km" [38.6.mi.~] each) of land throughout the
British Isles during a 5-year period and reported the
presence or absence of each bird species (Sharrock
1976). Twelve transparent overlays, ordered sepa-
rately, facilitate correlation of bird distribution with
selected environmental factors. Sampling blocks have
been different sizes in other countries, depending
on the size of the area to be sampled and the stand-
ard maps available. In France, the sampling unit was
20 by 27 km ( 12.4 by 16.8 mi.; Yeatman 1976) and
in Denmark it was 5-km (3-1 -mi.) square (Dybbro
1976). Several countries modified the method to
include some indication of abundance rather than
merely presence or absence.

No large-scale atlas has been attempted in the
U.S. because the U.S. Fish and Wildlife Service's
Breeding Bird Survey (BBS) provides an annual sam-
ple of changing abundance of each species and also
gives a density of coverage roughly comparable with
that of the projected European atlas of the 1980s.
Nevertheless, atlas studies have been initiated in
several states (Laughlin 1982; Laughlin et al. 1982).
These state atlases will provide information on pres-
ence or absence of the various species in many for-
ested areas. In the Rocky Mountains, they are often
termed "Latilong" studies, a word originally coined
by Skaar ( 1969) to denote 1° of latitude and longi-
tude blocks (Figure 1 ).

Although western state atlases use grids larger
than the British 10-km (6.1 -mi.) model, in most east-
ern states a 5-km (3.1 -mi.) grid (six blocks on a
7 '/j-minute topographic map) is used. The Maryland
Ornithological Society, however, uses 2 '/2-km ( 1
l/2-mi.) "quarter blocks" (about 600 ha [1,500 a.]).
The quarter blocks make it possible to pinpoint the
location of rare species and others of special interest
and are much better for outlining areas where a

Songbirds

297

particular species is not present (Klimkiewicz and
Robbins 1974). This accuracy is particularly impor-
tant where commercial or residential communities
are expanding or where habitats are being lost to
other types of development. In areas where large for-
ests are being destroyed by changes in land use,
quarter-block atlas data have been of immense value
in showing the degree various breeding species dis-
appear when forested areas are fragmented into
smaller tracts. Raynor (1983) proposed a method for
evaluating atlas coverages.

Christmas Bird Count (CBC)

The CBC is the best known and probably most
used source of information on geographical distribu-
tion of nongame birds in winter (Table 3). Started
in 1900, the counts now involve over 1,300 circles,
each 24 km (15 mi. ) in diameter, which are covered
by varying numbers of birders in an 8-hour period,
sometime between December 20 and January 2 each
year. Dawn-to-dusk (or longer) counts are preferred.

Cedar Waxwing

Breedinq

Miqr.

Winter

W b b R

b N b Habitat:

R,Sb
U

R.Ag.PJ
FC

U,R,Ag
Irr

W R N

W b

B b N

W N

R W

N W W Abundance:

Figure 1. Example of bird distribution by Latilong blocks in Colorado.
298 Songbirds

Counts less than 8 hours long, except in arctic areas
or at sea, are not acceptable. Results are summarized
on standardized forms which include entries for
weather, habitat description, and quality of the food
supply. Summaries are published annually in Ameri-
can Birds (Arbib 1978) (Figure 2). In 1980, 1,320
counts were included, covering a vast area from
Newfoundland to Alaska, south to Panama and Ha-
waii. In 1982, 36 countries and 6 continents were
involved (Arbib 1982). Popular evaluations of the
CBC have been written by Robbins ( 1966), Bock
(1979), and Wing and Jenks (1939).

Individual birds reported on CBCs in 1979
ranged from 92 at Churchill, Manitoba, to
22,352,044 at Pine Prairie, Louisiana. Sixty-five
counts had 1 50 or more species. The most meaning-
ful data are expressed in number of birds observed
per party-hour. Other bases, such as birds per mile
and birds per observer, have proven unreliable
(Bock 1979).

Winter Bird Population Study

In 1948, the National Audubon Society inaugu-
rated an annual Winter Bird Population Study which,
like the CBC, is published in American Birds. The
purpose of this type of count is to estimate the aver-
age number of birds using a particular habitat. Many
observers use the same plot in which they have con-
ducted a breeding bird census by spot-mapping.
Most plots range from 6 to 20 ha (14.8 to 49.4 a.).
Plots are visited 6 to 10 times in midwinter, and the
totals for each species are averaged (Kolb 1965).
Results are expressed in terms of birds per kilometer
squared [ 1 km ] and birds per 40.5 ha (100 a. ). In
1979-80, 92 counts were reported in American
Birds (Cink and Boyd 1981 ), only a fraction of the
1,300 Christmas counts reported that winter. Several
visits per winter are necessary because populations,
and even the species present in a given plot, vary
from day to day and from morning to afternoon.
Finches (Carpodacus sp. ), robins (Turdus sp. ), and
waxwings {Bombycilla sp. ), for example, often range
over many square kilometers in the course of 2 or 3
days. Other species, in mixed flocks, may range in
and out of a study plot. Such flocks tend, especially
on cold days, to favor sunny exposures, concentrate
in better feeding sites, and avoid windy areas
(Robbins 1978).

Difficulties notwithstanding, the Winter Bird
Population Study does enable one to compare popu-
lations of different habitats and, to a lesser degree,
to follow population trends over a period of years.
Webster ( 1966) analyzed the results of 248 winter
studies in forest habitats and 25 studies in grasslands.
By plotting species richness against population den-
sity, he found that southern pine forests mixed with
oaks or gums tended to have a higher species rich-
ness than other eastern forests. Webster, however,

Table 3. Key references to uses of Christmas Bird
Counts in North America.

Topic Considered

References

Evaluation of
method

Arbib (1967), Bock & Lep-
thien (1975b), Burtt & Burtt
(1982), Bystrak(1974),
Preston (1958), Tramer
(1974)

General application

Bock (1979, 1984)

Nuthatches

Bock & Lepthien (1972,
1974, 1976d)

Woodpeckers

Bock & Lepthien (1975a)

Finches

Bock & Lepthien (1976a,
1976b)

Various species

Bock & Lepthien (1972,
1976b, 1976c), Bocketal.
(1977), Bock& Root
(1981), Bock& Smith
(1971)

1176. Aspen, Colo. 39°15'N 106°54'W,
center n.e. corner Sec. 29, R85W, T9S, as de-
scribed 1982; elevation 7000 to 8200 ft; habi-
tat coverage: residential 35%, oak/
serviceberry/chokecherry 21%, willow/
cottonwood riverbottom 15%, spruce fir 12%,
aspen grove 9%, sage/ rabbi tbrush 3%, open
meadow 3%, pinyon juniper 2%. — Dec. 19; 8
a.m. to 5 p.m. Clear. Temp. 9° to 46°F. Wind
SE-W, 0-5 m.p.h. Snow cover 0 to 24 in. Wa-
ter mostly frozen. Wild food crop excellent.
Nineteen observers, 11-14 in 8 parties, 5-8 at
feeders. Total party-hours, 37 (18 on foot, 5
by car, 14 on skis) plus 19 hours at feeders;
total party-miles, 119.5 (13.5 on foot, 93 by
car, 13 on skis).

• Canada Goose 5 1 ; Mallard 304; Am. Wi-
geon 1; Wood Duck 1; Goshawk 1; Cooper's
Hawk 2; Great Horned Owl 1; Com. (Red-
sh.) Flicker 10; Hairy Woodpecker 5; Downy
Woodpecker 4; Gray Jay 4; Steller's Jay 47;
Scrub Jay 12; Black-billed Magpie 121; Com.
Raven 5; Black-capped Chickadee 116;
Mountain Chickadee 27; White-breasted
Nuthatch 7; Brown Creeper 1; Dipper 7; Am.
Robin 121; Townsend's Solitaire 3; Cedar
Waxwing 67; Starling 95; House Sparrow 356;
Red-winged Blackbird 54; Evening Grosbeak
46; Cassin's Finch 37; Pine Grosbeak 6;
(Gray-crowned) Rosy Finch 24; (Brown-
capped) Rosy Finch 4; Pine Siskin 1; Rufous-
sided Towhee 1; Dark-eyed (Slate-col.) Junco
23; Dark-eyed (Oregon) Junco 36; Dark-eyed
(Gray-headed) Junco 13.

Total, 33 species (3 additional races); 1615
individuals. (In count area count week but not
seen count day: Bald Eagle ad., Com. Snipe,

Blue Jay 9
Crow 1; Bh
tain Chicka
in 19; Wat
1761; Hous
128; Red-u
Blackbird 2
Cowbird 1
170; Pine S
eyed (Slate
gon) Junco
Junco 1; T
Sparrow 25
Total, 54
individuals
Barry Knap
W. 9th PI
Ward, Judy

1178. I

104°38*W, <
described 1
habitat covi
19; 7:30 a.r
to 60°F. Wii
to 6 in. Wat
fair. Fifteer
party-hours,
party-miles,
• Coopei
Rough-legg<
Marsh Hawl
2; Com. Sni[
Owl 1; Coi
Woodpeckei
Horned Lar
3; Black-bill
Com. Crow

Figure 2. Example of winter bird population study
taken from American Birds, January/February
1984.

Songbirds

299

did not list actual species involved but merely cited
17 years of data published originally in Audubon
Field Notes.

This method can be used to obtain indexes of
winter use. However, because winter bird popula-
tions vary enormously from year to year in any loca-
tion, a minimum of 2 years is required for a
meaningful study. In a critique of this method, Rob-
bins (1972) showed that in forest habitats six traps
were sufficient to obtain a stable minimum estimate
of the total wintering bird population, but that at
least 8 to 10 trips were required to obtain such esti-
mates for individual species. Brewer ( 1972) and
Engstrom and James (1981) also evaluated the
technique.

Spot-Mapping Census

The spot-mapping technique, also called the plot
census or simply mapping census, was first used in
North America by Williams (1936) and in Sweden
by Enemar ( 1959). Basically, this count involves 8 to
10 census trips through an area of specified size and
preferably uniform habitat, which has been surveyed
and mapped with a grid system (Figure 3). On each
visit, the position of each bird seen or heard is re-
corded on the plot map. Kendeigh ( 1944), Lack
(1937), and Udvardy (1957) gave good historical
accounts, described the method in detail, and in-
cluded comprehensive bibliographies. An important
feature of spot-mapping is to designate, with appro-
priate symbols, those individual birds that are heard
singing at the same time (simultaneous registra-
tions). These, in conjunction with clusters of single
registrations, make it possible to outline the approxi-
mate territorial limits of each male bird and make
reasonable estimates of the total number of territo-
rial males of each species present in the area.

Spot-mapping has been widely adopted in
Europe and North America. Procedures have been
standardized by the International Bird Census Com-
mittee (1969) so that results obtained in different
countries can be compared. Spot-mapping has been
widely used in England since 1962 to monitor bird
population changes (Batten and Marchant 1977).
The British Trust for Ornithology refers to such
counts as the Common Bird Census (their CBS, not
to be confused with CBC, Christmas Bird Count).
A quantitative description of habitat (James and Shu-
gart 1970) is now a standard feature of many of the
breeding bird censuses in forest habitats published in
American Birds.

Results from over 200 spot-mapping censuses
were published in American Birds in 1981 (Van
Velzcn 1981 ). Many private consultants and govern-
ment biologists regularly use the spot-mapping tech-
nique, although their results are not published in
American Birds. Dickson ( 1978) compared the tech-

nique with winter bird-population count techniques
(also referred to as "mean detections per count")
and Palmgren's (1930) "summation method" and
concluded that spot-mapping gave the highest popu-
lation estimates.

Robbins (1978) considered spot-mapping to be
the most accurate of the various bird census meth-
ods because it —

( 1 ) gives the greatest opportunity to record all
species breeding in the area,

(2) most closely approximates the absolute num-
ber of breeding pairs,

( 3 ) is more accurate in estimating whether the
birds recorded are inside or outside plot
boundaries, and

(4) involves less observer bias.

The chief disadvantage of the technique is the
amount of time required to set up the plot and con-
duct a minimum of eight census trips.

The results of 228 breeding bird censuses con-
ducted in Canada were summarized in three catalogs
by Erskine (1971, 1972, 1976). Censuses published
in American Birds since 1937 are on magnetic tape
at the Migratory Bird and Habitat Research Labora-
tory, U.S. Fish and Wildlife Service. They provide
a valuable resource for comparing habitat, bird spe-
cies, or other variables included in the computer
record. The files contain information on 1,000 plots
(Robbins 1978).

Ladder-backed woodpecker.

300

Songbirds

1A

1B

1C

1D

IE

1F

1G

1H

?A

7"

X

t

3A

yV

X

t

X

\ /

\ /
\ /

i /
I /
1/

X *

**

XI

4A

\ X

•

•
t-

\ ^

X^

5A

>w *

1
\ 'I

— *y / x

«"

X ^

6A

/ *

X

*

' ^\

7A

v *

*

•*

•

8A

• ^^^

Figure 3. Spot-map for the yellow-rumped warbler (Dendroica coronata) in a mixed-coniferous forest.
(Blocks are 50 m by 50 m producing a grid plot of 12.25 ha. Total breeding bird density is 3 territories
x 2 = 6 breeding yellow-rumped warblers per 12.25 ha or a density of 19.6/40 ha [19.6/ 100a.]).

Songbirds

301

TRANSECT LINE

• = detected
o = undetected

TRANSECT LINE

Figure 4. In line transect sampling, the observer
records either the right angle distance (x), or
both the flushing distance (r) and the sighting
angle (O) for each bird detected. Bird density is
estimated mathematically using the distribution
of distances (x or r).

Figure 5. In belt transect sampling, all birds are
assumed to be detected within a specified, fixed
distance (w) from the transect line of length L.
Birds outside the belt are not recorded. Bird
density = n/2Lw, where n is the number of
birds recorded.

Transects

Transect methods (Figures 4 and 5) involve
counting birds on one or both sides of a line
through one or more habitats. Usually either the
width of the transect is defined or the distance to
each bird encountered is estimated. The transect
method was first used extensively in the U.S. in
1906-09 by S.A. Forbes and A.O. Gross (Graber and
Graber 1963) and in Finland in 1941-56 by Merikal-
lio (1958). In both studies, the transect results were
used to estimate total populations by habitat for an
entire state or nation. The Forbes and Gross study
was later repeated by Graber and Graber ( 1963) to
show bird population changes over a 50-year period
in Illinois.

One advantage of the transect method is that a
relatively large area can be sampled in a short time.
Flack (1976), for example, used 45-m (150-ft)-wide
and 1,097-m (l,200-yd)-long transects to compare
breeding bird populations in 4 1 aspen forests in
9 western states and provinces in 1966-69.

The transect method may be used throughout
the year, but the results are less accurate outside the
breeding season. However, the chief disadvantage of
the transect method is that a single coverage of a
transect does not permit a good estimate of the
number of birds missed. Anderson and Pospahala
(1970), using data from 2,574 km (1,600 mi.) of
transects, generated a curvilinear (quadratic) equa-

tion to show the fraction of waterfowl nests missed
at various distances from the center line of the tran-
sect. However, they pointed out, that to adequately
correct for the number of fixed objects missed, one
needs a large sample and must also assume that all
the objects closest to the center of the transect are
detected.

Emlen (1971) also considered how the error in
the transect count may be reduced by first estimat-
ing the lateral distance to each bird encountered and
then deriving a coefficient of detectability. He as-
sumed that no bird close to the observer goes unde-
tected. In actual practice, a large number of birds
within a few meters of the transect line may be un-
detected (Jarvinen and Vaisanen 1976), especially in
a mature forest habitat. Even during the height of the
breeding season, there are enormous differences in
the singing behavior and conspicuousness of birds. A
noisy, active species such as the tufted titmouse
(Parus bicolor) may be recorded on 68% of the vis-
its if within 50 m (54.7 yd) of the observer. Species
such as the ruby-throated hummingbird (Archilochus
colubris), worm-eating warbler (Helmitheros vermi-
vorus), and even the American redstart (Setophaga
ruticilla ) may be recorded only 36% to 39% of the
time (Stewart et al. 1952). Emlen (1977) estimated
the number of unrecorded males of common species
by running each transect five times, plotting all sing-
ing birds on maps, outlining the territory of each,
and determining what European workers call the "ef-
fectivity" of a single trip for each species. He then

302

Songbirds

used the computed effectivity for correcting his
breeding season transect results.

The transect method is effective for comparing
the abundance of a given species in two or more
plots of similar habitat. Also, unless visibility is
strongly influenced by the structure of the habitat,
the transect method may be used to compare abun-
dance of a given species from one habitat to another
(Robbins 1978). It is not, however, a desirable
method for comparing abundance of two species
that may not be equally conspicuous, unless appro-
priate corrections are made, species by species.
These corrections can be made by taking a series of
transect counts through plots where the population
has been estimated by other methods (Ferry and
Frochot 1970; Enemar and Sjostrand 1970).

Although most transect workers record birds
per unit of distance or area — kilometer (or mile) or
square kilometer (or 100 acres) — a few have pre-
ferred to use units of time, such as birds per 10
hours (Colquhoun 1940). Conner and Dickson
(1980) discussed sampling design and statistical
treatment of data gathered by transect censusing.

North American Breeding Bird Survey
(BBS)

The North American Breeding Bird Survey
(BBS) (Figure 6) was developed by the U.S. Fish and

Figure 6. Breeding Bird Survey summaries are
published in American Birds.

Wildlife Service to monitor bird population changes
in North America over a period of years (Bystrak and
Robbins 1978; Erskine 1978; Robbins and Van Vel-
zen 1967, 1969, 1974). Each survey route is a series
of 50, 3-minute point counts at 800-m (V^-mi.) inter-
vals along a 39.4-km (24 V^-mi.) roadside transect
selected randomly. Coverage extends from half an
hour before sunrise to about 4 hours after sunrise,
embracing the period of greatest bird activity. At
each of the 50 points (counts), all birds heard and
all birds seen within 400 m ( XA mi. ) of the counting
position are tallied. The BBS, which now embraces
the populated areas of Canada and all of the U.S.
except Hawaii, provides an annual sample from
1,700 or more roadside transects (Robbins 1978).

The BBS results are used primarily for statistical
analysis of population changes over the years and
for mapping relative breeding densities throughout
the North American range of a species. A 1 5-year
summary of the BBS results is presented by Robbins
etal. (1986).

For intensive local studies, BBS routes can be
laid out in a non-random way so that all or most
secondary roads within the area of interest are in-
cluded in a sample. The route is termed "mini-route"
because 25 instead of 50 stops are made. By using a

Songbirds

303

shorter route, observers are able to cover the route
before working hours in the morning. Covering each
route twice (once in each direction) and combining
the results of the two counts eliminated most of
the difference in activity resulting from time of day.
D. Bystrak and others (Klimkiewicz and Solem 1974)
used mini-routes to map relative abundance of
breeding birds throughout two Maryland counties as
part of a Breeding Bird Atlas program for these
counties. Although the mini-route technique was
designed for roadside use, with slight modifications,
it could be used to cover forested areas by horse-
back or off-road vehicles. Such application would
make it possible to map distribution of breeding
birds over a wide area in a relatively short time. Dif-
ferences in bird populations could then be corre-
lated with differences in vegetation obtained from
aerial surveys or by ground survey methods.

The Indexes Ponctuels d'Abondance (IPA)
or Point Count Method

The point count or IPA method was developed
in France by Ferry and Frochot ( 1970) as a means of
obtaining indexes of abundance for comparing bird
populations of different habitats (or of the same hab-
itat in different locations) during the breeding sea-
son (Robbins 1978).

The IPA counts by the French ornithologists
consist of establishing a network of points regularly
distributed through the habitat to be studied. An
observer then stands at each designated spot for 20
minutes in the early morning in good weather and
notes all birds seen or heard. Each spot is censused
twice during the breeding season. The higher of
the two counts of pair numbers is used as an index
of abundance for each species. Each singing male,
occupied nest, or family of birds out of the nest
counts as one pair, whereas a bird merely seen or
heard calling counts as half a pair. The efficiency of a
20-minute stop seemed satisfactory to the French
investigators because during the last 5 of the 20
minutes only 3% more species and 9% more individ-
uals were recorded in forest habitat.

In Denmark, Jorgensen ( 1974) conducted 81
IPA censuses on eight mornings from mid-May to
mid-June. The 1 3 V2 hours of effective field work
was about 50% less than would have been needed
for covering one census plot by the spot-mapping
method. Jorgensen compared the density of each
species in different habitats, using the Mann-Whitney
U-test. He concluded that the IPA method was well-
suited to a study of forest succession in which statis-
tical comparisons are necessary or desirable. He
summarized habitat use (based on 15 to 18 counts
in each habitat ) in terms of a list of dominant spe-

cies, each making up 5% or more of the registra-
tions, and subdominants (2% to 5% ). Then, using
only the dominant species, he computed similarity
indexes for the various habitats using the formula:

S =

2c

(a + b)

Where.

S = the index.

a and b = the numbers of species in each sample.

c = the number of species common to the
two samples.

Density-Frequency Relationship. Blondel (1975)
introduced a further modification of the IPA method,
Echantillonnage Frequentiel Progressif (EFP). The
EFP method uses the presence or absence of a
species on each of the 20-minute IPA counts to
determine frequency of that species in each plot.
Comparison of the IPA with EFP figures allows one
to determine for each species the relationship
between its density and frequency. The frequency of
a species is closely correlated with the logarithm of
its density; the lower the frequency, the better the
correlation. Rotenberry and Wiens (1976) found
a similar correlation between density and frequency,
using BBS roadside transect data.

Blondel claimed that the EFP method, which is
highly standardized, is very useful for a rigorous
statistical interpretation of data. He used the EFP
method to calculate ecological profiles and niche
breadth for each species and analyzed the structure
of bird communities according to the structure of
the vegetation. For each community, he determined
the species richness, the species diversity index (H'),
the equitability (J'), and the level of fit to Galton's
log-normal model. He also discussed the influence of
reforestation on bird communities. Blondel con-
cluded that the EFP method is "very well adapted to
solve problems of theoretical and applied ecology
at the community level, and can be used fruitfully
for environmental monitoring" (Robbins 1978). Un-
fortunately, the method has not been tested in the
West but seems worthy of consideration.

Comparison with Spot-Mapping. In comparing
point counts (IPA) with mapping census in the
Bialowieza Forest in Poland, Tomialojc et al. (1978)
found that point counts overestimated the
population when the density was low and
underestimated it when the density was high. IPA
counts also require better trained observers and
involve more problems in the separation of migrants
or other non-breeding birds from breeding
individuals than mapping censuses.

304

Songbirds

Transect and Point Count Combination. Bond
(1957) used a method that was essentially a
combination of the transect and point count
methods to compare bird populations in 64 upland
hardwood stands in Wisconsin. After entering a
woodland, he walked about 50 m (55 yds) along a
transect line. At this point, he stopped for 5 minutes
and counted all birds seen or heard ahead of him. He
then walked ahead slowly for 5 minutes, averaging
150 to 175 m (164 to 191 yds). He repeated this
procedure until he had made five 10-minute counts
from each forest interior. Two early morning visits
were made to each woodlot, and the highest count
for each species was used. In these counts, he
detected 76% , 78% , and 70% as many pairs as he
found by spot-mapping censuses in three of the same
woodlands. Palmgren (1930) and Kendeigh (1944)
found that 81% and 63% of their birds in spot-
mapping censuses were detected on the first two
visits.

Winter Transects

The Maryland Ornithological Society (MOS)
devised a winter bird survey in central Maryland,
using walked transects (Robbins 1970). Preliminary
tests showed that the winter bird survey method was
not practical because of heavy traffic on many roads
during the first few hours after sunrise. Also, because
the birds were not singing, those that could not be
seen from the roadside could not be detected. The
walked transects, away from roads, would reduce the
problem caused by traffic noises.

The winter bird survey method involves tran-
sects of 8 km (5 mi. ) that are covered on foot dur-
ing the first 4 hours after sunrise. One route was
established at the center of each 7 '/2-minute U.S.
Geological Survey map of central Maryland, repre-
senting an 1 1- by 14-km (6.8- by 8.7-mi. ) grid. An ef-
fort was made to lay out each route in the form of a
square, 2 km (1.2 mi.) to a side. Many routes could
not be formed into a square because of streams,
ponds, buildings, high fences, or other obstacles; de-
spite changes in shape, the total length of 8 km ( 5
mi. ) was maintained. By timing their walking speed
for the first quarter of the route, observers were able
to return to the starting point within a few minutes
of the prescribed 4-hour period. Separate counts
were kept for each hour, and birds identified at a dis-
tance greater than 402 m ( 'A mi.) were recorded in
a separate column on the form (Robbins 1978). The
winter survey should be applicable for key wintering
areas in the Southwest.

Variable-Circular Plots

in studies of endangered species in dense Hawaiian
forests (Ramsey and Scott 1979; Scott and Ramsey
1981; Scott et al. 1981), oak-pine woodlands (Verner
and Ritter 1985), ponderosa pine forest (Szaro and
Balda 1982), and desert habitats (Szaro and Jakle
1982). De Sante (1981) found that the variable cir-
cular plot underestimated the densities of breeding
birds in California coastal scrub by 2% to 70% (de-
pending on the species of bird ) when compared
with estimates derived from a combination of color
banding, spot-mapping, and nest monitoring. De
Sante (1985) found similar discrepancies in lodge-
pole (Pinus contorta) populations of birds in the
Sierras.

For many species, tape recordings can be played
back to induce songs from silent territorial males
(Down 1970). The technique may increase census
accuracy (especially for species with low song activ-
ity) with little expenditure of time and can be used
to determine territorial boundaries. Repeated use
of tape recordings during the breeding season, how-
ever, can bias results because birds may alter their
habits or their territorial boundaries if they believe a
competing member of the same species has a terri-
tory nearby (Robbins 1978; McNicholl 1981). Re-
cordings have proven useful in censusing various
songbirds (Falls 1981; Johnson et al. 1981; Marion et
al. 1981). Although recordings are probably most
useful for non-songbirds such as rails and owls, they
have been used successfully to monitor Kirtland
and golden-cheeked warblers as well as trogons.

OBSERVATION POINT

• BIRDS PRESENT
' BIRDS OBSERVED

Reynolds et al. (1980) described a count, using
a combination of a transect and circular plots (Fig-
ure 7). This count method has been widely used

Figure 7. Variable circular-plot method is based on
principles of the variable-strip transect method.

Songbirds

305

Mark and Recapture Methods

Several investigators in the U.S., Sweden, and
France have marked and recaptured birds as a means
of studying the effectiveness of other census meth-
ods (De Sante 1981, 1985; Robbins 1978). Capture,
mark, and recapture methods are often called Lin-
coln or Peterson indexes (Davis and Winstead
1980).

Trapping and banding in itself is neither an effi-
cient nor a highly accurate way of measuring entire
breeding bird populations. It is, however, a very
effective way in determining how many pairs of cer-
tain species are present and in distinguishing mi-
grants from summer residents and, to some degree,
non-breeding from breeding individuals. Color-band-
ing or radiotelemetry can be used to define the
ranges of individual birds and also can point out
errors in judgment that occasionally occur when the
observer relies entirely on the mapping method
(Robbins 1978).

In France, Frochot et al. (1978) compared spot-
mapping, IPA, and capture-recapture methods in a
100-ha (247-a.) oak forest plot. Comparing the re-
sults for 1 2 common species, he estimated 49. 9
breeding pairs per 10 ha (24.7 a.) by the mapping
method and 47.0 by the IPA method (with appropri-
ate corrections for conspicuousness ). Banding data

sufficient for computation of population estimates
were available for only 4 of the 1 2 species, but for 3
of these 4, the estimate from capture-recapture was
higher than obtained by either of the other two
methods. Frochot et al. (1978) reported that while
the IPA census required less than 10 hours of prime
time in the early morning, the mapping census re-
quired 43 hours and the banding study, 400. They
emphasized three major advantages of the capture-
recapture method:

( 1 ) It permits a census of females and young as
well as of singing males;

(2) It can be used for testing mapping and IPA
methods; and

(3) It gives additional information about daily
range and habitat use of the individual birds.

In Colorado, Porter (1973) also used capture-recap-
ture methods to check other census methods in
grasslands.

Hewitt (1967) devised a roadside count to esti-
mate breeding populations of red-winged blackbirds
(Agelaius phoeniceus), which basically is a Lincoln
Index without having to capture and mark. It should
work well for other blackbirds found in western
marshes.

^^

Western bluebird.

Red-breasted nuthatch.

306

Songbirds

Nest Monitoring

Perhaps nest monitoring would seem an obvious
way to accurately determine breeding bird popula-
tions. In actual practice, however, it is seldom possi-
ble to find enough active nests of a species to use
this method for measuring the breeding population,
especially in forest habitats.

A good example of a nesting study of a single
species is an investigation of wood thrushes (Hylo-
cichla mustelina) made by Longcore and Jones
(1969) in a 14.4 ha (35.6 a.) Delaware woodlot.
Systematic nest searches were conducted during a
3-month period in 1965 and 1966. Grid lines located
at 45.7-m (150-ft) intervals were traversed at least
once every 3 days, except for a 2 -week period in
July 1965. A total of 142 wood thrush nesting at-
tempts were documented, of which 54 ( 38% ) were
successful. Nest finding was supplemented by inten-
sive banding efforts every 2 to 3 weeks during the
1966 breeding season, which resulted in the banding
of 46 wood thrushes. The aim of this study was to
determine reproductive success rather than to meas-
ure the population, but it gives an idea of how much
effort would be required to gather enough informa-
tion about a single species to get an accurate meas-
urement of the breeding population. Similar
approaches should work for western thrushes.

De Sante (1981) used nest monitoring in con-
junction with color-banding and spot-mapping to
estimate breeding densities of eight species of Cali-
fornia coastal scrub at Point Reyes National Seashore.
De Sante (1985) used a similar approach to count
individuals of 19 species of breeding birds in a subal-
pine lodgepole pine forest in the Sierras.

Conclusion

Many techniques are available to inventory and
monitor songbird numbers and the habitats they
utilize. For breeding birds, the best overall method
involves spot-mapping territorial males, but consider-
able time and effort are required. Eight to 10 early
morning counts of 10- to 20-ha (24.7- to 49.4-a.)
plots are recommended. Additional time is required
to analyze the habitat. Walked-line transects or vari-
able-circular plots along transects can also be used
to estimate avian populations. These usually require
fewer coverages than spot-mapping, but greater skill
and training are needed to estimate distances to
birds detected. Playback of recorded calls can be
used to elicit responses from more elusive species
but must be used consistently or not at all. Roadside
counts, such as the North American Breeding Bird
Survey (BBS), can sample larger areas but only give
indexes to bird numbers. BBS results can be corre-
lated to habitat if such data are recorded for each
stop. A variety of point-counts and capture, mark,
and recapture methods are available but, especially if

trapping is involved, require considerable effort.
The results of Christmas Bird Counts and winter bird
population studies sponsored by the National Audu-
bon Society are published regularly in American
Birds and are valuable sources of information on
winter distribution and trends in bird numbers. Ta-
ble 2 summarizes the comparative advantages and
disadvantages of several monitoring methods.

Rather than attempting to monitor all species of
birds in many different habitats, one might want to
concentrate on just a few species in selected habi-
tats. Often threatened or endangered species in "crit-
ical" habitats are stressed. Szaro and Balda (1982)
recommended selecting bird species for use as "indi-
cator" species by considering:

(1) residency status (i.e., summer, winter, or
permanent residence),

( 2 ) foraging and/or nesting substrate,

(3) adequate data base,

(4) ease of monitoring,

(5) sensitivity to habitat perturbations,.

(6) sensitivity to environmental fluctuations,

(7) the condition or range of conditions a given
species will indicate, and

(8) biogeographic considerations based on frag-
mentation of the habitat.

Graul et al. (1976) and Graul and Miller (1984)
recommended an ecological-indicator approach for
management, where the species in an area having
the most exacting ecological requirements are moni-
tored. This "steno" (as opposed to "eury") species
approach is based on the idea that by managing to
create habitat conditions for the species with the
narrowest requirements, a manager will provide for
all species. Other related approaches have stressed a
featured-species concept where a single species,
not necessarily stenotypic, is selected, and a unit of
habitat managed primarily to benefit or feature that
species. Graul and Miller (1984) reviewed still other
ecosystem management approaches in which song-
birds might be emphasized. An example of a steno-
typic songbird that might be monitored as an early
warning of excessive human disruption of its envi-
ronment would be the loggerhead shrike {Laniiis
ludovicianus; Bystrak 1983). The species has dem-
onstrated a considerable population decline, which
suggests a narrow range of tolerance, thus qualifying
it as an indicator species. The BBS provides good
population trends for loggerhead shrike (Robbins et
al., in press).

Songbirds

307

LITERATURE CITED

ABBOTT, I. 1976. Comparisons of habitat structure and
plant, arthropod, and bird diversity between mainland
and island sites near Perth, Western Australia. Austra-
lian J. Ecol. 1:275-280.

ANDERSON, AH. and A. ANDERSON. 1973- The cactus
wren. Univ. Arizona Press, Tucson. 226pp.

ANDERSON, DR. and R.S. POSPAHALA. 1970. Correction
of bias in belt transect studies of immobile objects.
J. Wildl. Manage. 34:141-146.

ANDERSON, S.H. 1979. Habitat structure, succession and
bird communities. Pages 9-21 in Proceedings of
Workshop: Management of North Central and North-
eastern Forests for Nongame Birds, U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. NC-51. Minneapolis, MN.

and H.H. SHUGART, Jr. 1974. Habitat selection

of breeding birds in an east Tennessee deciduous for-
est. Ecology 55:828-837.

ARBIB, R.S., Jr. 1967. Considering the Christmas count.
Audubon Field Notes 21(l):39-42.

. 1978. The National Audubon Society and amateur

ornithologists. Pages 26-29 in McCrimmon, Jr., DA.
and A. Sprunt, IV, eds. Proceedings of a Conference
on the Amateur and North American Ornithology.
Natl. Audubon Soc. Res. Dep. Ithaca, NY. 80pp.

. 1982. The Christmas bird count and the world.

Am. Birds 36:365-368.

ARMSTRONG, E.A. 1955. The wren. Macmillan Co. New
York, NY. 312pp.

AUSTIN, O.L., Jr., ed. 1968. Life histories of North Ameri-
can cardinals, grosbeaks, buntings, towhees, finches,
sparrows, and allies. U.S. Natl. Mus. Bull. 237, 3 vols.
1889pp.

BAILEY, A.M. and R.J. NIEDRACH. 1965. Birds of Colo-
rado. Denver Mus. Nat. Hist., 2 vols. 895pp.

BAKUS, G.J. 1959a. Observations on the life history of the
dipper in Montana. Auk 76:190-207.

. 1959b. Territoriality, movements, and population

density of the dipper in Montana. Condor 61:410-425.

BALDA, R.P. 1975. Vegetation structure and breeding
bird diversity. Pages 59-80 in Smith, DR., ed. Pro-
ceedings of the Symposium on Management of Forest
and Range Habitats for Nongame Birds. U.S. Dep.
Agric, For. Serv., Gen. Tech. Rep. WO-1, Washington,
DC.

and G.C. BATEMAN. 1972. The breeding biology of

the pinon jay. Living Bird 1 1:5-42.

BARLOW, J.C. 1962. Natural history of the Bell vireo ( Vi-
reo bellii) Audubon. Univ. Kansas Publ., Mus. Nat.
Hist. 12:241-296.

BATTEN, LA. and J.H. MARCHANT. 1977. Bird population
changes for the years 1975-76. Bird Study 24:159-
164.

BEASON, R.C. 1974. Breeding behavior of the horned lark.
Auk 91:65-74.

and E.C. FRANKS. 1973- Development of young

horned larks. Auk 90:359-363.

BENT, AC. 1939. Life histories of North American wood-
peckers. U.S. Natl. Mus. Bull. 174:1-334.

. 1940. Life histories of North American cuckoos,

goatsuckers, hummingbirds, and their allies. U.S. Natl.
Mus. Bull. 176:1-506.

— . 1942. Life histories of North American flycatchers,

larks, swallows, and their allies. U.S. Natl. Mus. Bull.

179:1-555.
— . 1946. Life histories of North American jays, crows,

and titmice. U.S. Natl. Mus. Bull. 191:1-495.
— . 1948. Life histories of North American nuthatches,

wrens, thrashers, and their allies. U.S. Natl. Mus. Bull.

195:1-475.
— . 1949. Life histories of North American thrushes,

kinglets, and their allies. U.S. Natl. Mus. Bull. 196:1-

454.
— . 1950. Life histories of North American wagtails,

shrikes, vireos, and their allies. U.S. Natl. Mus. Bull.

197:25-38.
— . 1953- Life histories of North American wood war-
blers. U.S. Natl. Mus. Bull. 203:1-734.
— . 1958. Life histories of North American blackbirds,

orioles, tanagers, and allies. U.S. Natl. Mus. Bull. 211:1-
549.

BLONDEL, J. 1975. L'analyses des peuplements d'oiseaux,
elements d'un diagnostic ecologique. La Terre et la
Vie. 29:533-589.

BOCK, C.E. 1970. The ecology and behavior of the Lewis
woodpecker (Asyndesmus lewis). Univ. California
Publ. in Zoology 92:1-100.

. 1979. Christmas bird count. Nat. Hist. 88(10):7-11.

. 1 984. Geographical correlates of abundance versus

rarity in some North American winter landbirds. Auk
101:266-273-

, J.H. BOCK, and L.W. LEPTHIEN. 1977. Abundance

patterns of some bird species wintering on the Great
Plains of the U.S.A. J. Biogeography 4:101-110.

and L.W. LEPTHIEN. 1972. Winter eruptions of red-
breasted nuthatches in North America, 1950-1970.
Am. Birds 26:558-561.

and . 1975a. A Christmas count analysis of

woodpecker abundance in the United States. Wilson
Bull. 87:355-366.

and . 1975b. Patterns of bird species diver-
sity revealed by Christmas counts versus breeding
bird surveys. Western Birds 6.95-100.

and . 1976a. Growth in the eastern house

finch population, 1962-1971. Am. Birds 30:791-792.

and . 1976b. Synchronous eruptions of bo-
real seed-eating birds. Am. Nat. 110:559-571.

and . 1976c. Geographical ecology of the

common species of Buteo and Parabuteo wintering in
North America. Condor 78:554-557.
— and T.L. ROOT. 1981. Winter abundance patterns
of landbirds in the U.S. and southern Canada. Am.
Birds 35:891-897.

and R.B. SMITH. 1971. An analysis of Colorado

Christmas counts. Am. Birds 25:945-947.

BOND, R.R. 1957. Ecological distribution of breeding birds
in the upland forests of southern Wisconsin. Ecol.
Monogr. 27:351-384.

BOYD, R.L. 1976. Behavioral biology and energy expendi-
ture in a horned lark population. Ph.D. dissertation,
Colorado State Univ., Fort Collins. 194pp.

BRAZIER, F.H. 1964. Status of the mockingbird in the
northern Great Plains. Blue Jay 22:63-75.

BREWER, R. 1972. An evaluation of winter bird popula-
tion studies. Wilson Bull. 84:261-277.

BURTT, HE. and B.P. BURTT. 1982. Reliability of the
Christmas bird count. Redstart 49:90-93-

BYSTRAK, D., ed. 1974. Wintering areas of bird species
potentially hazardous to aircraft. Nad. Audubon Soc, NY.

308

Songbirds

— . 1983. Loggerhead shrike (Lanius ludovicianm).
Pages 301-310 in Armbruster, J.S., ed. Impacts of Coal
Surface Mining on 25 Migratory Bird Species of High
Federal Interest. U.S. Dep. Inter., Fish and Wildl. Serv.
FWS/OBS-83/35. 348pp.

and C.S. ROBBINS. 1978. Bird population trends

detected by the North American breeding bird sur-
vey. Polish Ecol. Studies 3(4):13M43.

CACCAMISE, D.F. 1974. Competitive relationships of the
common and lesser nighthawks. Condor 76:1-20.

CALDER, W.A. 1973- Microhabitat selection during nesting
of hummingbirds in the Rocky Mountains. Ecol.
54:127-134.

CALL, M.W. 1981. Terrestrial wildlife inventories: Some
methods and concepts. U.S. Dep. Inter., Bur. Land
Manage. Tech. Note 349:1-171.

CAPEN, D.E., ed. 1981. The use of multivariate statistics in
studies of wildlife habitat. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM087. 249pp.

CINK, C.L. and R.L. BOYD 1981. Thirty^third winter bird-
population study. Am. Birds 35:21-45.

CODY, ML. 1981. Habitat selection in birds: The roles of
vegetation structure, competitors, and productivity.
BioScience 31:107-113.

COLQUHOUN, M.K. 1940. Visual and auditory conspicu-
ousness in a woodland bird community: A quantita-
tive analysis. Proceedings, Zoological Soc. London
110:129-148.

CONNER, R.N. and J.G. DICKSON. 1980. Strip transect
sampling and analysis for avian habitat studies. Wildl.
Soc. Bull. 8:4-10.

CORNWELL, G.W. 1963. Observations on the breeding
biology and behavior of a nesting population of
belted kingfishers. Condor 65:426-431.

DAVIS, D.E., ed. 1982. CRC handbook of census methods
for terrestrial vertebrates. CRC Press, Inc. Boca Raton,
FL. 397pp.

and R.L. WINSTEAD. 1980. Estimating the numbers

of wildlife populations. Pages 221-245 in Schemnitz,
S.D., ed. Wildlife Management Techniques Manual.
The Wildlife Soc. Washington, DC. 686pp.

DEGRAFF, R.M., tech. coord. 1978a. Proceedings of the
workshop on management of southern forests for
nongame birds. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. SE-14.

1978b. Proceedings of workshop on nongame bird

management in the coniferous forests of the western
U.S. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. PNW-
64.

and K.E. EVANS, compilers. 1979. Management of

north-central and northeastern forests for nongame
birds. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep. NC-
51.

and N.G. TILGHMAN, eds. 1980. Workshop pro-
ceedings, management of western forests and grass-
lands for nongame birds. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. INT-86.

DEHAVEN, R.W. and P.J. DEHAVEN. 1973. A contribution
toward a bibliography on the starling. U.S. Dep. Inter.,
Fish and Wildl. Service. Denver, CO. 92pp.

DESANTE, D.F. 1981. A field test of the variable circular-
plot censusing technique in a California coastal scrub
breeding bird community. Pages 177-185 in Ralph,
C.J. and V.M. Scott, eds. Estimating the numbers of
terrestrial birds. Studies Avian Biol. 6.

1985. A field test of the variable circular-plot cen-
susing technique in a Sierran subalpine forest breed-

ing bird community. Abstract of paper presented at
fourth joint meeting of Wilson and Cooper Ornithol-
ogical Societies, Univ. Colorado, Boulder, June 4-9.
Upp.

DICKSON, J.G. 1978. Comparison of breeding bird census
techniques. Am. Birds 32:10-13-

DLXON, K.L. 1961. Habitat distribution and niche relation-
ships in North American species of Parus. Pages 179-
216 in Blair, W.F., ed. Vertebrate Speciation. Univ.
Texas Press, Austin.

DOWN, D.D. 1970. Indexing population densities of the
cardinal with tape-recorded calls. Wilson Bull. 86:83-
91.

DYBBRO, T. 1976. De danoke gorglefulgles udbredelse.
Dansk Ornith Forening, Copenhagen. 293pp.

EIPPER, AW. 1956. Differences in vulnerability of the
prey of nesting kingfishers. J. Wildl. Manage. 20:177-
183-

EMLEN, J.T 1971. Population densities of birds derived
from transect counts. Auk 88:323-342.

. 1977. Estimating breeding season bird densities

from transect counts. Auk 94:455-468.

ENEMAR, A. 1959. On the determination of the size and
composition of a passerine bird population during the
breeding season. Var Fagelvard, Supplement 2:1-114.

and B. SJOSTRAND. 1970. Bird species densities

derived from study area investigations and line tran-
sects. Swedish Nat. Sci. Res. Council, Ecol. Res.
Comm. Bull. 9:33-37.

ENGSTROM, R.T. and F.C. JAMES. 1981. Plot size as a
factor in winter bird-population studies. Condor
83:34-41.

ERSKINE, A.J. 1971. A preliminary catalogue of bird cen-
sus plot studies in Canada. Canadian Wildl. Serv. Prog-
ress Notes 20:1-78.

. 1972. A preliminary catalogue of bird census plot

studies in Canada. Part 2, Canadian Wildl. Serv. Prog-
ress Notes 30:1-42.

. 1976. A preliminary catalogue of bird census plot

studies in Canada. Part 3, Canadian Wildl. Serv. Prog-
ress Notes 59:1-24.

. 1978. The first 10 years of the cooperative breed-
ing bird survey in Canada. Canadian Wildl. Serv. Rep.,
Serial 42:1-59.

FALLS, J. B. 1981. Mapping territories with playback: An
accurate census method for songbirds. Pages 86-9 1 in
Ralph, C.J. and J.M. Scott, eds. Estimating the numbers
of terrestrial birds. Studies Avian Biol. 6.

FEARE, C. 1984. The starling. Oxford Univ. Press. New
York, NY. 315pp.

FERRY, C. and B. FROCHOT. 1970. Lairfaune nidofrea-
trice d'une foret de chenes pedoncules en bourgogne
etude de deux successions ecologiques. La Terre et
la Vie 24:153-250.

FLACK, J.A.D. 1976. Bird populations of aspen forests in
western North America. Ornith. Monogr. 19. 97pp.

FRANZEB, K.E. 1977. Inventory techniques for sampling
avian populations. U.S. Dep. Inter., Bur. Land Manage.
Tech. Note 307. 17pp.

FROCHOT, B., D. REUDET, and Y LERUTH. 1978. A com
parison of three different methods of census applied
to the same population of forest birds. Polish Ecol.
Studies 3(4>.71-76.

GIBB, J.A. 1980. New Zealand ornithology during the past
50 years. Bull. British Ornith. Club 100:93-96.

GOODWIN, D. 1976. Crows of the world. Cornell Univ.
Press. Ithaca, NY. 354pp.

Songbirds

309

GRABER, R.R. 1971. Illinois birds: Turdidae. Illinois Nat.

Hist. Surv. Biol. Notes 75:1-44.
. 1972. Illinois birds: Hirundenidae. Illinois Nat.

Hist. Surv. Biol. Notes 80:1-36.
. 1973- Illinois birds: Laniidae. Illinois Nat. Hist.

Surv. Biol. Notes 83:1-18.
. 1974. Illinois birds: Tyrannidae. Illinois Nat. Hist.

Surv. Biol. Notes 86:1-56.
and J.W. GRABER. 1963. A comparative study of

bird populations in Illinois, 1906-1909 and 1956-

1958. Illinois Nat. Hist. Surv. Bull. 28:377-528.

-, and E.L. KIRK. 1970. Illinois birds: Mimi-

del. Illinois Nat. Hist. Surv. Biol. Notes 68:1-38.

GRABER, V.W., R.R. GRABER, and E.L. KIRK 1983. Illinois
birds: Wood warblers. Illinois Nat. Hist. Surv. Biol.
Notes 118. 144pp.

GRAUL, WD. and G.C. MILLER. 1984. Strengthening eco-
system management approaches. Wildl. Soc. Bull.
12:282-289-

, J. TORRES, and R. DENNEY. 1976. A species-eco-
system approach for nongame programs. Wildl. Soc.
Bull. 4:79-80.

GRISCOM, L. and A. SPRUNT, Jr. 1957. The warblers of
America. Devin-Adair Co. New York, NY. 356pp.

HARRISON, H.H. 1975. A field guide to birds' nests.
Houghton Mifflin Co., Boston, MA. 257pp.

. 1984. Wood warbler's world. Simon & Schuster,

New York, NY. 335pp.

HEWITT, D.H. 1967. A road-count index to breeding pop-
ulations of red-winged blackbirds. J. Wildl. Manage.
31:39-47.

HILDEN, O. 1965. Habitat selection in birds. Ann. Zoology
Fennici 2:53-75.

HOLMES, R.T., RE. BONNEY, and S.W. PACALA. 1979.
Guild structure of the Hubbard Brook bird commu-
nity: A multivariate approach. Ecol. 60:512-520.

and S.K ROBINSON. 1981. Tree species preference

of foraging insectivorous birds in a northern hard-
woods forest. Oecologia 48:31-35.

INTERNATIONAL BIRD CENSUS COMMITTEE. 1969
Recommendations for an international standard for a
mapping method in bird census work. Bird Study
16:249-255 (also 1970, Audubon Field Notes 24:723-
726.)

JACKMAN, S.M. and J.M. SCOTT. 1975. Literature review
of 23 selected forest birds of the Pacific Northwest.
U.S. Dep. Agric, For. Serv. Portland, OR. 382pp.

JAMES, F.C. 1971. Ordinations of habitat relationships
among breeding birds. Wilson Bull. 83:215-236.

and H.H. SHUGART, Jr. 1970. A quantitative

method of habitat description.

JARVTNEN, O. and R.A. VAISANEN. 1976. Finnish line
transect censuses. Ornis Fenn 53:115-118.

JOHNSGARD, PA. 1983. The hummingbirds of North
America. Smithsonian Inst. Press, Washington, DC.
303pp.

JOHNSON, R.R., B.T. BROCON, L.T. HAIGHT, and J.M.
SIMPSON. 1981. Playback recordings as a special
avian censusing technique. Pages 68-75 in Ralph, C.J.
and J.M. Scott, eds. Estimating the numbers of terres-
trial birds. Studies Avian Biol. 6.

JORGENSEN, OH. 1974. Results of IPA-censuses on Dan-
ish farmland. Acta Ornithologica 14:310-321.

KARR, JR. 1980. History of the habitat concept in birds
and the measurement of avian habitats. Pages 991-997
in Nohring, R., ed. Acta XVII Int. Congress Ornitholo-
gica, Berlin.

KENDEIGH, S.C. 1944. Measurement of bird populations.
Ecol. Monogr. 14:67-106.

KESSEL, B. 1957. A study of the breeding biology of the
European starling (Sturnus vulgaris L. ) in North
America. Am. Midi. Nat. 58:257-331.

KLIMKIEWICZ, M.K and C.S. ROBBINS. 1974. The breed-
ing bird atlas of Montgomery County, Maryland, USA.
Acta Ornithologica 14:446-458.

and J.K SOLEM. 1974. First year of breeding bird

atlas, Howard County, Maryland. Maryland Birdlife
30:27-35.

KNORR, O.A. 1961. The geographical and ecological dis-
tribution of the black swift in Colorado. Wilson Bull.
73:155-170.

KOLB, H. 1965. The Audubon winter bird-population
study. Audubon Field Notes 19:432-434.

LACK, D. 1937. A review of bird census work and bird
population problems. Ibis 79:369-395.

LAUGHLIN, S.D., ed. 1982. Proceedings of the northeast-
ern breeding bird atlas conference. Vermont Inst,
of Nat. Sci., Woodstock, VT.

, D.F. KIBBE, and P.F.J. EAGLES. 1982. Atlasing the

distribution of the breeding birds of North America.
Am. Birds 35:6-19.

LEA, R.B. 1942. A study of the nesting habits of the cedar
waxwing. Wilson Bull. 54:225-237.

LONGCORE, JR. and R.E.JONES. 1969. Reproductive
success of the wood thrush in a Delaware woodlot.
Wilson Bull. 81:396-406.

MACARTHUR, R.W. and J.W. MACARTHUR. 1961. On bird
species diversity. Ecol. 42:594-598.

MARION, W.R., T.E. OMEARA, and D.S. MAEHR. 1981. Use
of playback recordings in sampling elusive or secre-
tive birds. Pages 81-85 in Ralph, C.J. and J.M. Scott,
eds. Estimating the Numbers of Terrestrial Birds.
Studies Avian Biol. 6.

MASER, C, J.W. THOMAS, and R.G. ANDERSON. 1984.
The relationship of terrestrial vertebrates to plant
communities and structural conditions. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. PNW-172, 2 parts.
237pp.

MCCLELLAND, B.R. and S.S. FRISSELL. 1975. Identifying
forest snags for hole-nesting birds. J. Forestry 73:414-
417.

MCNICHOLL, M.K 1981. Caution needed in use of play-
backs to census bird populations. Am. Birds 35:235-
236.

MEENTS, J.K, JR. RICE, B.W. ANDERSON, and R.D.
OHMART. 1983- Nonlinear relationships between
birds and vegetation. Ecol. 64:1022-1027.

MERIKALLIO, E. 1958. Finnish birds, their distribution
and numbers. Fauna Fennica 5:1-181.

MILLER, A.H. 1931. Systematic revision and natural history
of the American shrikes. Univ. California Publ. in
Zoology 38:11-242.

NERO, R.W. 1984. Redwings. Smithsonian Institute Press,
Washington, DC. 160pp.

NORRIS, R.A. 1958. Comparative biosystematics and life
history of the nuthatches {Sitta pygmaoa and Sitta
pusilla). Univ. California Publ. Zoology 56:1 19-300.

OHMART, R.D. 1973. Observations on the breeding adap-
tations of the roadrunner. Condor 75:140-149.

ORIANS, G.H. 1980. Some adaptations of marsh-nesting
blackbirds. Princeton Univ. Press, Princeton, NJ.
295pp.

PACKARD, G.C. 1966. Evolution of North American house
sparrows in relation to altitude and aridity. Ph.D.

310

Songbirds

1

thesis, Univ. Kansas, Lawrence. 86pp.
PALMGREN, P. 1930. Quantitative Untersuchesogen uber

die Vogelfauna in den Waldern Sudfinnlands mit be-

sonderer Berucksichtigring Alands. Acta Zod. Fenn.

7:1-218.
PEERING, F.H and S.M. WALTERS. 1962. Atlas of the

British flora. Botanical Soc. British Isles. Nelson Press,

New York, NY. 32pp.
PHILLIPS, A.R., J. MARSHALL, and G. MONSON. 1964. The

birds of Arizona. Univ. Arizona Press, Tucson. 220pp.
PICKWELL, G.A. 1931. The prairie horned lark. Trans.

Acad. Sci., St. Louis, MO. 27:1-153.
PORTER, D.K. 1973. Accuracy in censusing breeding

passerines on the short -grass prairie. M.S. thesis, Co-

loado State Univ., Fort Collins. 107pp.
, MA. STRONG, J.B. GIEZENTANNER, and R.A.

RYDER. 1975. Nest ecology, productivity, and growth

of the loggerhead shrike on the short-grass prairie.

Southwestern Nat. 19:429-436.
PREBLE, N.A. 1957. Nesting habits of the yellow-billed

cuckoo. Am. Midi. Nat. 57:474-482.
PRESTON, F.W. 1958. Analysis of the Audubon Christmas

counts in terms of the lognormal curve. Ecol. 39:620-

624.
PRICE, F.E. and C.E. BOCK. 1983. Population ecology of

the dipper {Cinclus mexicanus) in the Front Range

of Colorado. Cooper Ornith. Soc, Studies Avian Biol.

7. 84pp.
PUTNAM, L.S. 1949. Life history of the waxwing. Wilson

Bull. 61:141-181.
RALPH, C.J. and J.M. SCOTT, eds. 1981. Estimating num-
ber of terrestrial birds. Studies Avian Biol. 6. 630pp.
RAMSEY, F.L. and J.M. SCOTT. 1979. Estimating popula-
tion densities from variable circular plot surveys.

Pages 155-181 in Cormack, R.M., G.P. Patel, and D.S.

Rolson, eds. Sampling Biological Populations. Int.

Co-op Publ. House, Fairfield, MD.
RAYNOR, G.S. 1983. A method for evaluating quality of

coverage in breeding bird atlas projects. Am. Birds

37:9-13.
REYNOLDS, R.T., J.M. SCOTT, and R.A. NUSSBAUM. 1980.

A variable circular -plot method for estimating bird

numbers. Condor 82:309-313-
RICE,J.C, R.D. OHMART, and B.W. ANDERSON. 1983.

Habitat selection attributes of an avian community: A

discriminant analysis investigation. Ecol. Monogr.

53:263-290.
ROBBINS, C.S. 1966. The Christmas count. Pages 154-163

in Stefferud, A., ed. Birds in Our Lives. U.S. Dep. In-
ter., Fish and Wildl. Serv. Washington, DC. 56 lpp.
. 1969. The breeding bird survey of 1967 and 1968.

U.S. Dep. Inter., Fish and Wildl. Serv. Spec. Sci. Rep. —

Wildl. 124:1-107.
. 1970. Winter bird survey technique tested in

Maryland. Maryland Birdlife. 26:11-20.
. 1972. An appraisal of the winter bird-population

study technique. Am. Birds 26:688-692.
. 1974. Progress report on the North American

breeding bird survey. Acta Ornithologica 14:170-191.
. 1 978. Census techniques for forest birds. Pages

142-163 in Proceedings of Workshop: Management of

southern forests for nongame birds. U.S. Dep. Agric,

For. Serv. Gen. Tech. Rep. SE-14.
, D. BYSTRAK, and PH. GEISSLER. 1986. The breed-
ing bird survey: its first 15 years, 1965-1979. U.S.

Dep. Inter., Fish and Wildl. Serv., Resour. Publ. 157:1-

154.

and W.T. VAN VELZEN. 1967. The breeding bird

survey, 1966. U.S. Dep. Inter., Fish & Wildl. Serv.

Spec. Sci. Rep— Wildl. 102:1-43-
ROOT, R.B. 1967. The niche exploitation pattern of the

blue-gray gnatcatcher. Ecol. Monogr. 37:317-350.
ROTENBERRY, J.T. 1980a. Temporal variation in habitat

structure and shrub-steppe bird dynamics. Oecologia

47:1-9.
. 1980b. Habitat structure, patchiness, and avian

communities in North American steppe vegetation: A

multivariate analysis. Ecol. 61:1228-1250.

and J.A. WIENS. 1976. A method for estimating

species dispersion from transect data. Am. Midi. Nat.
95:64-78.

ROTH, R.R. 1981. Vegetation as a determinant in avian
ecology. Pages 162-174 in Proceedings First Welder
Wildl. Foundation Symp.

SCOTT, J.M. and F.L. RAMSEY. 1981. Effects of abundant
species on the ability of observers to make accurate
counts of birds. Auk 98:610-612.

, , and C.B. KEPLER. 1981. Distance estima-
tion as a variable in estimating bird numbers. Pages
366-371 in Ralph, C.J. and Scott, J.M. , eds. Estimating
the Numbers of Terrestrial Birds. Studies Avian Biol.
6.

SEDGWICK, J.A. 1981. Breeding bird and small mammal
habitat relationships in northwestern Colorado. PhD
dissertation, Colorado State Univ., Fort Collins. 1 39pp.

SERVENTY, D.L. 1980. Developments in Australian orni-
thology. Bull. British Ornith. Club. 100:89-93-

SHARROCK, J.T.R. 1975. Dot distribution mapping of
breeding birds in Europe. Ardeola 21:797-810.

. 1976. The atlas of breeding birds in Britain and

Ireland. British Trust for Ornith., Tring. Herts. 477pp.

SKAAR, P.D. 1969. Birds of Bozeman Latilong. Published
by author, 501 S. Third, Bozeman, MT. 131pp.

SMITH, DR. 1975. Proceedings of the symposium on

management of forest and range habitats for nongame
birds. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
WO-1.

STEWART, RE., J.B. COPE, C.S. ROBBINS, and J.W. BRAI-
NERD. 1952. Seasonal distribution of bird populations
at the Patuxent Research Refuge. Am. Midi. Nat.
47:257-363.

SUMMERS-SMITH, J.D. 1963. The house sparrow. Collins.
London. 269pp.

. 1 967. Bibliography of the genus Passer. Intl. Stud-
ies of Sparrows. Warsaw, Poland. 82pp.

SZARO, R.C. and R.P. BALDA. 1982. Selection and moni-
toring of avian indicator species: An example from
a ponderosa pine forest in the Southwest. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. RM-89. 8pp.

and M.D. JAKLE. 1982. Comparison of variable

circular-plot and spot-mag methods in desert riparian
and scrub habitats. Wilson Bull. 94:546-550.

TAYLOR, DM. and CD. LITTLEFIELD. 1984. Willow fly-
catcher and yellow warbler responses to reduced
cattle use of willow habitat. Abstract of paper pre-
sented at 54th Ann. Meeting, Cooper Ornith. Soc.
Humboldt State Univ., Areata, CA. June 19-23- 26pp.

TOMIALOJC, L., W. WALANKIEWICZ, and T WESO-

LOWSKI. 1978. Methods and preliminary results of
the collective bird census work in the primeval forest
of the Bialowieza National Park. Polish Ecol. Studies
3(4):215-224.

TRAMER, E.J. 1974. An analysis of the species density of
U.S. landbirds during the winter using the 1971

Songbirds

311

Christmas bird count. Am. Birds 28:563-567.
TRIMBLE, S. 1975. Non-game birds of the West: An anno-
tated bibliography. U.S. Dep. Inter., Bur. Land Manage.

Tech. Note 269. 320pp.
UDVARDY, M.D.F. 1957. An evaluation of quantitative

studies in birds. Cold Spring Harbor Symp. Quantita-
tive Biol. 22:301-311.
U.S. DEPARTMENT OF AGRICULTURE, FOREST SER

VICE, Rocky Mountain Region. 1982. Wildlife and fish

habitat relationships. 2 vols. Denver, CO.
, Southwestern Region. 1975. Endangered and

unique fish and wildlife of the Southwestern National

Forests. Albuquerque, NM. 203pp.
VAN VELZEN, W.T., ed. 1981. Forty -fourth breeding bird

census. Am. Birds 35:46-112.
VERBEEK, N.A.M. 1970. Breeding- ecology of the water

pipit. Auk 87:425-451.
VERNER, J. 1985. Assessment of counting techniques.

Chapter 8 in Current Ornithology 2:247-302.
and A.S. BOSS. 1980. California wildlife and their

habitats: western Sierra Nevada. U.S. Dep. Agric, For.

Serv. Gen. Tech. Rep. PSW-37. 439pp.
and J.V. RITTER. 1985. A comparison of transects

and point counts in oak-pine woodlands of California.

Condor 87:46-68.
WEBSTER, J.D. 1966. An analysis of winter bird-population

studies. Wilson Bull. 78:456-461.

WIENS, J.A. 1969. An approach to the study of ecological
relationships among grassland birds. Ornith. Monogr.
8:1-93.

. 1973- Pattern and process in grassland bird com-
munities. Ecol. Monogr. 43:237-270.

1983- Avian community ecology: An iconoclastic

view. Pages 355-403 in Brush, AH. and G.A. Clark, Jr.,

eds. Perspectives in Ornithology. Cambridge Univ.

Press. New York, NY. 560pp.
WILLIAMS, A.B. 1936. The composition and dynamics of a

beech-maple climax community. Ecol. Monogr. 6:317-

408.
WILLSON, M.F. 1974. Avian community organization and

habitat structure. Ecol. 55:1017-1029.
WING, L. and M. JENKS. 1939. Christmas censuses: The

amateur's contribution to science. Bird Lore 41:343-

350.
WINTERNITZ, B.L. 1976. Temporal change and habitat

preference of the montane breeding birds. Condor

78:382-393.
YEATMAN, L. 1976. Atlas des oiseaux nicheur de France

de 1970 a 1975. Soc. Ornithology de France, 55 Rue

de Buffon, Paris. 283pp.
YOUNG, H. 1955. Territorial behavior and nesting of the

eastern robin. Am. Midi. Nat. 53:329-352.
ZELENY, L 1976. The bluebird: How you can help its fight

for survival. Indiana Univ. Press, Bloomington. 170pp.

312

Songbirds

16
RAPTORS

Michael N. Kochert

U.S. Bureau of Land Management

Snake River Birds of Prey Research Project

Boise, ID 83705

Editor's Note: The importance of raptors is widely
recognized. Interest in these species has resulted in
many studies and much literature. At times, biolo-
gists can have trouble wading through this mate-
rial to find what they need. This chapter
summarizes the available information on raptors
and refers the biologist to major studies and
reports.

INTRODUCTION

Raptors are difficult to accurately count because
they often nest in inaccessible areas at relatively
low densities. They can be secretive and difficult to
find, and are wide-ranging and rapid-moving. Because
of these difficulties, a census (complete count) is
rarely possible. In most raptor surveys, population
and productivity estimates are based on samples.
Because data are sometimes hard to obtain, it may
be impractical to obtain adequate sample sizes and
results may differ depending on how the data were
collected. These problems can be compounded by
poor sampling design and methods of interpretation
(Steenhof and Kochert 1982), and may result in a
waste of time and money.

Surveys should provide adequate information to
determine population changes. The type and inten-
sity of an inventory or monitoring effort may be
dictated by personnel, funding, or logistics. Because
of these restrictions, biologists may be required to
use low resolution surveys. However, statistically
sound sampling procedures must be used regardless
of the resolution.

When designing an inventory or monitoring
effort, biologists should obtain general information
about the raptor species or group they intend to
survey and the habitats in which it occurs (Fuller
and Mosher 1981 ). They also must consider the
biases that influence survey results. Observer compe-
tence, experience in identifying raptors and conduct-
ing surveys, and knowledge of raptor behavior are
important factors influencing survey results and can
be sources of variability. This variability is com-
pounded when several species in many different hab-
itats are surveyed simultaneously.

This chapter provides information necessary to
design a raptor inventory and monitoring effort. The
first section presents the principal habitat features
related to the major species and the use of these
features. The second section briefly describes the
major techniques for measuring raptor populations
and evaluates their limitations, biases, and utility
in an inventory or monitoring program.

'

Raptors

313

HABITAT FEATURES CORRELATED WITH
SPECIES GROUPS

Raptors, like other animals, are associated with
specific habitat features. Often the occurrence of a
raptor species can be predicted by the presence
of certain habitat characteristics. Knowledge of the
relationships between raptors and habitat features
can assist biologists in designing and conducting an
inventory or monitoring effort. This section summa-
rizes the main habitat features for the major raptor
species in the western United States (Table 1 ) and
discusses how biologists should apply these features
in an inventory or monitoring effort. Although west-
ern habitats are emphasized, eastern habitats are
described if the species also occurs in the eastern
United States.

Vegetation and physical (topographic and phys-
iographic) features are the two most important
habitat features that influence the abundance and
distribution of raptors (Newton 1979). These fea-
tures influence availability of nest sites and foraging
habitat. Some raptors are adapted to forests (e.g.,
accipiters) whereas others are adapted to open grass-
lands and shrublands (e.g., prairie falcons and ferru-
ginous hawks; Table 1 ). Some species occur in a
wide variety of topographic and physiographic situa-
tions and others require a relatively narrow set of
physical features. Great-horned owls, for example,
nest in numerous types of topographic and physio-
graphic situations, but nest dispersion of cliff-nesting

species, such as prairie falcons, is directly influenced
by the distribution of rock outcroppings and escarp-
ments. Other raptor species are closely associated
with water or riparian areas (Table 1 ).

Nesting habitats are emphasized in Table 1 be-
cause more is known about nesting requirements
than requirements during other portions of raptors'
life cycles, and successful nesting is critical in main-
taining population stability (Newton 1979). Nesting
habitat features in Table 1 are presented in the form
of a hierarchy. The vegetation category includes
the major vegetation types (Kuchler 1964; Brown et
al. 1980) in which a species occurs throughout its
range. Physical features represent specific physio-
graphic and topographic situations within the vegeta-
tion types which are associated with a particular
species. Habitat characteristics reflect special features
of the vegetation where the species nests, and can
be related to age class (e.g., old-growth) or vegeta-
tive structure (e.g., closed, dense canopy, or open
low shrubs). For example, northern goshawks and
Cooper's hawks both nest in conifer forests in the
west; however, goshawks occur in dense mature
stands with a multi-layered canopy and Cooper's
hawks occur in even-aged second growth stands
with a dense canopy (Table 1). Associations with
farmland and urban development are included in this
category of habitat characteristics. Farmland habitat
includes cultivated lands or pasture lands, whereas
urban habitat consists of small towns, large cities,
and suburban areas.

Cliffs furnish nest sites and overlook hunting areas.

314

Raptors

Within these areas of specific habitat characteris-
tics, different raptors use different substrates for
nesting. Some species have a very narrow range of
requirements whereas others nest on many different
substrates. Knowledge of these requirements aids
in locating nests and predicting the occurrence of
nesting birds. Also, certain populations can be re-
stricted by the availability of nesting sites (Newton
1979), especially for tree nesting raptors in the
Great Plains and those that use snags.

Availability of food is the. second factor that
regulates raptor abundance and distribution (Newton
1979). Knowledge of foraging habitat requirements
can help biologists predict raptor occurrence. This
knowledge is important in managing and maintaining
foraging habitat that can be critical to raptor survival
and nesting success. The ranges in foraging habitat
presented in Table 1 are based on a few studies in
the literature and should be considered guidelines
instead of absolute rules.

Wintering habitat may be as important as nest-
ing habitat because habitat quality may influence
juvenile survival and the condition of breeding fe-
males the following spring (Newton 1979). Winter-
ing habitat is less predictable than nesting habitat
because raptors usually use more varied habitats in
the winter. Biologists should be aware that species
such as bald eagles, northern harriers, and long-eared
owls communally roost during the winter, and pro-
tection and management of these roosts are impor-
tant in maintaining wintering habitat for these
species.

Little is known about migration routes and habi-
tat use during migration. Certain raptor species,
however, use migration corridors dictated by geo-
graphic features or barriers (e.g., mountain ranges or
water bodies) which cause migrating birds to con-
gregate (Newton 1979). In the western United
States, migrating raptors tend to follow north-south
mountain ranges where birds use deflective air cur-
rents caused by escarpments. Surveys of migrating
raptors are of little value for local inventory or moni-
toring efforts because the origin of the birds is
mostly unknown and migration counts are subject to
many biases (Fuller and Mosher 1981). These
counts, however, can have value on a regional basis
in determining general population trends.

Biologists should use the information presented
in Table 1 as guidelines for predicting possible pres-
ence of species and for developing inventory and
monitoring designs, especially those using stratified
sampling procedures. However, biologists should not
use these features as a sole means of assessing rela-
tive raptor abundance or habitat quality because
many environmental factors can influence actual
population numbers.

POPULATION MEASUREMENT TECHNIQUES

This section describes the major techniques
for surveying raptor populations that can be applied
in a BLM inventory and monitoring program. How-
ever, before designing a survey, biologists should
consult Steenhof ( 1986) and Fuller and Mosher
(1986) for further discussion of biases, limitations,
and application of these techniques. Discussions
in this section focus on the utility of each technique
in assessing relative population changes, including
evaluations of accuracy and precision. Accuracy is
the closeness of the value computed or measured by
the technique to the true or known value and is
expressed as a percentage of the true or known
value. Precision is the closeness of repeated meas-
urements of the same variable and is the amount of
variation expressed as a percentage of the mean.
Therefore, measurements with greater precision have
less variation in relation to the mean and thus
smaller precision values. I used coefficients of varia-
tion (Sokal and Rohlf 1981 ) to measure precision.
When that was not possible, and if 95% confidence
intervals were reported, I followed Postovit (1979)
and used the following formula which gave a meas-
ure roughly double the coefficient of variation.

Percent precision :

± 95% Confidence Interval
Mean

X100

Presence

A survey to determine presence is often the first
step in evaluating potential impacts of land use and
is also used to describe geographical range of a spe-
cies and changes in distribution (Fuller and Mosher
1981). Specific techniques for assessing presence are
not discussed in this section because they are essen-
tially the same as those used to assess relative abun-
dance, which I describe in a later section. Use of
information about presence of a species does not re-
quire careful enumeration of birds, and data collec-
tion and analysis requirements are more rigorous for
surveys to determine relative abundance.

The method used to determine presence will
depend on funding and logistics as well as the raptor
species to be surveyed. Although data collection
requirements are the least rigorous, methods are
subjected to biases with detectability being the main
bias (Fuller and Mosher 1981). The literature, habi-
tat models, and knowledgeable people are initial
information sources in assessing presence. Aerial sur-
veys are an expeditious way to determine presence
of conspicuous raptors over large areas. Surveys from
boats and land vehicles can be more efficient than
those from aircraft because of less cost and more
opportunity to see inconspicuous raptors. Vocaliza-
tions provide a quick means to detect the presence

Raptors

315

Table 1. Habitat features correlated with selected Falconiform species in the United States.

SPECIES

NESTING HABITAT

Vegetation

Physical Features

Habitat Characteristics

Black vulture
(Coragyps atratus)

Shrubland; grassland;
eastern and southern de-
ciduous, coniferous, and
mixed forest types.

Varied. Flatlands to rolling
hills and foothills, can-
yons, and ravines.

In or near open to semi-open
habitat including farmland.

Turkey vulture
(Cathartes aura)

Numerous shrubland,
grassland, coniferous,
and deciduous forest
types.

Varied. Flatlands to
mountainous to 3750 m
(12000 ft). Canyons,
ravines, and hillsides.

In or near open to semi-open
habitat including farmland.

California condor

(Gymnogyps

californianus)

Primarily chaparral vege-
tation type. One nest
located in a redwood
{Sequoia sp.) type.

Rugged mountainous
terrain between 610-1370
m (2000-4500 ft).

Open to semi-open habitats.

Mississippi kite

(Ictinia

mississippiensis)

Eastern deciduous forest,
grassland, and shrubland
types.

Level to rolling uplands;
also floodplain.

Tree groves, shelter belts;
gallery and riparian forests
near open habitat including
farmland and urban areas.

Black-shouldered kite
(Elanus caeruleus)

Grassland and oak
(Quercus sp.) savannas.

Coastal valleys and
grassy foothills.

Open to semi-open habitats
including farmland.

Northern goshawk
{Accipiter gentilis)

Coniferous, deciduous,
and mixed forest types.

Gentle slopes or flat
areas near a steep incline
or canyon often near
water. Northerly exposure
in western U.S. Southerly
exposure in Alaska.

Dense mature stands with a
multilayered canopy and
dense to open understory.

Cooper's hawk
(Accipiter cooperii)

Coniferous, deciduous,
and mixed forest types.
Also shrubland and
grassland types.

Flat terrain or gentle
slopes (30°), sometimes
near water. Low eleva-
tions. Sea level to 2700 m
(8700 ft) in western U.S.
on northerly exposures.

Wooded areas exceeding 6-8
ha (15-20 a.); occasionally
smaller areas or isolated
trees. Mature deciduous, con-
iferous, or mixed forest with
dense understory in the East.
In western mountainous, old
or even-aged second growth
stands (30-80 years) with a
dense canopy and sparse
ground cover. Also oakwood-
land habitat. Riparian forests
in grassland and shrubland
types.

Common barn-owl
(Tyto alba)

Nearly every shrubland,
grassland, and forest
type except in montane
areas and north of 50°.

Variable. Absent in moun-
tainous terrain.

Open to semi-open habitat.
Common in farmland and
urban habitats.

Western screech-owl
(Otus kennicottii)

Most forest associations,
shrubland, and grassland
types.

Variable below 2470 m
(8000 ft). Also riparian
areas.

Woodland habitat near open
areas. Also farmland and
urban habitats.

Whiskered screech-
owl
{Otus trichopsis)

Oak and oak-pine forests
in the Southwest.

Usually canyons.

Usually riparian habitat and
dense stands.

316

Raptors

FORAGING HABITAT

WINTERING HABITAT

SOURCES

Nesting Substrate

Ground, hollow logs, and
stumps; occasionally
caves and old buildings.

Open to semi-open areas
where carrion is available.

Similar to nesting;
also urban habitat.

Scott et al. (1977); Am.
Ornithol. Union (1983).

Cliff cavities and caves,
ground, hollow logs, and
trees near ground level.

See black vulture.

Similar to nesting.

Scott et al. (1977); Call
(1978).

Cavities, caves and crev-
ices of cliffs, and rarely
hollow trees.

Grassland and oak savanna
foothills up to 60-70 km (35-
45 mi.) from nests where
large mammal carrion is
available.

Same as nesting.

Wilbur (1978); Studer
(1983).

Trees.

Open areas near colonies or
above the forest canopy for
insects.

N/A*

Parker (1975, pers. com-
mun.); Parker and Ogden
(1979).

Trees.

Open pasturelands, grass-
lands, and forest canopy for
insects.

Same as nesting.

Waian and Stendall
(1970); Bammann (1975).

Tall trees, greater than 8
m (25 ft) high.

Same as nesting. Also eco-
tone between forest and
openings. Areas containing
rodents and large birds up to
2.8 km (1.8 mi.) from nests.

Varied forested and
open habitats. More
open than nesting
habitat.

McGowan (1975); Rey-
nolds et al. (1982); Rey-
nolds (1983).

Medium high trees,
greater than 6 m (20 ft)
high.

Same as nesting. Use areas
containing medium to large
passerines within 2.4 km (1.5
mi.) of nests.

Diverse habitats.
Woodland, grass-
land, shrubland,
farmland, and urban.
More open than
nesting.

Jones (1979); Millsap
(1981); Titus and Mosher
(1981); Reynolds et al.
(1982); Reynolds (1983);
Fuller and Partelow
(1983).

Cavities in trees, cliffs,
and banks. Wide variety
of man-made structures.

Open habitats containing
small rodents within 5.6 km
(3.5 mi.) of nests.

Similar to nesting.

Scott et al. (1977); Call
(1978); C. Marti (pers.
commun.); Hegdal and
Blaskiewicz (1984).

Cavities, woodpecker or
flicker holes in trees,
saguaro cactus, and oc-
casionally snags. Man-
made structure.

Open habitats containing
rodents and insects usually
within 0.5 km (0.3 mi.) of nest.

Similar to nesting.

Scott et al. (1977); Call
(1978); Johnson et al.
(1979); Marti (1979).

Cavities and flicker holes
in trees.

Areas near nest containing
insects.

N/A*

Marshall (1967); Scott et
al. (1977).

Raptors

317

Table 1. Habitat features correlated with selected Falconiform species in the United States (continued).

SPECIES

NESTING HABITAT

Vegetation

Physical Features

Habitat Characteristics

Flammulated owl
(Otus flammeolus)

Coniferous forest and
mixed coniferous-decid-
uous forest types.

Mountainous areas be-
tween 1000-3100 m
(3250-10000 ft). Usually
on dry hillsides or on
ridge tops.

Usually mature forest with
dense canopies and a brush
understory or intermixture
of oaks. Occasionally in sec-
ond growth forest. Rarely in
logged-over areas.

Northern hawk-owl
(Surnia ulula)

Northern boreal forests.

Open woods, parklands,
and barrens with low
shrubs and scattered
trees.

Forest edge areas or open
canopy (20-60% canopy
cover) stands.

Northern pygmy-owl
(Glaucidium gnoma)

Most coniferous associa-
tions throughout the
West.

Mountainous up to 3750
m (12000 ft).

Open to semi-open forests.

Elf owl

(Micrathene whitneyi)

Palo verde-saguaro asso-
ciations and occasionally
coniferous-deciduous
forests at higher
elevations.

Steep canyons, riparian
areas, hillsides, and flats.

Variable.

Sharp-shinned hawk
(Accipiter striatus)

Coniferous, deciduous, or
mixed forest types.

Canyons, valleys, and
riparian areas. In can-
yons, nest 50-100 m
(162-325 ft) upslopefrom
a stream.

Dense stands with well-devel-
oped canopy (60%) and
dense understory. Dense,
young (25-50 year), even-
aged, second growth stands
with single-layered canopy
in western coniferous forests.
Also in conifers within dense
mixed stands.

Osprey

(Pandion haliaetus)

Numerous shrubland,
grassland, coniferous for-
est, and deciduous forest
types.

Proximity to water (can
be up to 1 1 km or 7 mi.
away).

Varied habitats including
farmland and urban. Old
growth stands better than
even-age second growth
stands.

Bald eagle

(Haliaetus

leucocephalus)

Numerous shrubland,
deciduous forest, and
mixed forest types.

Proximity to water (usually
within 3.2 km or 2 mi.).

Uneven-aged, multi-layered
forests with an availability
of large trees with sturdy
branches.

Golden eagle
(Aquila chrysaetos)

Shrubland and grassland
types. Also tundra, coni-
ferous forests, and wood-
land-brushland.

Canyons, buttes, and
mountainous escarp-
ments between 600-3100
m (1950-10000 ft) in
elevation.

Open to semi-open habitats.
Sometimes riparian and farm-
land habitats.

318

Raptors

FORAGING HABITAT

WINTERING HABITAT

SOURCES

Nesting Substrate

Woodpecker or flicker

1 7-27 ha (43-63 a.) areas

N/A*

Bull and Anderson

holes in trees or snags.

around nests containing

(1978); Franzreb and

Also nest boxes.

insects.

Omart (1978); Winter
(1979); Reynolds and
Linkhart (1984).

Cavities in trees, some-

Open areas containing micro-

Same as nesting.

Grossman and Hamlet

times snags. Occasion-

tines, small rodents, and

Sometimes open

(1964); Scott etal.

ally corvid nests in trees

birds.

habitat in temperate

(1977); Meehan and

or on cliffs.

areas.

Ritchie (1982).

Woodpecker holes in

Areas near nests containing

Open to semi-open

Scott etal. (1977); Call

trees or snags.

insects and small rodents.

habitats.

(1978); Hayward (1983).

Woodpecker holes in

Small area around nest con-

N/A*

Scott etal. (1977); Call

saguaro cactus or snags.

taining insects, particularly

(1978); Johnson etal.

Also cavities in dead

beetles, moths, and crickets.

(1979); Millsap (pers.

branches of living trees.

commun.).

Small trees (less than 6 m

Same as nesting areas con-

Diverse habitats

Piatt (1973); Jones

[20 ft] high).

taining small passerine birds

including riparian,

(1979); Millsap (1981);

within approximately 1.5 km

woodland, farmland,

Clarke (1982); Reynolds

(0.9 mi.) of nests.

and urban habitats.

et al. (1982); Reynolds

More open than

(1983).

nesting. Perches

have substantial ar-

boreal cover.

Tops of trees, rock and

Slow-moving water usually

N/A*

Johnson and Melquist

dirt pinnacles, cactus,

within a few km of the nest.

(1973); Zarn (1974c);

and numerous man-made

Can be up to 1 1 km (7 mi.)

Melquist (1975).

structures (snags or trees

away.

with dead tops

preferred).

Primarily trees, occasion-

Primarily slow-moving water

Same as nesting, but

Call (1978); Fraser

ally cliffs. Nest below

within a few km of the nest.

also in farmland

(1978); Steenhof (1978);

crown in super canopy

During winter, bodies of

and urban habitats.

Steenhof et al. (1980);

trees. Occasionally man-

water; occasionally upland

Open water most

Anthony etal. (1982);

made structures.

habitats. Winter foraging

important; however,

Keister and Anthony

areas can be up to 27 km (17

shrubland also used.

(1983); U.S. Fish and

mi.) from roost.

Diurnal perches
primarily in trees
close to a food
source. Night roosts
more protected than
day roosts. Stands
generally have large,
old, open-structured
trees with horizontal
branches.

Wildlife Service (1983).

Primarily cliffs, sometimes

Open habitats where rabbits

Same as nesting. In

Snow (1973); Wrakestraw

trees, and occasionally

and small rodents are avail-

the West below 1400

(1973); U.S. Bureau of

ground and man-made

able, usually within 7 km (4.4

m (4500 ft) in eleva-

Land Management

structures.

mi.) of nests.

tion. In the East

(1979); Kochert (1980);

coastal plains and

Millsap and Vana (1984).

wetlands.

Raptors

319

Table 1. Habitat features correlated with selected Falconiform species in the United States (continued).

SPECIES

NESTING HABITAT

Vegetation

Physical Features

Habitat Characteristics

Northern harrier (Cir-
cus cyaneus)

Grassland and shrubland
associations; also large
open expanses in decid-
uous forests.

Relatively flat and open
terrain from sea level
to 3200 m (10,400 ft) in
elevation.

Open habitat with short vege-
tation. Mainly wetland and
riparian habitats, but also
shrub and grass upland and
farmland habitats.

Red-tailed hawk
(Buteo jamaicensis)

Nearly every open to
semi-open vegetation
type. Absent in tundra
and uncommon in dense
forests.

Canyons, hillsides, and
level terrain. In East oc-
cupy forested uplands,
ridges, and steep slopes.
Riparian floodplain in
Alaska.

In or near open to semi-open
habitats. Nests situated for
unobstructed access in ripar-
ian and dense forested areas.

Swainson's hawk
{Buteo swainsoni)

Shrubland and grassland
types.

Low to moderate eleva-
tions, valleys, foothills,
and level uplands. In
areas moist enough to
support trees.

Small tree groves in open
habitats, including farmland.

Ferruginous hawk
{Bueto regal is)

Grassland, shrubland,
and juniper-pinyon (Juni-
perus-Pinus) woodland
types.

More xeric sites than
other sympatric buteos.
Avoid mountainous areas,
steep canyons, and high
cliffs.

Open habitat. Avoid heavily
farmed areas.

Red-shouldered

hawk

(Buteo lineatus)

Deciduous forest associ-
ations in East. Riparian
and sycamore (Platanus)
woodland in West.

Mosaic lowlands, riparian
areas, floodplains, and
valleys.

Mature forests with fairly
closed canopy and large
amount of floodplain edge in
the East. In the West, riparian
habitat with taller-than-aver-
age trees.

Broad-winged hawk
(Buteo platypterus)

Deciduous and mixed
deciduous forest types.

Wet, poorly drained
areas. More upland than
red-shouldered hawk
areas.

Much like red-shouldered
hawk. Nest close to small
streams, lakes or swampy
areas.

Rough-legged hawk
(Buteo lagopus)

Tundra and taiga.

Foothills and mountain
valleys with bluffs and
rock formations.

Open habitat.

Zone-tailed hawk
(Buteo albonotatus)

Various shrubland types
in the Southwest; also the
Arizona Pine Forest type.

Rugged terrain. Deep,
broad canyons and table-
lands up to 2750 m (9000
ft) in elevation.

Large trees required. Usually
in perennial and intermittent
streams; occasionally in table-
lands away from water.

Gray hawk
(Buteo nitidus)

Riparian forests in mes-
quite and shrubland
types.

Floodplain areas.

Presently mature gallery trees.
Historically mesquite trees.

320

Raptors

FORAGING HABITAT

WINTERING HABITAT

SOURCES

Nesting Substrate

Ground. In rank
vegetation.

Open habitats up to 18 km
(11.2 mi.) from the nest.

Same as nesting, but
at lower elevations.
Use uplands more
often. Will commun-
ally roost in
depressions.

Call (1978); Apfelbaum
and Seelbach (1983);
Thompson-Hanson
(1984).

Trees preferred, also
cliffs and man-made
structures. Higher and in
taller trees than other
sympatric Buteo species.

Open habitats containing
lagomorphs, small rodents,
and snakes, usually within 3
km (1.9. mi.) of nest.

Same as nesting.
Absent from high
elevations.

Brown and Amadon
(1968); U.S. Bureau of
Land Management
(1979); Schmutz et al.
(1980); Cottrell (1981);
Millsap (1981); Titus and
Mosher (1981); Bednarz
and Dinsmore (1982).

Primarily trees and small
shrubs, occasionally
man-made structures.
Closer to the ground than
other sympatric Buteo
species.

Open habitats with short veg-
etation containing small ro-
dents and insects within 3 km
(1 .9 mi.) of the nest and a
home range of approximately
9 km2 (3.5 mi.2).

N/A*

Fitzner (1980); Schmutz
et al. (1980); Cottrell
(1981); Bechard (1982);
Gilmer and Stewart
(1984).

Low rocky outcrops (pre-
ferred), trees in small
groves, on cut banks,
and a variety of man-
made structures.

Open habitats with short veg-
etation containing ground
squirrels, pocket gophers,
and rabbits usually within 2-3
km (1.2-1.9 mi.) of nest.

Same as nesting
habitat, also
pastureland.

Wakely (1978); Schmutz
et al. (1980); Cottrell
(1981); Millsap (1981);
Evans (1983).

Trees, larger than those
used by broad-winged
hawks. Farther below
canopy than red-tailed
hawks.

Closed to semi-closed wood-
land in and near wet areas
containing small rodents,
reptiles, amphibians, and
crayfish in an area 0.6-0.9
km (0.2-3.5 mi.2) around
nest. Up to 1.0-2.3 km (0.6-
1.4 mi.) from nest.

Similar to nesting;
also forest edge and
openings.

Brown and Amadon
(1968); Wiley (1975);
Titus and Mosher (1981);
McCrary (1981); Bednarz
and Dinsmore (1982).

Trees (see red-shoul-
dered hawk).

Closed to semi-open habitat
in an approximately 13 km2
(5.1 mi.2) range containing
small mammals, reptiles,
amphibians, and large
insects.

N/A*

Brown and Amadon
(1968); Matray (1974);
Fuller (1979); Titus and
Mosher (1981).

Primarily river bluffs,
lesser degree upland
outcroppings and es-
carpments. Occasionally
the ground, sometimes
trees.

Wet meadows, bogs,
marshes, open riparian areas,
pastures. Also shrub and
grass uplands containing
small rodents.

Open shrub and
grass habitats in
temperate areas (see
foraging habitat).

Zarn (1975); Bildstein
(1978); BLM (unpubl.
data).

Primarily trees, rarely
cliffs.

Upland habitats, open shrub-
land containing an abun-
dance of lizards up to 27 km
(16.0 mi.) from nests.

N/A* Rare winter
sightings in south-
west U.S.

Call (1978); Millsap
(1981, pers. commun.);
Fuller (1983).

Trees.

Thorn shrub-mesquite bosque
with an abundance of lizards.

N/A* Rare winter
sightings in south-
west U.S.

Millsap (pers. commun.).

Raptors

321

Table 1. Habitat features correlated with selected Falconiform species in the United States (continued).

SPECIES

NESTING HABITAT

Vegetation

Physical Features

Habitat Characteristics

Common black-hawk

{Buteogallus

anthracinus)

Shrubland types of the
Southwest.

Riparian and floodplain
between 430-1850 m
(1400-6000 ft) in eleva-
tion. Perennial streams
essential.

Mature riparian habitat. Tree
groves preferred over single
trees.

Harris' hawk

(Parabuteo

unicinctus)

Mesquite (Prosopis sp.)
and Palo verde (Cerci-
dium sp.) vegetation
types.

Flatlands with washes
and ravines that give way
to low hills.

Structurally complex and
diverse vegetation. Multilay-
ered and semi-closed shrub.

Gyrfalcon
(Falco rusticolus)

Tundra, occasionally in
taiga.

Foothills and mountain
valleys with rock forma-
tions, bluffs, and rock
outcroppings.

Varied. Vast expanses of
open habitat.

Peregrine falcon
(Falco peregrinus)

Nearly all shrubland,
grassland, and forest
types. Rare in alpine
areas.

Flat terrain to rugged
canyons. Usually on high
cliffs (62 m or 200 ft), in
areas of open expanses
and high topographic
relief, and near (within 4.8
km or 3.0 mi.) permanent
or semi-permanent
sources of water.

Open habitats, including
farmland and urban habitats.

Prairie falcon
(Falco mexicanus)

Arid and semi-arid shrub-
land and grassland
types. Sometimes the
open parklands of coni-
ferous forest types.

Canyons, buttes, moun-
tainous escarpments,
outcroppings to eleva-
tions of 3100 m (10,000
ft).

Open habitat with short
vegetation.

Merlin

(Falco columbarius)

Shrubland, grassland,
boreal forest, and coastal
coniferous forest types.

Varied. Uncommon in

rugged-mountainous

terrain.

Near open habitat. Forests
broken by expanses of open
country. Openings in dense
forests. Tree groves, riparian
forests.

American kestrel
(Falco sparverius)

Nearly all shrubland,
grassland, and forest
types.

Variable terrain associ-
ated with open areas up
to 3750 m (12,000 ft).

Wide variety of open to semi-
open habitats, including farm-
land and urban habitats.

Spotted owl Northern
(Strix spotted
occidentalis) owl

Douglas-fir (Pseudotsuga
menziesii), cedar (Thuja
sp.), and hemlock (Tsuga
sp.) forests.

Usually below 1700 m
(5500 ft) elevation on
75% slopes half way
down hillside.

Multilayered old growth (100-
200+ years) forests with
deep ravines and 50-90%
canopy closure.

California

spotted

owl

Fir-pine/Douglas-fir for-
ests and oak woodlands.

Steep slopes.

Heavily timbered old growth.

Mexican

spotted

owl

Spruce, Douglas-fir, and
pinyon-juniper forests.
Aspen and maple stands.

Deep narrow timbered
canyons below 2770
m (9000 ft) in elevation

Timbered riparian habitat.

322

Raptors

FORAGING HABITAT

WINTERING HABITAT

SOURCES

Nesting Substrate

Trees.

Riparian habitat (mostly in the
water) containing fish, am-
phibians, reptiles, and small
mammals within 0.5 km (3.1
mi.) of nests.

N/A*

Schnell (1979); Millsap
(pers. commun).

Palo verde, mesquite,
other trees, and saguaro
cactus (Cereus
giganteus).

Semi-closed areas (more
closed than those used by
red-tailed hawks) containing
rabbits, medium to small
rodents and birds (quail) in a
5 km2 (2.0 mi.2) area around
the nest.

Same as nesting,
also occur in
farmland.

Call (1978); Mader (1976,
1978); Millsap (1981).

Primarily cliffs (6-90 m)
occasionally trees in
taiga.

Open habitats with an abun-
dance of hare and ptarmigan.

Same as nesting.
Sometimes in the
northern contermi-
nous U.S.

Cade (1982); Mindell
(1983).

Cliffs (60 m), rarely in
trees.

Open habitats. Usually ripar-
ian and shoreline areas. Also
upland habitats containing
shorebirds and passerines.
Usually within 5 km (3.1 mi.)
of nests, but up to 24 km (15
mi.).

N/A*

Porter and White (1973);
Cade (1982); Ellis (1982).

Cliffs and rock outcrop-
pings, rarely abandoned
raptor nests in trees.

Open areas of low vegetation
containing ground squirrels
and passerine birds up to 24
km (15 mi.) from nests.

Similar to nesting,
also farmland
habitats.

U.S. Bureau of Land
Management (1979);
Cade (1982); Becker and
Ball (1983a).

Abandoned corvid nests
in trees, occasionally in
tree cavities or on the
ground.

Open shrub and grass habi-
tats containing passerine
birds up to 9 km2 (5.6 mi.2)
from nests. Also forage for in-
sects above forest canopy.

Coastal, marsh,
grassland, farmland,
and urban habitats
where passerine and
shore birds
concentrate.

Trimble (1975); Scott et
al. (1977); Cade (1982);
Becker and Ball (1983b);
Becker (1984); Millsap
(pers. commun.).

Old flicker holes or cavi-
ties in banks and cliffs,
buildings, abandoned
magpie nests, and nest
boxes.

Open habitats mostly in
marshland, grassland, sa-
vanna, shrubland, open for-
est, farmland, and urban
habitats.

Similar to nesting.
Absent in high
elevations.

Scott et al. (1977); Cade
(1982).

Tree cavities or other
raptor nests.

Old growth stands, primarily
second growth stands sec-
ondarily containing flying
squirrels (Glaucomys sp.) and
woodrats {Neotoma sp.) in a
0.5-3.4 km2 (0.2-1.3 mi. ) area
around nest.

Similar to nesting.

Forsman (1983a); Fors-
man et al. (1984).

Tree cavities, other raptor
nests, cavities in cliffs;
occasionally on the ground.

Same as nesting.

Lower elevations.
Tend to be more
open.

Zarn (1974a); Layman
(pers. commun).

Tree cavities, cliff ledges
and cavities, empty rap-
tor nests.

Same as nesting.

Similar to nesting.

Zarn (1974a); C. Marti
(pers. commun.).

Raptors

323

Table 1. Habitat features correlated with selected Falconiform species in the United States (concluded).

SPECIES

NESTING HABITAT

Vegetation

Physical Features

Habitat Characteristics

Barred owl
(Strix varia)

Eastern deciduous and
mixed coniferous forests;
boreal and montane coni-
ferous forests.

In lowlands in East. In
West, features similar to
spotted owl habitats.

In the West, resembles north-
ern spotted owl habitat; how-
ever, less affiliation with old-
growth forests. In East, dense
stands in swamps and river
bottoms.

Boreal owl
(Aegolius funereus)

Boreal and montane
coniferous forests.

Variable from flat boreal
uplands to mountainous
terrain.

Dense, closed canopy stands.

Northern saw-whet

owl

(Aegolius acadicus)

Most forest types up to
3400 m (10,000 ft) in
elevation.

Variable.

Dense, closed canopy stands
with smaller trees, fewer
openings; greater canopy
than those used by boreal
owls.

Great horned owl
(Bubo virginianus)

Nearly every grassland,
shrubland, and forest
type. Absent in tundra.

Variable.

Wide variety of open to semi-
open habitats including farm-
land and urban habitats.

Great gray owl
(Strix nebulosa)

Northern boreal forests
and montane coniferous
forests.

Flat lowlands to moun-
tainous terrain between
1850-2500 m (6000-8100
ft) in elevation.

Old growth conifer or mixed
deciduous stands in montane
forests. Mature poplar, and
to lesser degree, conifer
stands in boreal forests. Nests
usually near a wet meadow
or marsh.

Long-eared owl
(Asio otus)

Most shrubland, grass-
land, and forest types.

Variable, level, to moun-
tainous terrain.

Semi-open to dense wooded
areas, tree groves, and dense
shrubs near open areas in-
cluding farmlands. Nests
usually within a few meters of
the woodland edge.

Burrowing owl
(Athene cunicularia)

Arid and semi-arid shrub-
land and grassland
types.

Level to rolling hills.

Open habitats including farm-
land and urban areas with
short vegetation and an avail-
ability of burrows.

Short-eared owl
(Asio flammeus)

Most shrubland and
grassland types north of
37° north latitude. Also
in tundra.

Open terrain. Absent from
mountainous areas.

Open habitat including grass-
land, shrubland, meadow,
marsh, and farmland habitat.

Snowy owl
(Nyctea scandiaca)

Tundra.

Open flats or slopes.
Nests usually on highest
point; occasionally on
a rocky ledge.

Open habitat. Nests some-
times near marshes or small
lakes.

324

Raptors

FORAGING HABITAT

WINTERING HABITAT

SOURCES

Nesting Substrate

Large hollow trees; occa-
sionally corvid and raptor
nests. Sometimes snags.
Also nest boxes.

Wooded areas with an open
understory in a home range
between 0.8-3.6 km2 (0.3-1.4
mi.2).

Similar to nesting.

Nicholls and Warner
(1972); Scott et al.
(1977); Apfelbaum and
Seelbach (1983); Taylor
and Forsman (1976);
Boxall and Stepney
(1982).

Woodpecker holes in
snags.

Semi-open areas containing

microtines.

Similar to nesting.

Scott et al. (1977); Mee-
han and Ritchie (1982);
Hayward (1983).

Woodpecker holes in
snags and trees. Also
nest boxes.

Areas containing small ro-
dents and insects in a range
up to 0.5 km2 (0.2 mi.2).

Similar to nesting.

Forbes and Warner
(1974); Scott etal.
(1977); Hayward (1983).

Abandoned large bird
nests in trees and on
cliffs. Also cliff ledges
and cavities.

Open to semi-open areas
containing medium to small
mammals in a 1 .3-3.7 km2
(0.5-1.4 mi.2) area.

Similar to nesting.

Call (1978); Fuller (1979);
Am. Ornithol. Union
(1983).

Abandoned raptor or
corvid nests in trees.

Marsh and wet meadows
containing microtines.

Similar to nesting.
Occasionally farm-
land and urban
habitat.

Winter (1979); Nero
(1980).

Abandoned corvid nests
in small scrubby trees
or large bushes.

Open habitats containing
microtines and other small
rodents in 1.8-3.7 km2 (0.7-1.4
mi.2) range.

Similar to nesting.

Call (1978); Marti (1979);
Wijnandts (1984); Marks
(1984).

Abandoned mammal
burrows. Occasional cav-
ities in small basalt
outcroppings.

Open habitat with short vege-
tation containing small ro-
dents, insects, lizards, and
passerine birds.

Similar to nesting.

Zarn (1974b); Marks and
Ball (1983).

Ground

Open habitat (see character-
istics) containing microtines in
a 0.2-1.2 km2 (0.07-0.5 mi.2)
range.

Similar to nesting.

Clark (1975); Call (1978).

Ground

Open uplands and meadows
with a supply of small
rodents.

Same as nesting.
Also coastal, grass-
land, farmland, and
urban habitats.

Grossman and Hamlet
(1964); Lein and Weber
(1979).

"Migrates outside the United States

Raptors

325

of secretive raptors over large areas, and walking
surveys are useful for locating raptors that are diffi-
cult to detect.

Relative Abundance — Non-Nesting Surveys
Road Counts.

Description. Road surveys can be conducted
by either the continuous count or point count
methods (Fuller and Mosher 1986). Routes either
wander through different habitats (Craighead and
Craighead 1956), circle or cut through an area
(Craig 1978), or consist of transects arranged to
completely cover an area (Marion and Ryder 1975).

When conducting continuous counts, observers
repeatedly drive specified routes in a vehicle at
speeds between 16-40 km/h (10-25 mph) and count
all birds seen on either side of the road. They record
the perpendicular distance from the transect for
each bird seen. Because birds are counted on either
side of the road two observers are necessary.

For the point count method, observers stop at
systematically placed stations (usually 0.8 km or 0.5
mi. apart) along the road (Fuller and Mosher 1986).
During a prescribed time period (usually 3 min),
they count all raptors seen or heard within a prede-
termined radius of the station. The radius at stations
can be fixed (usually 0.4 km [0.25 mi.]), variable,
or unlimited. In either situation, biologists must re-
cord the distance from the station of each bird seen
or heard. An advantage of this method is that each
station (or circular plot) can be treated as an inde-
pendent sample with its own habitat characteristics,
thus increasing sample size and enhancing statistical
analysis.

Before conducting road surveys, biologists
should list raptor species and habitat types they need
to survey. They should delineate habitat types on
maps and compute the linear distance of each type
along the road. Transects should be run at least once
to delineate the habitat types, to assess effective
detection distances, and to identify any biases that
may affect the count. Routes should be selected so
that all major communities are represented in the
sample area (Craighead and Craighead 1956; Millsap
1981).

Road counts should be made under similar
weather conditions because weather conditions can
affect raptor activity (Craighead and Craighead
1956). The time, weather (at least precipitation,
wind, cloud cover, and temperature), and vehicle
odometer reading (read to 0.16 km [0.1 mi] and es-
timated to nearest 0.08 km [0.05 mi.]) should be
recorded at the beginning and end of the survey. For

each sighting, observers should record the species,
age class, sex (if possible), and activity (perched or
flying) of the raptor as well as the time, habitat type,
vehicle odometer reading (read to 0.16 km [0.1 mi.]
and estimated to nearest 0.08 km [0.05 mi.]), and
perpendicular distance from the bird to the road
(Millsap 1981).

All road count methods have inherent biases for
which biologists must compensate (Fuller and
Mosher 1986). The season, time of day, weather, and
activity of the birds can bias counts (Fuller and
Mosher 1981). To compensate for this, researchers
have alternated the direction of each survey run,
restricted runs to specific times of the day, or run
transects at several times of the day (Craig 1978;
Thiollay 1978; Millsap 1981). Repeated runs of the
transect in a sampling period increase accuracy, and
at least three repeated runs are desirable. The esti-
mated number of birds per transect for each survey
period is the maximum number of birds seen on a
single run. Man-made structures such as fences, pow-
erlines, and towers provide perches that draw birds
along census routes and cause inflated population
estimates. If possible, routes adjacent to these struc-
tures should be avoided, especially when contrasting
abundance among areas. This bias may not be critical
when assessing population changes in the same area
over time. Temporary influxes of birds, concentra-
tions of birds, and interspecific interactions may also
bias counts and should be considered in data analysis
(Fuller and Mosher 1981).

Because the detectable distance of a raptor var-
ies with species and habitat type, substantial biases
occur when comparisons are made among areas
or habitats. To compensate for this problem, survey-
ors can reduce strip width and only count birds
within a narrow width which reflects detection
(Millsap 1981 ). A better alternative is to record all
birds seen and their perpendicular distances from
the transect center line regardless of their distance,
and to analyze these data with the line transect com-
puter program TRANSECT (Burnham et al. 1980).
This program compensates for variable detection
distances and is available to all BLM employees on
the computer at the BLM Service Center. The main
problem with program TRANSECT is that the ideal
sample size is 2s 40 individuals (Burnham et al.
1980; Mikol 1980). Transects can be analyzed with
less than 40 individuals, but resolution is greatly
reduced. The problem can be compensated for by
(1) increasing transect length, (2) increasing the
number of repetitions, or ( 3 ) pooling species and
habitats (Millsap pers. commun.).

A criticism of road surveys is that transects usu-
ally are not random and data are not collected for
statistical analysis. However, appropriate methods
can reduce this problem. If many roads exist in a
large area of homogeneous habitat, investigators can

326

Raptors

randomly select transects from the array of available
roads. Often these conditions do not exist, and divid-
ing the transects into segments provides a viable
alternative. Transects should be established in habitat
and land use types that need to be surveyed. Tran-
sect segments can be delineated by the actual dis-
tance of specific habitat types along the transect,
resulting in segments of unequal length. These seg-
ments of specific habitats are conceptually linked
together and treated as one long transect. Data are
analyzed with line transect programs, and compari-
sons among years and habitat types can be made
(Mikol 1980).

Another method divides transects into equal 1.6
km (1.0 mi.) segments (Koplin pers. commun.). Each
segment is treated as an independent sample plot
with specific habitat, land use, and .visibility charac-
teristics. The entire transect is surveyed, and the
exact location of each bird is recorded during re-
peated runs. The mean number (and variance) of
birds per sample plot is calculated; comparisons are
made among years within and among habitats, as
well as among habitats within years. To test for
clumping, one can compare the number of plots
with various numbers of birds with a Poisson Model.
More simply, biologists can compute the coefficient
of dispersion (Sokal and Rohlf 1981 ) of the plots.

in woodland habitat, and 5 to 26% in heavily for-
ested habitat. However, when effective transect
width was adjusted to habitat specific detection dis-
tances within each vegetation type, accuracy aver-
aged 90% (Millsap pers. commun.).

Craighead and Craighead (1956:49) and Fuller
and Mosher ( 1981 ) discussed factors that affect ac-
curacy of road surveys. Results are more reliable for
determining relative densities of conspicuous species
(buteos, vultures, and eagles) with similar detectabil-
ity in open country of homogeneous habitat types.
Accuracy greatly decreases when many raptor spe-
cies are sampled in habitats that differ greatly in
vegetation and terrain. This gives relatively inflated
density estimates in open areas compared with esti-
mates in the more vegetated areas.

Although road surveys can be inaccurate, they
are precise. Coefficients of variation calculated from
seven surveys, which counted medium-large raptors
in a 0.4 km (0.25 mi.) strip in open to semi-open
habitats, ranged from 6 to 14% (Craighead and
Craighead 1956; Schnell 1967; Kochert et al. 1975,
1976; Craig 1978). Millsap and LeFranc (pers. com-
mun. ) found that their road transect data were more
precise when they adjusted effective transect width.

CD = Jl
Y

This value will be near 1 if the distribution is
essentially Poisson (random), >1 in clumped samples,
and < 1 in uniformly distributed samples. Habitat-spe-
cific sample plots can also be linked to form a long
conceptual transect and analyzed with the line tran-
sect program (Mikol 1980). The advantage of this
method is that each segment can be considered a
replication, thus increasing sample size. In addition,
sample plots can be randomly selected on a stratified
basis.

Accuracy and Precision. Road counts appear
to have variable accuracy which may be affected
by transect width. Craighead and Craighead (1956)
made repeated vehicle surveys during the winter
in a partially forested area in Michigan. They be-
lieved that 92% of the buteos, 43% of the northern
harriers, 33% of the Cooper's hawks, and 33% of the
American kestrels were located by using a belt of
0.4 km (0.25 mi.) on both sides of the road. In Vir-
ginia, B. Millsap and M. LeFranc (pers. commun.)
found much lower accuracy from preliminary results
of one summer-fall field season. Models of perched
raptors representing various size classes were ran-
domly distributed within a 0.4-km (0.25-mi.) strip on
either side of the road. By using this belt, the pro-
portion of models detected ranged (depending on
size) from 11 to 64% in grassland habitat, 2 to 32%

Equipment. No special equipment is required
except binoculars, spotting scope (minimum 20x),
thermometer, wind meter, watch, field forms, and
a vehicle. Biologists should use a range finder to
assist in estimating perpendicular distances from the
transect.

Cost. Road transects are inexpensive, and large
areas can be surveyed by two people in a short time.
Two people in the Snake River Birds of Prey Area
covered 112 km (70 mi.) of transect in 4 hours
(1 person-day). This is a reasonable estimate for a
survey day considering travel time to and from the
survey site.

Raptors

327

Training. At least one observer must be
trained in raptor identification and should be able to
identify color phases and age classes of those species
in the survey area. Short courses in raptor identifica-
tion are recommended.

Discussion. Road counts are a fast, inexpensive
means of surveying raptors over large areas. The
method, however, is ineffective for nest surveys and
monitoring owls and forest raptors. It is a reliable
means for assessing occurrence, species composition,
and age structure of populations of suitable species.
The technique is useful for assessing long-term popu-
lation trends through relative measures; however, it
is plagued with many biases (Fuller and Mosher
1981, 1986). For meaningful surveys, biologists must
work within the limitation of those biases, primarily
detectability and observer biases, and use only expe-
rienced observers on surveys.

Aerial Surveys.

Essentially two types of aerial surveys can be
used to census large raptors. Random transects are
used to survey raptors that are scattered throughout
a survey area. Searches of specific habitats (e.g., lakes
and riparian areas) are used to survey species that
congregate in specific habitats. These surveys are
used primarily to census wintering birds.

Description. Random aerial transects have
been used primarily for golden eagles and occasion-
ally for other large raptors (Fuller and Mosher
1981). Starting from a randomly selected section
line, investigators randomly select sample transects
from all possible transects that can be placed at fixed
0.8-km (0.5-mi.) intervals in the survey area (Wrak-
estraw 1972; BLM unpubl. data). Transects should be
of equal length (usually 80 km [50 mi.]) and sepa-
rated enough to avoid duplicate sightings (at least
6.4 km [4.0 mi.]). However, because of irregular
survey area boundaries, some investigators have used
transects of unequal lengths (Boeker and Bolen
1972; Craig 1974). They considered all transect lines
collectively as a large transect which was divided
into 80-km (50-mi.) segments. This approach is valid
only if the transects and the beginning point for the
first segment are randomly selected; however, biolo-
gists should avoid this approach because segments
may not be totally independent.

Transects should be flown at about 15 m (50 ft)
to 92 m (300 ft) above ground level (AGL) at
speeds between 160-190 km/h (100-120 mph)
(Boeker and Bolen 1972; Boeker 1974). All birds
seen on each side of the plane, regardless of dis-
tance, should be counted and their perpendicular
distance from the transect should be recorded. This
increases analytical flexibility, and data can be ana-
lyzed within a fixed belt (Boeker and Bolen 1972),
or by using line transect analysis (Mikol 1980).

Because birds are counted on both sides of the
aircraft, two observers and a pilot are necessary. The
right side observer sits in the right front seat and
searches from the center line out to the right side of
the aircraft, scanning both the sky and the ground.
The left side observer sits in the left rear seat of the
aircraft and searches the left side of the aircraft. To
reduce biases, surveys must be flown in as few con-
secutive days as possible, but days should not be
excessively long. After 3-4 hours of flying, fatigue
and eyestrain affect a person's ability to see birds
(Roseneau 1972); a survey day should be 6-7 hours
maximum of flying, interrupted with a necessary
1-hour rest period. A 1600-km (1000-mi.) survey is
most efficiently flown in 3 consecutive days.

Kochert et al. (1983) analyzed 3 years of golden
eagle transect data by using both the belt transect
method and line transect analysis (program TRAN-
SECT, Burnham et al. 1980). The belt method gave
consistently lower estimates of wintering population
levels than TRANSECT — they believed the results
from line transect analysis yielded a more realistic
view of the actual population size. Belt transects
typically underestimate population densities when
the size of the belt is too wide to detect all animals
within the belt (Franzbeb 1981); however, effective
transect width varies from survey to survey with
program transect, depending on probabilities of de-
tection (Burnham et al. 1980). Although TRANSECT
can be compromised by small sample sizes (see
Road Count section), it appears to be a reliable ana-
lytical method.

Aerial surveys of specific habitats are used pri-
marily for bald eagles, and biologists can either com-
pletely search habitat patches or randomly sample
them. In designing the initial inventory, they should
identify habitat areas from aerial photographs where

Aerial surveys can easily identify bald eagle nests in forest
habitats.

328

Raptors

birds are known to concentrate and should then
chart survey routes on maps. U.S. Fish and Wildlife
Service (1983) presents detailed guidelines for win-
tering bald eagle surveys and stresses the need for
subsequent surveys throughout the winter. Areas can
be surveyed by either helicopters or fixed-wing air-
craft at a recommended height of 31-93 m ( 100-300
ft) AGL and speed between 96-120 km/h (60-70
mph). The best time to survey feeding activity is 1-3
hours after daylight. Entire habitat areas should be
flown to count all birds present; large areas can be
covered by systematic parallel transects or concen-
tric contours of the habitat border (Hancock 1964).
Where potential bald eagle foraging area is vast, sur-
veys along systematic transect lines should be spaced
2.4 km (1.5 mi.) apart and extend no more than 27
km (15 mi.) from roosts (U.S. Fish and Wildlife
Service 1983).

Often areas are too large to survey all habitats.
In this situation, biologists should consider surveying
a random sample of quadrats of the area (King et al.
1972; Grier 1977; Grier et al. 1981). This method
provides population estimates with confidence limits
and allows statistical comparison of subsequent sur-
veys (Hodges et al. 1979). If the objective is to sur-
vey a large amount of linear habitat (i.e., a river),
surveyors should segment the linear habitat and ran-
domly sample the segments, preferably on a stratified
basis.

Gilmer et al. (1981) and Grier et al. (1981)
described aircraft certification, preflight preparations,
safety precautions, and special considerations for
aerial surveys. Surveys should be flown on clear days
with no precipitation. Though not mandatory, winds
should be less than 32 km/h (20 mph). Safety is a
prime concern, and only experienced pilots and
aircraft with sufficient power should be used. Heli-
copters should not be used in canyons and areas
with many powerlines. Pilot attitude and experience
can also influence error (Hoskinson 1976).

Accuracy and Precision. Caughley (1974)
discussed biases of aerial transect surveys. He be-
lieved that the method had moderate to low accu-
racy because many animals were not seen; accuracy
deteriorated progressively with increasing transect
width, cruising speed, and altitude. Although inaccur-
ate, transect surveys from fixed-wing aircraft are
precise. Precision (based on 95% confidence inter-
vals) of fixed-wing transect surveys in Wyoming and
Idaho for golden eagles ranged between 14.0 and
16.5% (Wrakestraw 1972, 1973; BLM unpubl. data).

Fixed-wing aircraft surveys of bald eagles in
specific habitats on flat terrain appear to be fairly
accurate. Steenhof ( 1976) surveyed an area for bald
eagles from both the ground and a fixed-wing aircraft

on the same day and obtained nearly the same re-
sults. Also, helicopter surveys may be more accurate
than fixed-wing surveys (U.S. Fish and Wildlife
Service 1983).

Slight biases may occur in assessing age struc-
ture of bald eagles from airplanes. Hancock ( 1964)
believed that perched adult bald eagles were more
conspicuous than immatures and assumed that he
underestimated immatures by 20-35%, compared
with only 10-15% for adults. These biases may not
be significant for golden eagle surveys because plu-
mage difference between adults and sub-adults are
not striking, and all perched birds must be flushed
and inspected at close range before they are aged.

Perched adult bald eagles are mice as easy to see as im-
mature ones.

Equipment. No special equipment is required
except for the aircraft. Certain aircraft may have
advantages over others. Readers should see the
Equipment section of Aerial Nesting Surveys dis-
cussed in this chapter. Some observers prefer to use
a small cassette tape recorder to record data. Others
carry binoculars; however, I have found them of
limited value for aerial surveys. Also, hearing protec-
tors can reduce fatigue.

Training. Observers must be able to identify
eagles and other raptors to species and age class. See
Training in the Road Transect section.

Cost. Cost can vary considerably depending on
type of survey and aircraft. In the Birds of Prey Area
it has taken 20 hours (including ferry time) for two
observers to survey 20 transects 80 km (50 mi.)
long (BLM unpubl. data). The same survey in a heli-
copter would take about 30 hours for two people.
Costs of surveying specific habitats can vary greaUy
depending on the amount of ferry time and aircraft
used.

Raptors

329

Discussion. Both random aerial transects and
aerial surveys of specific habitats provide a reliable
and expeditious means of assessing raptor presence,
including smaller conspicuous species over large
areas. Random aerial transects are an effective means
to count large raptors over broad areas as demon-
strated by the statewide survey of golden eagles
in Wyoming (Wrakestraw 1972, 1973). This method
is also reliable for assessing age structure of golden
eagles for population dynamics studies. Although the
method has inherent biases, they affect accuracy
more than precision (Caughley 1974), making it a
reliable means of assessing relative densities and
monitoring long-term population trends. Aerial sur-
veys of wintering habitat can be an accurate and
reliable means of surveying bald eagle concentra-
tions, and randomly sampled quadrats could be a re-
liable way to monitor numbers of wintering raptors
in large areas.

Weather, raptor activity, time of day, foliage
condition, and snow cover can affect raptor detecta-
bility and bias surveys. Biases can be reduced by
standardizing survey times, season, and weather con-
ditions. Because observer competence is a major
source of variability in survey results, the same expe-
rienced observers ) should conduct all surveys in a
particular area with the same pilot (U.S. Fish and
Wildlife Service 1983).

Counts at Roosts and Colonies.

All kite species in North America nest either
colonially or semi-colonially, and many falconiforms
and strigiforms use communal roosts (Newton 1979;
Allen and Young 1982). This section focuses on
techniques used to locate colonies or communal
roosts and to estimate bird numbers. Often the same
techniques are used to survey roosts and colonies.

Description. Before searching for roosts or
colonies, biologists should identify potential nesting
or roosting habitat from aerial photographs and vege-
tation maps. A common method of locating roosts
or colonies employs non-random ground or aerial
searches of these potential areas for concentrations
of raptors (Bildstein 1979; Fuller and Mosher 1986;
Glinski and Ohmart 1983). A few investigators, how-
ever, have searched for roosts or colonies by run-
ning systematically spaced transects (Sykes 1979).
Roost searches are most effective during the last 90
minutes of daylight (Keister 1981). Eagle roosts may
be identified by low level aerial photography. Roosts
or colonies can also be located by following birds
back from feeding areas to the roost or colony by di-
rect observation or with the aid of radio telemetry
(Keister 1981). The location of suspected roosts or
colonies can be obtained by observing birds from
a vantage point, and the exact location and relative
size of roosts can be determined from the distribu-
tion of feces and castings (Keister 1981).

The most common method of enumerating indi-
viduals in a roost or colony is by directly counting
the number of birds (Parker 1975; Steenhof 1976;
Keister 1981). Night roost counts for diurnal raptors
should be made between 60 minutes before sunset
to 30 minutes after sunset (Keister 1981). Evening
counts are usually more accurate than morning
counts (Hein 1961; U.S. Fish and Wildlife Service
1983). Most Mississippi kite nesting colonies are
small enough that investigators can enter the colony
and count the number of nesting attempts (Parker
pers. commun.); however, caution should be taken
to minimize disturbance.

Bald eagles are easily counted at winter roost concentra-
tions.

Sometimes roosts are inaccessible, or birds can-
not be seen (e.g., northern harriers), and biologists
must estimate numbers of individuals by counting
birds as they fly to and from roosts (Hein 1961; Bild-
stein 1979; U.S. Fish and Wildlife Service 1983).
This technique has an advantage in that it minimizes
disturbance. Hein (1961) gave detailed procedures
for conducting roost flight counts. Roosting flight
counts should be made from 90 minutes before sun-
set to 30 minutes after sunset, with a subsequent
count the next morning from 30 minutes before sun-
rise to 90 minutes after sunrise. The maximum num-
ber of the two counts should be used. Sometimes
several counters stationed around the roost are re-
quired when birds depart in different directions
(Weller et al. 1955).

Biologists sometimes use flight counts to census
nesting colonies (Fuller and Mosher 1981 ). Glinski
and Ohmart (1983) estimated Mississippi kite num-
bers in colonies by counting birds during courtship,
foraging, and predator-mobbing flights. They used
these numbers and counts of nesting attempts to
estimate the nonbreeding component. Accurate
counts, however, can only be made by many peri-
odic visits to the colony or roost starting early in the
season (Bildstein 1979; Parker and Ogden 1979).
Although long-eared owls can be counted at roosts,

330

Raptors

distribution, and relative size of colonies and roosts
are more important for most monitoring and inven-
tory programs than absolute counts or changes in
absolute numbers of individuals. Changes in numbers
of individuals may not reflect changes in the popula-
tion. Many colonial raptors tend to be nomadic
(Newton 1979), and changes in bird numbers in a
colony or roost may only reflect a local or regional
shift. To account for any inter-roost shifts, investiga-
tors should periodically make simultaneous counts at
all roosts in an area.

Because many factors influence counts at roosts
(e.g., weather and shifts in prey abundance), counts
should be considered indexes (Bildstein 1979; Fuller
and Mosher 1981; Keister 1981; U.S. Fish and Wild-
life Service 1983). Although direct counts are in-
dexes, and indirect counts (e.g., flight counts) may
underestimate the number of birds (U.S. Fish and
Wildlife Service 1983), both methods provide a reli-
able means of assessing relative abundance. Counts
at night roosts provide a more accurate index of
wintering bald eagle abundance than day counts on
feeding areas (U.S. Fish and Wildlife Service 1983).
Compared with other census techniques, roost and
colony counts are very reliable. As with any census
technique, there are problems with detectability, and
when communally roosting birds bunch up they are
difficult to count. However, accuracy of counting
methods can be increased by taking the average of
numerous repeated counts made during a sampling
period.

Long-eared owl.

Millsap (pers. commun. ) found that morning counts
of birds flying to the roost were very similar to di-
rect counts of birds in the roost. This method may
also work for short-eared owls. Some investigators
have counted these owls by walking through roosts
and flushing individuals (Clark 1975), but this is not
desirable because of disturbance.

Keister (1981) discussed other methods of
enumerating roosting eagles such as thermal infrared
remote sensing, radar, and infrared photography.
These methods have limited utility for a BLM moni-
toring or inventory program. However, in the Birds
of Prey Area, biologists have experimented with
standard photography using 1000 ASA print film to
count communally roosting ravens on transmission
towers and have had encouraging results (BLM un-
publ. data).

Discussion. This method is effective in assess-
ing presence of birds in specific areas. It is also use-
ful in population dynamics studies, and combined
with other surveys over a large area, it may also be a
means of assessing long-term trends. The number,

Relative Abundance — Nesting Surveys
General Considerations.

Survey Design. When designing an inventory,
investigators should gather all available relevant
historical information on the area to be surveyed
(Fuller and Mosher 1981). This information can be
obtained through the literature, records (e.g., Federal
and State agencies, museums, or the Laboratory of
Ornithology at Cornell University), and inquiries of
local residents, falconers, wildlife managers and
biologists, and amateur ornithologists. Because many
species have relatively restricted nest site require-
ments and use structures that last many years (e.g.,
ledges), historic data can lead one to specific nest
sites. Historic information can provide data on
species occurrence and distribution.

Inventory or monitoring objectives will deter-
mine the sampling approach. In certain unique habi-
tats or areas of special concern, like the Birds of
Prey Area, biologists may be required to design sys-
tematic searches of the entire area staggered over
the nesting period (U.S. Bureau of Land Management
1979). However, in most situations, investigators

Raptors

331

will not be able to survey the entire area, and the
most efficient approach in terms of time and people
is to sample the area in a simple random or stratified
random manner.

Biologists should consider quadrat sampling to
estimate populations over large areas (King et al.
1972; Grier 1977; Postovit 1979; Grier et al. 1981;
Gilmer and Stewart 1984). Quadrats are randomly
selected from grids on maps or aerial photos. Selec-
tions can be simple random or stratified to habitat
characteristics or quality. Investigators can survey
either the entire quadrat or just the primary nesting
habitat (e.g., shorelines or cliffs) within the quadrat.
Quadrat sampling provides population estimates with
confidence limits (Grier 1977), and quadrats can be
easily monitored with statistical comparisons among
surveys (Hodges et al. 1979). Surveys that include
more than primary habitat may be more precise
(Leighton et al. 1979).

To survey long stretches of linear habitat (e.g.,
cliff, shoreline, or riparian areas), biologists can di-
vide the habitat into equal segments and select seg-
ments to survey on a stratified random basis. Each
segment can be treated as a sample plot for statisti-
cal comparisons. Random sampling of segments gives
an inference of the entire habitat stretch sampled,
but non-randomly-sampled segments give informa-
tion only on the segment sampled.

Locating Nests and Nesting Areas. Successful
location of raptor nests or nesting areas requires
knowledge of the distribution, habitat relationships,
nesting chronology, spacing, and nesting behavior of
the species or group of species to be surveyed. This
knowledge and understanding can be derived from
several sources (Craighead and Craighead 1956; Call
1978; Newton 1979; and various technical notes
available from the BLM Service Center).

Typical ferruginous hawk nest on rock outcrop.

For a territory to be classified "occupied," it
must satisfy the criteria for occupancy (see Produc-
tivity section). There are a variety of visible signs
that advertise raptor occupancy and nesting attempts
(Craighead and Craighead 1956; Call 1978; Jones
1979; Fuller and Mosher 1981, 1986; Reynolds 1982;
Forsman 1983b). These signs differ with species and
nesting habitat. In cliff areas, rocks stained with feces
(whitewash) are the primary feature. Decorated or
repaired nests, perches, plucking perches, and regur-
gitated pellets are also important. In forested areas,
stick nests, perches, molted feathers, plucking
perches, and prey remains below a nest are common
signs of occupancy.

Behavioral observations are also important for
locating nests. These include courtship displays, food
carries, food transfers, copulation, nest building, ac-
tive defense and calling by adults, and food begging
by young (Craighead and Craighead 1956; Call 1978;
Jones 1979; Fuller and Mosher 1981; Reynolds 1982;
Mindell 1983)- A single observation of an unrepaired
nest does not mean the site is vacant (Fuller and
Mosher 1981). Nesting areas should be visited more
than once to confirm vacancy.

Biases. Many factors influence detection of
nests and nesting areas (Fuller and Mosher 1981,
1986). Visibility biases are influenced by light condi-
tions, time of day, weather, type of nesting habitat,
seasonal changes in the habitat, seasonal changes in
the behavior of raptors, and observer competence
and experience (Craighead and Craighead 1956;
Henny et al. 1974, 1977; Postovit 1979; Grier et al.
1981). Raptor behavior can bias observations; secre-
tive raptors that stay away from nests are easily
missed, whereas extremely aggressive birds are easily
found (Craighead and Craighead 1956; Call 1978;
Postovit 1979). Also, pairs arriving late and those
that have multiple broods (e.g., barn owls) or renest
can influence accuracy because they are often
missed.

Aerial Surveys.

Description. Boeker (1970), Hickman (1972),
Roseneau (1972), Grier (1977), Fraser (1978),
Leighton et al. (1979), and Grier et al. (1981)
described techniques for surveying cliff and tree
nesting raptors with fixed-wing aircraft. They
discussed types of aircraft, safety precautions, and
costs, and stressed the need to retain only
experienced pilots. Observers fly predetermined
routes and search for nests, presence of adults, or
other signs of occupancy. Surveys for conspicuous
bald eagle and osprey nests are flown relatively fast
at speeds between 112 and 160 km/h (70-100 mph)
at about 100 m (325 ft) AGL (Grier et al. 1981).
Cliff nesting surveys tend to be flown slower

332

Raptors

(between 72 and 112 km/h [45 and 70 mph]) at
distances of 15-62 m (50-200 ft) from the cliff
(Boeker 1970; Hickman 1972; Roseneau 1972).
Observers fly slightly below the rim, scan the cliff,
and search for evidence of occupancy. Often a
second pass is required to determine the nest
contents, and these approaches should be made
going "down hill" after checking air movement
(Boeker 1970). Additional passes in a tiered fashion
descending down the cliff are necessary to
completely cover large cliff faces.

White and Sherrod (1973) and Carrier and
Melquist (1976) described survey techniques and
safety precautions for surveying from a helicopter.
Techniques for locating nests are similar to those for
fixed-wing aircraft. A desired speed to survey cliffs
is 48-70 km/h (30-40 mph), and open nests on pow-
erline structures can be adequately surveyed at 80-
96 km/h (50-60 mph). Once a suspected nest is lo-
cated, observers can hover or slow to speeds less
than 32 km/h (20 mph) to provide a better view of
the nest. I have found that glare on the helicopter
bubble reduces one's ability to see nests; this can be
remedied by removing the door. White and Sherrod
(1973) advised approaching nests from upwind to
avoid hitting flushed birds; however, Carrier and
Melquist (1976) did not believe this was a problem.
Many times downwind approaches are necessary in
canyon surveys for hovering or maneuvering at slow
speeds. Biologists and pilots should always be alert
because occasionally birds will attack the aircraft or
fly in its path (Nelson pers. commun.).

Surveys can be systematic grids of an area, ran-
dom quadrat samples of an area, or random or non-
random surveys of nesting habitat (see General Con-
siderations). Systematically spaced transects are used
primarily to survey entire areas for tree-nesting spe-
cies (Petersen 1979). Both fixed-wing and helicopter
nest surveys should be made with at least two ob-
servers on calm (winds less than 24 km/h [15 mph])
days (Carrier and Melquist 1976; Grier et al. 1981).
Helicopters, however, are more stable and effective
in winds greater than 24 km/h (15 mph) (Call
1978). Surveys can be conducted at any time of day,
but mornings are usually the best time for canyon
surveys because of the reduced wind conditions.
Although Grier et al. ( 1981 ) found no difference in
visibility due to season, if possible, observers should
survey tree-nesting raptors before foliation in decid-
uous forests. Also, possible errors can be avoided
by observers recording observations independently.

Accuracy and Precision. Two types of error
characterize aerial nesting surveys: failure to locate
occupied sites and failure to properly classify them.
Estimates are conservative at best because some
birds and nests are not seen, and visibility biases are

common and often unavoidable (Caughley 1974;
Grier 1977).

Accuracy varies with species surveyed, observ-
ers, and number of flights in each survey. Henny
et al. (1974, 1977) calculated that only about 64%
of the occupied osprey sites on various substrates
were located in a single, fixed-wing flight. Accuracy
for similar single-flight surveys of cliff-nesting golden
eagles was estimated at 81% (Phillips et al. 1984).
Accuracy of single flight fixed-wing surveys for occu-
pied breeding areas of tree nesting bald eagles in
three areas ranged between 84 and 85% (Leighton
et al. 1979; Grier et al. 1981; Fraser et al. 1983).
Accuracy increased to 98% by adding a second flight
(Grier et al. 1981). Although differences were not
significant, the proportion of bald eagle tree nests
found by experienced observers ranged from 67 to
87% (Grier et al. 1981 ). Specific data are lacking for
the smaller, less conspicuous raptors (gyrfalcons,
rough-legged hawks, and red-tailed hawks). Accuracy
of surveys for these species may be lower than those
for eagles, considering accuracy in locating osprey
pairs (Henny 1974, 1977) and raven tree nests
(12%; Grier et al. 1981).

Little has been published on accuracy of nest
surveys from helicopters. In a single-flight survey,
Wier ( 1982) located seven of the eight (88% )
golden eagle, rough-legged hawk, and gyrfalcon nest-
ing attempts on a cliff in Alaska. Birds of Prey Area
personnel located 96% of a sample of 48 golden-
eagle nesting attempts during a single flight (BLM
unpubl. data). Accuracy of occupancy classifications
may be related to the proportion of pairs laying eggs.
In a single-flight occupancy survey for cliff-nesting
golden eagles during incubation, Birds of Prey Area
personnel correctly classified 88% of 77 sites when
only 67% of the pairs laid eggs; however, all sites
were correctly classified during a year when all pairs
bred (BLM unpubl. data). Most of the misclassified
nesting areas contained nonbreeding pairs.

Timing and classification criteria can also influ-
ence accuracy of aerial surveys. Surveys for occu-
pancy are most accurate after completion of all
clutches, but before the first brood hatches (Fraser
et al. 1983; Steenhof 1986). Fraser (1978) increased
accuracy in classifying bald eagle breeding areas for
occupancy from 85% to 89% when criteria required
that only one adult instead of two be seen in the
breeding area. However, sites where one bird is ob-
served must satisfy the criteria for occupancy (see
Productivity section), and biologists must note
whether one or two birds were seen on their sur-
veys so data can be compared with other studies. If
they find no evidence of egg laying, no freshly deco-
rated nests, and only a single adult, the site should
be revisited, preferably on the ground, to verify oc-
cupancy (Steenhof 1986).

Raptors

333

Fixed-wing aerial surveys are fairly precise for
locating occupied bald eagle territories (Grier 1977;
Leighton et al. 1979). However, the precision of
extrapolating population estimates from random aer-
ial survey blocks appears low. Leighton et al. (1979)
calculated that their aerial surveys of random blocks
could not detect population changes of less than
27% . Surveyors can refine precision by stratified ran-
dom sampling (Grier 1977). However, Grier and
Hamilton (1978) observed no reduction in variance
when they subsampled clusters of random samples
on a stratified basis using optimum allocation of sam-
ples. Fraser (1978) found that when he included a
single adult bald eagle at the nest, or single adults or
pairs away from the nest in his criteria for site occu-
pancy, it increased accuracy but reduced precision.

Equipment. Often local availability will deter-
mine the type of aircraft to be used. Smaller fixed-
wing aircraft (e.g., Piper Super Cub) may be better
for raptor nest surveys than larger ones (Cessna 180
or 182) because smaller craft are slower, more ma-
neuverable, and offer better observer visibility (Hick-
man 1972; Roseneau 1972; Grier et al. 1981). For
helicopter surveys, the smaller jet-turbine aircraft
(e.g., Hiller/Soloy) may be more effective because of
lower hourly cost and the ability to hover in unsta-
ble air conditions. However, if surveys involve a
large amount of ferry distance, the faster medium-
sized helicopters (e.g., Bell Jet Ranger) are more
efficient.

Cost. Depending on the aircraft, fixed wing
surveys cost 1/3 (Cessna 180) to 1/10 (Piper Super
Cub) of an equivalent ground survey. A Super Cub
took 25% less survey time per linear mile (1.6 km)
of cliff to locate golden eagle nests than did the
Cessna 180 and 98% less time than ground surveys
(Boeker 1970; Hickman 1972). Quadrat surveys
for tree nesting bald eagles covered 150 km" (58
mi. )/h using a Cessna 180 and 71 km (28 mi. )/h
with a Piper Super Cub (King et al. 1972; Grier
1977; Hodges et al. 1979). By cluster sampling, Grier
and Hamilton (1978) reduced their flight time by
15%. Considering salaries, travel, and per diem, heli-
copter surveys cost about 5% less than that of equiv-
alent ground surveys, and they were completed 10
times faster (Carrier and Melquist 1976; BLM un-
publ. data).

Discussion. Aerial surveys are most successful
with large conspicuous raptors that have high visibil-
ity, limited nesting habitat, and synchronized nesting
cycle (Henny et al. 1977; Call 1978; Leighton et al.
1979). Although aerial nesting surveys have been
successful for ospreys, gyrfalcons, rough-legged
hawks, and goshawks (McGowan 1975; Call 1978;
Fuller and Mosher 1 98 1 ), they are most accurate for
bald and golden eagles. Accuracy appears to de-
crease for smaller, less conspicuous raptors. Al-
though peregrine falcons, snowy owls, great horned

owls, and red-tailed hawks have also been surveyed
by aircraft, these surveys have usually supplemented
ground searches (Fuller and Mosher 1981). The
technique has limited value on cavity-nesting raptors
(e.g., prairie falcons, owls; Call 1978).

Aerial surveys can be more effective than
ground surveys because they cost less, they are an
expeditious means of surveying large areas, and they
provide ready access to remote country. Investiga-
tors, however, must consider many factors when
choosing an aircraft. Although more costly in com-
parison, helicopters provide more visibility (they can
hover to allow a better view of nests), better maneu-
verability and stability, and in many instances,
greater accuracy than fixed-wing aircraft. Further-
more, helicopters can land in the field for ground in-
spection of nests.

Aerial surveys have many biases and limitations
(Fuller and Mosher 1981, 1986). Detectability and
classification biases are most important, and can be
reduced by supplementing flights with follow-up
ground searches or flying additional surveys. Also,
problems with precision can be reduced with larger
sample sizes. The technique is fairly repeatable, and
if biologists work within the limitations of these
biases, it can be a reliable means of surveying large
areas.

Calling Surveys.

Detection of calls has been used primarily to
survey owls, but recently the technique has been ap-
plied to some falconiforms (Fuller and Mosher
1 98 1 ). Every North American owl species, except
the common barn-owl, snowy owl, short-eared owl,
and northern hawk-owl, has responded to human
broadcasted owl vocalizations (Foster 1965; Martin
1973; Call 1978; Fuller and Mosher 1981; Hayward
1983). Falconiforms presently known to respond
to these calls are the red-shouldered hawk, red-tailed
hawk, broad-winged hawk, goshawk, Cooper's hawk,
sharp-shinned hawk, common black-hawk, gray hawk,
and zone-tailed hawk (Fuller and Mosher 1981; Ro-
senfield et al. 1985; Mosher et al. unpubl. ms.; Mill-
sap pers. commua).

Description. Two basic methods are used to
locate raptors by their calls. The first is to listen for
calls and note their location. This method, used
mainly to supplement other survey techniques, en-
tails walking or driving through an area and periodi-
cally listening for calls. The other method is to elicit
responses either by imitating calls or broadcasting
recordings of calls; responses can be either return
calling or the silent approach of a bird. This method
involves periodic broadcasting of calls followed by
periods of silence during which surveyors look and
listen for responses (Forsman et al. 1977; Fuller and
Mosher 1981; Forsman 1983b).

334

Raptors

Raptors can be surveyed by the continuous tran-
sect method (Forsman et al. 1977; Forsman 1983b)
or the calling station method (Fuller and Mosher
1981). In the continuous transect method, investiga-
tors walk along a road and play recorded calls at
fixed intervals (often every 15-20 sec). After a bird
responds, surveyors move a fixed distance (often 0.8
km or 0.5 mi.) down the road (presumably out of
the responding owl's territory) before calling again.
A relative density estimate of the number of pairs
per linear distance is calculated from this method.

Is it live
or
I s it Memorex ?

The calling station method involves broadcast-
ing calls from established stations. Stations can be
non-randomly placed in specific habitat types or
placed at fixed distances to ensure full coverage of
an area (Siminski 1976; Smith 1978; Springer 1978).
Stations can be systematically placed along a road
or transect (sometimes called point count transects;
Simpson 1972; Ellison 1980; Mosher et al. unpubl.
ms.), along parallel transects to form a grid (Cink
1975), or in quadrats randomly selected from a grid
(Nowicki 1974). Calls are usually broadcast for 15
seconds followed by a 1 5-60 second listening period,
after which the sequence is repeated. Investigators
usually call at each stop for 4 to 1 5 minutes. Mosher
et al. (unpubl. ms.) recommended that observers
look and listen for an additional 5 minutes before
proceeding to the next stop. Also at each stop they
broadcast to one side of the road, turned the speaker
180 degrees and broadcast again, and repeated this
process until three recordings had been broadcast to
each side of the road. In contrast, Balding and Dibble
(1984) played calls through four speakers each
pointed in a different direction. Relative densities
derived by this method can be reported as the num-
ber of birds per calling station, number of birds per
linear distance, or number of birds in an area.

Call surveys should be conducted under rela-
tively calm conditions with no precipitation or fog
(Simpson 1972; Forsman et al. 1977; Smith 1978;
Springer 1978; Forsman 1983b). Minimum data to be
recorded at each calling stop are weather (at least
temperature, precipitation, wind speed, and cloud
cover), moon phase, terrain, vegetation, and raptor
response. If a response is elicited, investigators
should record species, approximate location, time of
response, type of response (vocal or visual), habitat
type, and sex (if distinguishable). However, inexperi-
enced observers may not be able to distinguish be-
tween sexes.

Most researchers call owls at night, usually be-
tween 30 minutes after sunset to 1 hour before sun-
rise (Bell 1964; Simpson 1972; Nowicki 1974;
Forsman et al. 1977; Hayward 1983); however, Sim-
inski (1976) and Springer ( 1978) successfully called
great horned owls during daylight. Little information
exists on when owl calling is most successful. Some
workers believe time of night has no effect (Cink
1975); however, nesting barred owls appeared most
responsive near the middle of the night (Smith
1978), and wintering eastern screech-owls re-
sponded more towards dawn (Beatty 1977).

Calling for nesting diurnal raptors is usually
conducted between 30 minutes after sunrise to mid-
morning, and birds may be more responsive toward
mid-morning (Balding and Dibble 1984; Mosher et
al. unpubl. ms.; Millsap pers. commun.). Responses to
calls appear to be immediate, mostly within 5 min-
utes of broadcasting (Rosenfield et al. 1985; Mosher
et al. unpubl. ms.). Although hawks respond to calls
during all stages of the nesting cycle ( Mosher et
al. unpubl. ms.), Cooper's hawks appeared to be less
responsive during incubation (Rosenfield et al.
1985).

If the objective is to estimate the number of
birds present in an area, it is important to know the
range over which the sample is being taken. How-
ever, biologists should avoid using the absolute
measure, number of birds per unit area, because not
all birds respond to calls. However, some researchers
have calculated number of birds responding per
unit area by dividing the area contained in an arbi-
trarily delineated study area by the number of birds
responding to calls (Siminski 1976; Springer 1978;
Smith 1978). This measure is meaningless if the en-
tire area is not covered by vocalizations. Conversely,
it may overestimate the density if it draws birds in
from outside the study area. To estimate the number
of birds in a large area, Nowicki (1974) randomly
selected quadrats from the area and calculated a
density based on the number of owls responding,
divided by the area contained in all sample quadrats.
Total number of owls were estimated by multiplying
this figure by the size of the total study area. This
method is valid if quadrats are randomly selected,

Raptors

335

the entire area of the quadrat is covered by the calls,
and no owls from adjacent quadrats are drawn in
and respond to the calls.

Some researchers estimated the distance their
broadcast calls carried and used this "effective listen-
ing distance" to determine the size of their sampling
area (Cink 1975). This approach has many biases.
Vegetation and topography influence the distance
that calls carry, and effective listening distances
based on human detection may be inaccurate be-
cause of differences between raptor and human hear-
ing (Nowicki 1974; Fuller and Mosher 1981).
However, one should know the audio range and
effectiveness of equipment. Mosher et al. (unpubl.
ms.) measured sound levels from their equipment
and suggested that vocalizations would be audible to
about 800 m ( 2640 ft ), at which distance sound
levels were similar to background noise (30-40 dB);
however, they believed calls could carry up to 1.6
km (1.0 mi.).

Sometimes owls are difficult to locate, and when
many birds simultaneously respond to the calls, indi-
viduals or pairs are difficult to differentiate. Also
many owls, particularly flammulated owls, have ven-
triloquial abilities which make it extremely difficult
to pinpoint the source of the call without considera-
ble effort (Millsap pers. commun. ). To more pre-
cisely locate calling barred owls, Bell (1964)
triangulated from two stations. When western
screech-owl and elf owl nests were concentrated ( 50
m [165 ft] apart) in limited riparian zones, Johnson
et al. (1979) placed at least three people 50-100
m (165-330 ft) apart to differentiate among simulta-
neously responding pairs. At best, locations of re-
sponding owls are estimates, and often a reasonable
estimate is extremely difficult. This problem is com-
pounded because birds responding to calls may ap-
proach the caller beyond the limits of their normal
home range or territory (Reynolds pers. commun.).
This can impose great limitations when assigning
habitat types based on call responses.

Accuracy and Precision. Although certain
raptors respond to calls year-round (Foster 1965),
vocalization surveys are most reliable during the
breeding season (Fuller and Mosher 1981). Not all
individuals will respond to calls (Nowicki 1974;
Beatty 1977; Johnson et al. 1979; Balding and Dibble
1984). Siminski (1976), Forsman et al. (1977), and
Springer ( 1978) found that from 75 to 85% of the
nesting spotted owls and great horned owls in their
study areas responded to broadcasted calls. Male
owls appear to be more responsive to broadcasted
calls than females (Siminski 1976; Springer 1978).
Results on accuracy of this method on diurnal rap-
tors are preliminary (Mosher pers. commun.). How-
ever, Mosher et al. (unpubl. ms.) found a significant
positive correlation between the number of nesting

diurnal pairs and the number of responses to re-
corded calls of conspecifics. To increase accuracy,
they recommended calling on ten different occasions
during the nesting season and establishing more and
longer transects when surveying large areas.

Equipment. Necessary equipment consists of a
cassette tape player, amplifier, speaker, and cassette
tape recordings of calls. Surveyors have used many
kinds of tape recorders, amplifiers, and speakers that
have ranged from simple units to sophisticated com-
ponent systems. Examples are the Model 600 BM
Game Tape Caller (Springer 1978), Sony Model TC
100 Tape Recorder (Smith 1978), and Marantz
C-205 Tape Recorder with 10-in. (25-cm), 8-ohm
trumpet speakers (Mosher et al. unpubl ms.). Regard-
less of equipment, its audible range should be no
less than 0.4 km (1/4 mi.). Mosher et al. (unpubl.
ms. ) recommended a tape player with a frequency of
about 40 Hz to 12 kHz and a power output of about
1.2 watts at 1 kHz. Some owls have local dialects
which could make them less responsive to calls of
the same species from a different geographic region
(Siminski 1976; Fuller and Mosher 1981). Surveyors
should try to use calls recorded in the geographic
region where the survey is being conducted.

Taped vocalizations are available from the fol-
lowing sources:

Laboratory of Ornithology
Cornell University
Sapsucker Woods
Ithaca, NY 14850

National Audubon Society
Western Education Center
376 Greenwood Beach Road
Tiburon, CA 94920

National Geographic L.P. Records
National Geographic Society
17th & M Streets NW
Washington, DC 20036

Oregon State Office

Bureau of Land Management

P.O. Box 2965

Portland, OR 97208 (for spotted owl tapes)

Training. Investigators must be thoroughly
familiar with the repertoire of calls ( including any
differences between sexes ) of the species they are
surveying. Training for both new and experienced
surveyors is essential. Listening to recorded calls
from the many field guide series is the simplest rep-
ertoire of the species (Johnson et al. 1979). It may
be necessary to obtain taped calls from an expert on
a particular species.

336

Raptors

Cost. Vocalization surveys are relatively inex-
pensive to conduct, and fairly large areas can be
surveyed in a short period of time. Forsman et al.
(1977) surveyed an average of 12.8 km (8.0 mi.) of
road per night using the continuous transect
method. Ellison (1980) averaged 12 calling stations
per night. Using the procedure recommended by
Mosher et al. (unpubl. ms.), it is possible for one
person to complete 16 calling stations or cover 12.8
km (8.0 mi.) of transect in 4 hours if all stations
are easily accessible.

Discussion. Vocalization surveys result in sig-
nificantly more bird contacts than simply looking
and listening for inconspicuous raptors. They greatly
reduce the time required to find birds and may be
the only reliable means to locate owls and forest-
inhabiting hawks. Although responsiveness varies
with species, the method is reliable for assessing spe-
cies occurrence (Mosher pers. commun.). Results
suggest that calling from point counts can be useful
in developing indexes to local populations and per-
haps for estimates of density (Mosher et al. unpubl.
ms.). Accuracy, however, is reduced when the
method is used to count individuals in a given area.
Estimates of birds in an area should only be used if it
is certain that the broadcast calls covered the sample
area. Otherwise, a more realistic measure of relative
density would be number of birds per linear distance
(Forsman et al. 1977) or contacts per station
(Ellison 1980).

Although the method is used to assess relative
abundance of nesting raptors, it is valuable in supple-
menting road counts and other types of nest

searches (Fuller and Mosher 1981). Biologists use
nocturnal calling to obtain the general location of
spotted owl territories and diurnal calling to specifi-
cally locate roosts and nests (Forsman 1983b). To
locate diurnal raptor nests, biologists concentrate
their searches in areas where most contacts occur
(Mosher et al. unpubl. ms.) or search in the direc-
tion of departing birds after a contact has been made
(Rosenfield et al. 1985). Whenever possible, survey-
ors should supplement call surveys with ground
searches for birds or nests. These searches are essen-
tial to confirm breeding attempts, and they also serve
to assess survey accuracy.

Investigators should be cautious when using
broadcast calls alone to assess relative densities be-
cause several factors contribute to the difficulty in
estimating numbers, and many variables can influ-
ence the reliability of a vocal census. Fuller and
Mosher (1981) discussed the main variables, includ-
ing time of year, time of day, behavioral differences,
and differences in vocalizations and in responses
between sexes. Both sexes of the great horned owl,
spotted owl, screech-owl, and fiammulated owl re-
spond to calls, and there may be sexual differences
in call responses (Marshall 1939; Siminski 1976;
Forsman et al. 1977; Springer 1978; Smith 1978;
Johnson et al. 1979). They demonstrate sexual differ-
ences in the pitch of their calls; the male call is soft-
er and lower. Also, one cannot assume if a male
responds to a call that a pair is occupying a territory,
because many responding males may not be paired,
and unpaired male spotted owls and fiammulated
owls appear more responsive during the nesting
season (Reynolds pers. commun.).

Spotted owl.

Great horned owl.

Raptors

337

Equipment and environmental variables can also
affect response rates (Fuller and Mosher 1981).
Background noise produced by the equipment or the
environment (e.g., traffic noise) may interfere with
efficient transmission and should be avoided if possi-
ble. Effects of wind on sound transmission and the
observers ability to hear the calls vary with the habi-
tat type but have less of an influence in open habi-
tats. Steep terrain, thick vegetation, and running
water can greatly restrict the audible distance of the
call or the observer's ability to hear a response. Re-
sponse rates may be affected by the lunar cycle, and
owls may be more responsive during a bright, wax-
ing moon (Johnson et al. 1979). Ellison (1980) ad-
justed his systematic calling sequence monthly to
avoid synchronizing calling station visits with the
moon phase.

Surveyors need to understand ranges and disper-
sion of target species pairs. If call stops are too far
apart, some birds may be missed, and if too close,
the same pair may be counted twice. Mosher et al.
(unpubl. ms.) recommended stations spaced at 0.8
km (0.48 mi.) for larger raptors; however, points can
be only 0.25 km (790 ft) for smaller raptors such as
screech-owls. It is worthwhile to test the audio
range of equipment on raptors in a known location
in the vegetational and topographic areas to be
sampled.

Certain factors should be considered when
broadcasting calls. Johnson et al. ( 1981 ) believed
that if western screech-owls were censused too often
they could become habituated to broadcasts and be
less responsive; however, Mosher et al. (unpubl. ms.)
saw no indication of this with diurnal raptors. When
calling different species from the same station, Call
(1978) recommended that calls of the smallest spe-
cies be played first because response behavior of the
smaller species may be inhibited by the vocalization
of a larger owl. However, some researchers do not
recommend calling different species from the same
station because of the possible biases (Mosher pers.
commun.).

Ground Surveys.

I include searches by foot, horseback, boat, and
land vehicle (e.g., automobile, motorcycle, bicycle)
in this category. The methods involve similar tech-
niques, allow time for close inspection, and are often
used to complement each other.

Description. Biologists should be concerned
with collecting two types of data during ground
surveys: the general location of a pair (i.e., nesting
territory) and the exact nest location. The latter
requires more time and effort, and slightly different

338

Raptors

techniques. In most monitoring programs it is not
necessary to determine the exact nest location.
Therefore, biologists should only be concerned with
obtaining general locations of pairs unless specific
management needs require the exact locations of
nests. Biologists should use multiple techniques; one
negative observation does not necessarily mean that
a site is vacant. Subsequent visits and checks for
signs of occupancy should be made before conclud-
ing vacancy.

General Considerations) or defending adults is re-
quired to determine the exact nest location. To lo-
cate scrapes, some workers have flushed falcons by
clapping hands or throwing rocks. These techniques
should be used with caution because they disturb
birds, and they may not be totally reliable because
some birds do not flush when disturbed (BLM un-
publ. data). Sometimes the only way to locate nests
involves many hours of observing the cliff and
watching for adults or young.

Surveyors must be aware of the potential haz-
ards from disturbance caused by nest searches. Pre-
cautions are discussed by Fyfe and Olendorff ( 1976),
Call (1978), and Steenhof and Kochert (1982). Al-
though investigators need to locate nests before
hatching, they must observe them from a safe dis-
tance and avoid entering nests and disturbing parent
birds until the young can thermoregulate (Steenhof
and Kochert 1982).

Craighead and Craighead (1956), Call (1978),
Jones (1979), Reynolds (1982), and Forsman
(1983b) described survey techniques for forest-
nesting raptors. To determine general nest locations
in deciduous forests, biologists often canvass wooded
areas before foliation in the spring and look for stick
nests. These searches are often supplemented by
aerial surveys (Petersen 1979). Observations of aer-
ial courtship displays of raptors above the canopy
yield the general nest location for signs of occu-
pancy (see General Considerations) and defending
adults are required to locate the nest. Workers some-
times locate pairs and follow them to nests (Village
1984). Searches for forest-nesting raptors, especially
cavity nesting owls, are facilitated by calling surveys
(Forsman 1983b; Fuller and Mosher 1986). Occu-
pancy of suspected cavities by smaller raptors may
be confirmed by rapping the tree and observing the
bird at the cavity entrance (Call 1978); however,
this does not necessarily mean there is a nest in it.
Also, this method may not be reliable because cer-
tain species, such as flammulated owls, are seldom
brought to the cavity entrance by rapping (Reynolds
pers. commun. ). Exact nest locations of spotted owls
have been determined by baiting the birds with a
tethered mouse and following the owl with prey
back to its nest (Forsman 1983b).

Call (1978) and Mindell (1983) discussed sur-
vey techniques for cliff-nesting raptors; decorated
stick nests, whitewash on the cliff, and behavior of
adult birds are the most commonly used clues. Cliffs
should be surveyed early in the nesting season to
obtain general locations of breeding pairs and to
locate stick nests. These searches are often later aug-
mented by aerial surveys. If the nest or scrape is
not located during the early survey, a thorough
search of the general location for field signs (see

Searches for ground nesters sometimes require
unique techniques. General nesting locations of har-
riers and other ground-nesting raptors are obtained
by observing consistent nesting behavior (i.e., court-
ship flights, territorial defense, copulation, prey ex-
changes, and nest building) in an area; however,
behavior must be interpreted with caution because
unmated harriers sometimes exhibit courtship behav-
ior (Call 1978; Thompson-Hanson 1984). Exact nest
locations can be determined by observing the female
flying consistently to the same spot with prey during
the brood-rearing period or by systematically search-
ing a general nesting location, sometimes with a
trained dog (Call 1978; Thompson-Hanson 1984).
Burrowing owl nests are located by observing
perched owls, prey remains, and perches in areas
with burrows (Call 1978).

Accuracy and Precision. Little published in-
formation exists on accuracy and precision of
ground surveys for nesting raptors; however, accu-
racy varies with species, habitat, and methods. In
most instances, foot surveys are most accurate be-
cause observers are able to see inconspicuous nests
(Fuller and Mosher 1981), and accuracy increases
with each subsequent survey (Craighead and Craig-
head 1956; Postovit 1979). As with the other meth-
ods, surveys are most accurate after completion of
all clutches, but before the first brood hatches. Vil-
lage (1984) showed a 20-30% decrease in accuracy
when European kestrel (Falco tinnunculns) nests
were visited on foot late in the season because he
missed nonbreeding pairs and those which failed
early. Craighead and Craighead (1956) used both
foot and vehicle surveys of a deciduous forest area;
they believed that 85% of all raptor nests were eas-
ily found and the remaining 1 5% required numerous
repeated searches. Olendorff (1975) believed that
by using repeated vehicle and foot surveys he lo-
cated over 95% of the raptor pairs nesting on trees
and cliffs in a grassland area. Large conspicuous nests
are easily found; Hodges and Robards (1982) be-
lieved that a trained observer could locate 90% of
the bald eagle nests within 200 m (660 ft) of the
shoreline in a single pass in a boat and accurately as-
sess nesting status of 90% of the nests as well. Accu-
racy for boat surveys of cliff-nesting falcons would
probably be much less.

Raptors

339

Village (1984) presented two sources of error
in assessing nesting density: (1) study areas are nqt
chosen at random, and (2) nests may be more easily
missed in large study areas than small ones. This
problem is compounded because clusters of sight-
ings, normally recorded as one pair, may be two
pairs. Biases can be so large that they may have
masked any variation in breeding density.

Precision of ground surveys can vary greatly
depending on size of the sampling unit, sampling in-
tensity, and number of species surveyed. Postovit
(1979) refined precision of raptor surveys of a 223
km (93 mi. ) area 2.6 times (from 63 to 24% based
on 95% confidence intervals) by reducing sample
quadrat size from 2.56 km to 0.65 km (1.02 mi. to
0.26 mi. ) and increasing sampling intensity from
10 to 33% of the study area. However, surveys still
lacked sensitivity, and population changes less than
39% could not be detected. Because precision may
be increased by sampling a large proportion of an
area in a short period using a correction factor for
visibility bias (Postovit 1979), other more intensive
surveys may be more precise. Stratified random sam-
pling apparently does not increase precision when
surveying a collection of raptor species (Postovit
1979), but it increases precision when sampling for
a single species that occurs in restricted habitats
(Leighton et al. 1979; Grier et al. 1981 ).

Equipment. Binoculars (at least 7x35) and a
spotting scope (minimum 20x) with a sun shade are
essential.

Training. Observer competence is one of the
many factors that influence accuracy of nesting sur-
veys. Survey manuals (e.g., Call 1978) and short
courses are helpful; however, novice biologists
should survey with an experienced biologist until
they have developed a satisfactory "search image."

Cost. Cost will vary depending on species and
accessibility, and ruggedness and remoteness of the
terrain. Village (1984) believed that one full-time
worker could properly survey about 100 km2 (39
mi. ) of reasonably accessible habitat for European
kestrels during a nesting season. When travel costs
and salaries are considered, ground surveys cost
nearly the same as equivalent helicopter surveys and
more than fixed-wing surveys.

be surveyed quickly by automobile, and boats pro-
vide ready access to remote areas. However, these
surveys have some error, and questionable nesting
areas should be reinspected on foot.

Although ground surveys may be the most accu-
rate survey method for nesting raptors, they have
certain limitations and biases (Postovit 1979; Steen-
hof and Kochert 1982; Steenhof 1986; Fuller and
Mosher 1986). Observer and visibility biases are
most significant, and because of possible poor preci-
sion, especially with sampling of large areas, there
may be problems using single surveys to monitor
nesting populations. Biases can be reduced and pre-
cision increased by repeated surveys of sample areas,
increased sample size, and by using only experi-
enced surveyors. However, ground surveys are time-
consuming, and time is one of the main limitations
of the technique. This problem can be reduced by
augmenting ground surveys with more expedient
methods (e.g., aerial surveys).

Productivity Surveys

Data Collection and Analysis.

General Considerations. Data on raptor
reproductive rates are useful to managers for
assessing the status and reproductive health of a
raptor population. Information required, biases, and
timing of productivity surveys are briefly discussed
in this section. Postupalsky (1974), Steenhof and
Kochert (1982), Fraser et al. (1983), and Steenhof
(1986) discussed them in more detail.

Productivity Data. Clutch size and number of
young hatched usually are not necessary for produc-
tivity surveys; however, the following basic informa-
tion is used for estimates of reproductive success
and production (Steenhof 1986):

( 1 ) Number of pairs in the area This refers to
the number of occupied "breeding areas" or
"nesting territories" (Postupalsky 1974;
Steenhof 1986). Occupancy classification
requires at least one of the following observa-
tions during the nesting season:

Discussion. Ground surveys are most effective
for inconspicuous secretive raptors. This method,
especially on foot, is useful for intensive searches of
small areas of specialized habitat; however, these
searches are the most time-consuming of all. Surveys
on foot or horseback may be more accurate than
those by boat or automobile because the former al-
low for closer inspection. These techniques, how-
ever, should complement each other. Large areas can

(a) Evidence that an egg was laid (i.e., obser-
vation of an incubating bird, eggs, egg-
shell fragments, young, or a decorated
nest with fresh prey and feces).

( b ) Observation of two breeding- age birds
that appear to be paired within a nesting
territory.

340

Raptors

(c) Observations of one or more birds at-
tending a nest, engaging in reproductive
behavior, or defending an area.

Some researchers disagree with the last crite-
rion, but observation of one adult attending a nest or
defending an area usually constitutes enough evi-
dence that a pair is present (Steenhof 1986). Also
the presence of a freshly decorated nest usually is
evidence of occupancy for species such as golden ea-
gles and red-tailed hawks. Caution must be used,
however, because of the difficulty in distinguishing
new nesting material from old, and because winter-
ing birds in some areas decorate nests and then
move elsewhere (Steenhof 1986). Areas where only
decorated nests are observed should be revisited
on the ground to confirm occupancy.

(2) Number of breeding pairs. This is the num-
ber of pairs that lay eggs (see 1 above); how-
ever, this information is not mandatory.
Useful information can be obtained without
knowing the number of breeding pairs
(Steenhof and Kochert 1982). Biologists
should remember that observations of raptors
near a nest are not always evidence of a
breeding attempt (Fuller and Mosher 1981);
nonbreeding pairs sometimes decorate or
repair nests. Some birds may also assume in-
cubation posture without having laid an egg;
but because of the difficulty of distinguishing
these situations, birds observed in incubation
position should be considered "breeding"
for analysis (Steenhof 1986).

(3) Number of successful pairs. These are pairs
that raise at least one young to "acceptable"
fledging age which is defined as 80% of the
average age that most young leave the nest
(Steenhof and Kochert 1982; Steenhof 1986).
Full counts of broods at fledging are not
necessary.

(4) Number of young to reach acceptable fledg-
ing age. The number of young fledged per
pair is the most important measure of pro-
ductivity (Steenhof 1986). This value can be
calculated directly from all known pairs or
a random sample of preselected pairs in an
area. It can also be estimated indirectly by
combining independent estimates of percent-
age breeding, percentage success, and num-
ber fledged per successful nest (Steenhof and
Kochert 1982; Steenhof 1986).

Red-tailed hawk on nest.

Biases. As with other surveys, two broad types
of error can bias productivity surveys: measurement
error and sampling error (Steenhof 1986). Measure-
ment error occurs when data are incorrectly col-
lected or interpreted because of the technique,
visibility problems, or observer incompetence.
Proper timing of surveys can reduce measurement
error (Fraser et al. 1983). Also, by calculating a
measurement error rate and an estimated standard
error, true differences in productivity can be tested
where all breeding pairs are known and are repeat-
edly sampled in a given year (Fraser et al. 1984).

Sampling error occurs when the pairs sampled
do not represent the population. Biases can be re-
duced by randomly selecting pairs (either preselect-
ing known pairs or sampling areas), using only pairs
found during or before incubation, or using the May-
field Model (Mayfield 1961) to calculate success
(Steenhof and Kochert 1982). An unbiased estimate
of percentage of pairs breeding can be obtained only
from samples of traditional pairs selected before the
nesting season (Steenhof and Kochert 1982). Only
these nests and those found during incubation
should be used for calculating reproductive success
and productivity; inclusion of pairs found late yields
inflated estimates. The Mayfield Model allows work-
ers to maximize use of data sets with small sample
sizes by including nests found late in the season
(Steenhof 1986). A computer program is available to
handle the tedious calculations required by the May-
field approach (contact the Birds of Prey Research
Staff, Boise District, BLM for details).

Raptors

341

Timing of Surveys. For single species surveys,
a minimum of two surveys are required (Fraser et
al. 1983), and more may be necessary (Steenhof
1986). The first survey, to count occupied territories
and pairs with eggs, should be done during incuba-
tion, ideally after completion of all clutches but be-
fore hatching of the first brood. A middle survey may
be necessary if there is a wide latitude in nesting
chronology. It should be conducted after the oldest
broods reach acceptable fledging age but have not
left the nest. The function of this survey is to assess
fledging of the oldest broods and to age the remain-
ing young to establish the optimal time of the third
survey. The last or "productivity" survey should
be conducted after the youngest brood reaches ac-
ceptable fledging age but before the oldest chicks,
who were less than acceptable fledging age on the
middle survey, leave the nest. For multiple-species
surveys in the Birds of Prey Area, investigators found
it necessary to conduct a minimum of four surveys.
Because of the wide range in chronology, a second
occupancy survey is necessary to compensate for the
wide latitude in starting dates (BLM unpubl. data).

Survey Methods
Aerial Surveys.

Description. Techniques for assessing
productivity from aircraft are the same as those used
during aerial nesting surveys; however, the former
requires a closer inspection of the nest contents.

Accuracy and Precision. Aerial productivity
surveys are fairly accurate for large conspicuous
raptors. An average 82% of the breeding pairs were
correctly classified and 86% of the young were seen
during fixed-wing surveys of nesting bald eagles
(Fraser et al. 1983).

In the Birds of Prey Area, accuracy of helicopter
surveys for cliff nesting golden eagles was 86% for
assessing breeding status and 95% for counting
young (BLM unpubl. data). Accuracy can be en-
hanced by adjusting survey timing. By properly tim-
ing surveys, Fraser et al. (1983) increased accuracy
for both estimates by about 12%. Accuracy in the
Birds of Prey Area was also increased with properly
timed helicopter surveys.

Poole ( 1981 ) found fixed-wing surveys of os-
prey breeding pairs and fledged young were as accu-
rate as ground counts. However, accuracy may be
lower for smaller, less conspicuous raptors. During
helicopter surveys in the Birds of Prey Area, investi-
gators have been unable to obtain accurate brood
counts at some red-tailed and ferruginous hawk nests
because the parents covered their young (BLM un-
publ. data).

Discussion. Aircraft provide a fast and expedi-
tious way to collect productivity data for large con-
spicuous raptors, especially if surveys are optimally
timed. Multiple productivity surveys are often re-
quired and additional flights further increase accu-
racy (Grier et al. 1981 ). Aerial surveys can be
reliable for monitoring eagle productivity, especially
if all breeding sites are known and surveyed at the
proper time (Fraser et al. 1984). Lastly, helicopters
cost more than fixed-wing aircraft, but they are usu-
ally more accurate because they can hover to allow
more opportunity to correctly count broods.

Ground Surveys.

Description. Techniques for assessing
productivity from ground surveys are the same as
those used for nesting surveys except exact location
of the nest and inspection of its contents are
required. Inspections can be accomplished by distant
observations with the aid of telescopes or binoculars;
close observations from a nearby vantage point, or
with the aid of a mirror attached to a pole (Parker
1972); or by entering nests (Call 1978; Steenhof
1986).

Discussion. The usefulness of methods for
inspecting nests depends on survey purpose and en-
vironmental conditions (Steenhof 1986). Distant
observations are necessary and adequate to confirm
an incubating bird, but less useful for brood counts,
especially if the entire nest cannot be seen. Climbing
into the nest is undoubtedly the most accurate
method to obtain productivity data. However, it is
extremely time-consuming and causes disturbance to
the birds. Close inspections and distant observations
require less time but are subject to error. Foliage
and topographic features often do not allow full view
of the nest or complete brood counts; Millsap (pers.
commun.) minimized this error and used a 15 m
(49.5 ft) range pole and mirror to count eggs and
young in 1 50 tree nests. He spent an average of 3
minutes at each nest with very little disturbance and
high accuracy.

SYSTEMS FOR CORRELATING HABITAT
VARIABLES WITH POPULATION
MEASUREMENTS

Habitat models provide useful information for
management biologists. They can be used to predict
potential species occurrence and to evaluate poten-
tial habitat quality. The models are also useful in
designing an inventory. Biologists are cautioned from
using these models alone as a population measure
or as a management tool. They still need to inven-
tory the area of concern by using the techniques
described in this chapter to gather the information
necessary for management decisions.

342

Raptors

Climbing into raptor nests is the most accurate method of obtaining productivity information.

Few habitat models have been published. I will
discuss those that are available or being developed.
The U.S. Fish and Wildlife Service (FWS), Office of
Biological Services' Habitat Evaluation Procedures
(HEP) Group in Fort Collins, Colorado, has devel-
oped Habitat Suitability Index Models for the ferrugi-
nous hawk, prairie falcon, sharp-shinned hawk,
golden eagle, northern spotted owl, short-eared owl,
burrowing owl, and northern harrier. The only pub-
lished model (Jasikoff 1982) is on ferruginous hawks;
others are in draft form. Model narratives describe
habitat relationships for the raptor. The FWS's Divi-
sion of Ecological Services in Anchorage, Alaska, has
also developed a draft habitat model for bald eagles.
Ellis (1982) developed a system for evaluating the
potential for peregrine falcon habitat in Arizona.
It provides a good means for predicting possible oc-
currence, and it may be applied to other habitats
in the arid West. Mosher et al. ( 1985) developed a
management model to predict nesting habitat for
eastern woodland hawks. Lastly, Stalmaster (1983)
developed a detailed energetics model for bald ea-
gles wintering in the Pacific Northwest that deter-
mines eagle carrying capacity in salmon spawning

drainages and estimates use of salmon by eagles. This
model also provides broad scale management
recommendations.

DISCUSSION

Raptor inventory and monitoring programs are
frequently limited because raptors are often difficult
to detect and count. Raptors often occur at relatively
low densities in a diversity of habitats, and it is ex-
tremely laborious to obtain complete counts of nest-
ing pairs. Estimates usually must be obtained through
sampling, and adequate sampling is often costly and
arduous.

Design and implementation of a successful rap-
tor inventory or monitoring program is affected by
sampling, measurement, and interpretation problems.
Many sampling problems affect estimates of numbers
and population changes. Although reliable for assess-
ing abundance, the usefulness of many indirect tech-
niques (e.g., calling surveys) is limited if studies do
not relate the sample to the actual population size

Raptors

343

(Fuller and Mosher 1981). Often relative indexes are
of limited value because surveys are not conducted
in a standardized manner or samples are not ade-
quately collected for statistical analysis. However,
these problems can be eliminated by proper sam-
pling procedures. Real problems occur when logis-
tics prevent implementation of a proper sampling
design, when sampling intensity is inadequate to
allow the necessary comparisons to be made, or
when samples are biased and do not represent the
population.

Accurate measurement of raptor populations is
also compromised by many biases. The greatest bias
is detectability; not all birds can be found or seen.
Next is observer competence. Every technique dis-
cussed here is affected by these two biases. Their
effects can be reduced by adequate sampling inten-
sity and by using only experienced observers. Envi-
ronmental factors (e.g., weather) and biological
variability (e.g., changes in food or bird behavior)
also cause error, and each technique has its own
inherent measurement biases.

Interpretation problems usually stem from in-
complete knowledge of how raptors interact with
their environment. For example, lack of knowledge
of the effects of climate and food supply hampers

one's ability to distinguish true population changes
from regional shifts in distribution. Interpretation
errors also result from a poor selection of estimators.
Some variables may remain the same when the popu-
lation actually changes. For example, production of
successful pairs often will show no change when
productivity of the population actually has changed
(BLM unpubl. data). To conduct a meaningful inven-
tory and monitoring program, biologists must use
those features that best reflect the population (Fuller
and Mosher 1986; Steenhof 1986).

These problems can be reduced by using effi-
cient, standardized techniques suitable for use with a
diversity of raptor species and habitats. Standardiza-
tion of estimated variables, including measures of
variability (e.g., confidence limits, standard error), is
as important as standardization of survey techniques
(Fraser 1978). This allows valid comparisons to be
made among areas.

Through an understanding of the inherent biases
of the methods used to measure raptor populations,
surveys can be designed and implemented to give
results of optimal accuracy and precision. By recog-
nizing the limitations of the resultant data and the
necessity for careful interpretation, it is possible to
reach conclusions that are useful and effective inputs
to management decisions.

344 Raptors

LITERATURE CITED

ALLEN, H.L. and L.S. YOUNG. 1982. An annotated bibliog-
raphy of avian communal roosting. Washington Dep.
Game, Olympia. 177pp.

AMERICAN ORNITHOLOGISTS UNION. 1983- Check list
of North American birds, 6th ed. Allen Press.
Lawrence, KS. 877pp.

ANTHONY, R.G., R.L. KNIGHT, G.T. ALLEN, BR. MCCLEL-
LAND, and J.I. HODGES. 1982. Habitat use by nesting
and roosting bald eagles in the Pacific Northwest.
Trans. North Am. Wildl. Nat. Resour. Conf. 47:332-
342.

APFELBAUM, S.I. and P. SEELBACH. 1983. Nest tree, habi-
tat selection and productivity of seven North Ameri-
can raptors based on the Cornell University nest
record card program. Raptor Res. 17:97-113-

BALDING, T. and E. DIBBLE. 1984. Responses of red-
tailed, red-shouldered, and broad-winged hawks to
high volume playback recordings. Passenger Pigeon
46:71-75.

BAMMANN, A.R. 1975. Ecology of predation and social
interactions of wintering white-tailed kites. M.S. The-
sis, Humboldt State Univ. Areata, CA. 81pp.

BEATTY, W.H. 1977. Attracting screech owls. Redstart
44:102-104.

BECHARD, M.J. 1982. Effect of vegetative cover on forag-
ing site selection by Swainson's hawk. Condor 84:153-
159.

BECKER, DM. 1984. Reproductive ecology and habitat
utilization of Richardson's merlins in southeastern
Montana. M.S. Thesis, Univ. Montana, Missoula. 62pp.

and I.J. BALL. 1983a. Prairie falcon {Falco tnexi-

canus). Pages 138-153 in Armbruster, J.S. ed. Impacts
of Coal Surface Mining on 25 Migratory Bird Species
of High Federal Interest. FWS/OBS-83/85. U.S. Dep.
Inter., Fish and Wildl. Serv. Fort Collins, CO.

and I.J. BALL. 1983b. Merlin (Falco columbarius).

Pages 124-137 in Armbruster, J.S. ed. Impacts of
Coal Surface Mining on 25 Migratory Bird Species of
High Federal Interest. FWS/OBS-83/85. U.S. Dep. In-
ter., Fish and Wildl. Serv., Fort Collins, CO.

BEDNARZ, J.C. and J.J. DINSMORE. 1982. Nest-sites and
habitat of red-shouldered hawks and red-tailed hawks
in Iowa. Wilson Bull. 94:31-45.

BELL, R.E. 1964. A sound triangulation method for count-
ing barred owls. Wilson Bull. 76:292-294.

BILDSTEIN, KL. 1978. Behavioral ecology of red-tailed
hawks (Buteo jamaicensis), rough-legged hawks (B.
lagopus), northern harriers (Circus cyaneus), Ameri-
can kestrels (Falco sparverius ) and other raptorial
birds wintering in south central Ohio. Ph.D. disserta-
tion. Ohio State Univ., Columbus. 364pp.

1979. Fluctuations in the number of northern

harriers (Circus cyaneus hudsonius) at communal
roosts in south central Ohio. Raptor Res. 1 3:40-46.

BOEKER, EL. 1970. Use of aircraft to determine golden
eagle (Aquila chrysaetos) nesting activity. Southwest.
Nat. 15:136-137. "

1974. Status of golden eagle surveys in the west-
ern states. Wildl. Soc. Bull. 2:46-49-

and E.B. BOLEN. 1972. Winter golden eagle popu-
lations in the Southwest. J. Wildl. Manage. 36:477-
484.

BOXALL, PC. and P.H.R. STEPNEY. 1982. The distribution
and status of the barred owl in Alberta, Can. Field-
Nat. 96:46-50.

BROWN, D., C. LOWE, and C. PASE. 1980. Digitized sys-
tematic classification for ecosystems with an illus-
trated summary of the natural vegetation of North
America. Gen. Tech. Rep. RM-73- U.S. Dep. Agric, For.
Serv., Fort Collins, CO. 93pp.

BROWN, L.H. and D. AMADON. 1968. Eagles, hawks and
falcons of the world. McGraw-Hill. New York, NY.
945pp.

BULL, EL. and R.G. ANDERSON. 1978. Notes on flammu-
lated owls in northeastern Oregon. Murrelet 59:26-27.

BURNHAM, KP., DR. ANDERSON, and J.L LAAKE. 1980.
Estimation of density from line transect sampling
of biological populations. Wildl. Monogr.72. 202pp.

CADE, T.J. 1982. The falcons of the world. Cornell Univ.
Press. Ithaca, NY. 188pp.

CALL, M.W. 1978. Nesting habitats and surveying tech-
niques for common western raptors. Tech. Note 316.
U.S. Dep. Inter., Bur. Land Manage. Serv. Cen. Denver,
CO. 115pp.

CARRIER, WD. and WE. MELQUIST. 1976. The use of a
rotor-winged aircraft in conducting nesting surveys of
ospreys in northern Idaho. Raptor Res. 10:77-83-

CAUGHLEY, G. 1974. Bias in aerial survey. J. Wildl. Man-
age. 38:921-933.

CINK, C.L. 1975. Population densities of screech owls in
northeastern Kansas. Kansas Ornithol. Soc. Bull.
26:13-16.

CLARK, R.J. 1975. A field study of the short-eared owl
(Asio flammeus pontoppidan) in North America.
Wildl. Monogr. No. 47. 67pp.

CLARKE, R.G. 1982. Nest site selection by sharp-shinned
hawks in interior Alaska. Pages 155-162 in Ladd, W.N.
and P.F. Schempf, eds. Raptor Management Biology
in Alaska and Western Canada. FWS/AK/PROC-82. U.S.
Dep. Inter., Fish and Wildl. Serv. Anchorage, AK

COTTRELL, M.J. 1981. Resource partitioning and repro-
ductive success of three species of hawks (Buteo sp. )
in an Oregon prairie. M.S. Thesis, Oregon State Univ.,
Corvallis. 72pp.

CRAIG, G.R. 1974. Raptor populations and characteristics
studies. Job Prog. Rep. No. W- 1 24-R- 1 . Colorado Div.
Wildl., Denver. 36pp.

CRAIG, T.H. 1978. A car survey of raptors in southeastern
Idaho 1974-1976. Raptor Res. 12:40-45.

CRAIGHEAD, J.J. and EC. CRAIGHEAD. 1956. Hawks,
owls, and wildlife. Wildl. Manage. Inst. Washington,
DC. 443pp.

ELLIS, D.H. 1982. The peregrine falcon in Arizona: habitat
utilization and management recommendations. Rep. 1.
Inst. Raptor Stud. Oracle, AZ. 24pp.

ELLISON, P.T. 1980. Habitat use by resident screech owls
(Otus asio). M.S. Thesis, Univ. Massachusetts, Am-
herst. 86pp.

EVANS, D.L. 1983. Ferruginous hawk (Buteo regalis).
Pages 109-123 in Armbruster, ed. Impacts of Coal
Surface Mining on 25 Migratory Bird Species of High
Federal Interest. FWS/OBS-83/85. U.S. Dep. Inter., Fish
and Wildl. Serv. Fort Collins, CO.

FITZNER, R.E. 1980. Behavioral ecology of the Swainson's
hawk (Buteo swainsoni) in Washington. PNL-2754
UC-11. Battelle Pac. Northwest Lab. Richland, WA.
15pp.

Raptors

545

FORBES, J.E. and D.W. WARNER. 1974. Behavior of a
radio-tagged saw whet owl. Auk 91:783-795.

FORSMAN, ED. 1983a. Spotted owl (Strix occidentalism
Pages 243-255 in Armbruster, J.S. ed. Impacts of Coal
Surface Mining on 25 Migratory Bird Species of High
Federal Interest. FWS/OBS-83/85. U.S. Dep. Inter., Fish
and Wildl. Serv. Fort Collins, CO.

. 1983b. Methods and materials for locating and

studying spotted owls. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. PNW-162. 8pp.

, EC. MESLOW, and M.J. STRUB. 1977. Spotted owl

abundance in young versus old-growth forests, Ore-
gon. Wildl. Soc. Bull. 5:43-47.

and H.M. WIGHT. 1984. Distribution and biology

of the spotted owl in Oregon. Wildl. Monogr. 87.
64pp.

FOSTER, F. 1965. An early reference of the technique of
owl calling. Auk 82:651-653.

FRANZREB, KE. 1981. The determination of avian densi-
ties using the variable-strip and fixed-width transect
survey methods. Stud. Avian Biol. 6:139-145.

and R.D. OHMART. 1978. The effects of timber

harvesting on breeding birds in a mixed coniferous
forest type. Condor 80:431-441.

FRASER, J.D. 1978. Bald eagle reproductive surveys: accu-
racy, precision, and timing. M.S. Thesis. Univ. Minne-
sota, Minneapolis. 82pp.

, L.D. FRENZEL, J.E. MATHISEN, F. MARTIN, and

M.E. SHOUGH. 1983. Scheduling bald eagle reproduc-
tion surveys. Wildl. Soc. Bull. 11:13-16.
-, F MARTIN, L.D. FRENZEL, and J.E. MATHISEN.

1984. Accounting for measurement errors in bald
eagle reproduction surveys. J. Wildl. Manage. 48:595-
598.

FULLER, M.R. 1979- Spatiotemporal ecology of four sym-
patric raptor species. Ph.D. dissertation. Univ. Minne-
sota, Minneapolis. 220pp.

. 1983- Zone-tailed hawk (Buteo albonotatus).

Pages 98-108 in Armbruster, J.S. ed. Impacts of Coal
Surface Mining on 25 Migratory Bird Species of High
Federal Interest. FWS/OBS-83/85. U.S. Dep. Inter., Fish
and Wildl. Serv. Fort Collins, CO.

and J. A. MOSHER. 1981. Methods of detecting and

counting raptors: a review. Stud. Avian Biol. 6:235-
246.

and . 1986. Raptor survey techniques in

Millsap, B.A. and K.W. Kline, eds. Raptor Management
Techniques Manual. Vol. 1. Natl. Wildl. Fed. Washing-
ton, DC.

and JR. PARTELOW. 1983. Cooper's hawk. (Accipi-

ter cooperii). Pages 74-97 in Armbruster, J.S. ed.
Impacts of Coal Surface Mining on 25 Migratory Bird
Species of High Federal Interest. FWS/OBS-83/85.
U.S. Dep. Inter., Fish and Wildl. Serv., Fort Collins,
CO.

FYFE, R.W. and R.R. OLENDORFF. 1976. Minimizing the
dangers of nesting studies to raptors and other sensi-
tive species. Can. Wildl. Serv. Occas. Pap. 23. Can.
Wildl. Serv. Ottawa, Ont. 17pp.

GILMER, D.S., L.M. COWARDIN, R.L. DUVAL, L.M. MECH
UN, C.W. SHAIFFER, and V.B. KUECHLE. 1981. Proce-
dures for the use of aircraft in wildlife bio-telemetry
studies. Resour. Publ. 140. U.S. Dep. Inter., Fish and
Wildl. Serv., Washington, DC. 19pp.

and R.E. STEWART. 1984. Swainson's hawk nesting

ecology in North Dakota. Condor 82:12-18.

GLINSKI, R.L. and R.D. OHMART. 1983- Breeding ecology
of the Mississippi kite in Arizona. Auk 85:200-207.

GRIER, J.W. 1977. Quadrat sampling of a nesting popula-
tion of bald eagles. J. Wildl. Manage. 41:438-443.

, J.M. GERRARD, G.D. HAMILTON, and PA. GRAY.

1981. Aerial visibility bias and survey techniques
for nesting bald eagles in northwestern Ontario. J.
Wildl. Manage. 45:83-92.

and G.D. HAMILTON. 1978. Aerial census of bald

eagles in the West Patricia Planning Area. West Patri-
cia Land Use Plan. Wildl. Tech. Rep. 7. Ontario Min.
Nat. Resour., Ottawa. 25pp.

GROSSMAN, ML. and J. HAMLET. 1964. Birds of prey of
the world. Bonanza Books, New York, NY. 496pp.

HANCOCK, D. 1964. Bald eagles wintering in the south-
ern Gulf Islands, British Columbia. Wilson Bull.
76:111-120.

HAYWARD, G.D. 1983 Resource partitioning among six
forest owls in the River of No Return Wilderness,
Idaho. M.S. Thesis. Univ. Idaho, Moscow. 127pp.

HEGDAL, P.L. and R.W. BLASSKIEWICZ. 1984. Evaluation
of the potential hazard to barn owls of TALON (bro-
difacoum bait) used to control rats and house mice.
Environ. Toxicol. Chem. 3:167-179.

HEIN, E. 1961. Wood duck roosting flights at Paint Creek,
Iowa. Proc. Iowa Acad. Sci. 68:264-270.

HENNY, C.J., MM. SMITH, and V.D. STOTTS. 1974. The
1973 distribution and abundance of breeding ospreys
in the Chesapeake Bay. Chesapeake Sci. 15:125-133.

, MA. BYRD, J.A. JACOBS, P.D. McLAIN, MR. TODD,

and B.F. HALLA. 1977. Mid-Atlantic Coast osprey
population: present numbers, productivity, pollutant
contamination, and status. J. Wildl. Manage. 41:254-
265.

HICKMAN, G.L. 1972. Aerial determination of golden
eagle nesting status. J. Wildl. Manage. 36:1289-1292.

HODGES, J.I., J.G. KING, and F.C. ROBARDS. 1979. Re-
survey of the bald eagle breeding populations in
southeast Alaska. J. Wildl. Manage. 43:219-221.

and F.C. ROBARDS. 1982. Observations of 3850

bald eagle nests in southeast Alaska. Pages 37-46 in
Ladd, W.N. and P.F. Schempf, eds. Raptor Management
and Biology in Alaska and Western Canada. U.S. Dep.
Inter., Fish and Wildl. Serv. FWS/AK/Proc. 82. Anchor-
age, AK

HOSKINSON, R.L. 1976. The effect of different pilots on
aerial telemetry error. J. Wildl. Manage. 40:137-139.

JASIKOFF, T.M. 1982. Habitat suitability index models:
ferruginous hawk. U.S. Dep. Inter., Fish and Wildl.
Serv. FWS/BS-82/10.10. Washington, DC. 18pp.

JOHNSON, DR. and WE. MELQUIST. 1973. Unique, rare,
and endangered raptorial birds of northern Idaho:
nesting success and management recommendations.
Publ. No. Rl -73-021. Univ. Idaho and U.S. Dep. Agric,
For. Serv. Moscow, ID. 42pp.

JOHNSON, R.R., B.T. BROWN, L.T. HAIGHT, and J.M.

SIMPSON. 1981. Playback recording as a special avian
censusing technique. Stud. Avian Biol. 6:68-75.

, L.T. HAIGHT, and J.M. SIMPSON. 1979. Owl popu-
lations and species status in the southwestern United
States. Pages 40-59 in Schaeffer, P.P. and S.M. Ehlers,
eds. Owls of the West: Their Ecology and Conserva-
tion. Natl. Audubon Soc. Tiburon, CA.

JONES, S. 1979. Habitat management series for unique or
endangered species. Report No. 17 - The accipiters
goshawk, Cooper's hawk, sharp-shinned hawk. Tech.

346

Raptors

Note 335. U.S. Dep. Inter., Bur. Land Manage. Denver,
CO. 51pp.

KEISTER, G.P., Jr. 1981. An assessment of bald eagle com-
munal roosting in northwestern Washington. Wash.
Dep. Game, Olympia. 39pp.

and R.G. ANTHONY. 1983- Characteristics of bald

eagle communal roosts in the Klamath Basin, Oregon
and California. J. Wildl. Manage. 47:1072-1079.

KING, J., F.C ROBARDS, and C.J. LENSINK 1972. Census
of the bald eagle breeding population in southeastern
Alaska. J. Wildl. Manage. 36:1292-1295.

KOCHERT, M.N. 1980. Golden eagle reproduction and
population changes in relation to jackrabbit cycles:
implications to eagle electrocutions. Pages 71-86
in Howard, R.P. and J.F. Gores, eds. A Workshop on
Raptors and Energy Development. Idaho Chapter
Wildl. Soc, Boise.

, A.R. BAMMANN, R.P. HOWARD, J.H. DOREMUS, M.

DELATE, and D. DONAHUE. 1975. Reproductive
performance, food habits and population dynamics of
raptors in the Snake River Birds of Prey Natural Area.
Pages 1-50 in Snake River Birds of Prey Res. Proj.
Ann. Rep. U.S. Dep. Inter., Bur. Land Manage. Boise,
ID.

, A.R. BAMMANN, J.H. DOREMUS, M. DELATE, and J.

WYATT. 1976. Reproductive performance, food habits
and population dynamics of raptors in the Snake
River Birds of Prey Natural Area. Pages 1-57 in Snake
River Birds of Prey Res. Proj. Ann. Rep. U.S. Dep.
Inter., Bur. Land Manage., Boise, ID.
-, J.H. DOREMUS, K STEENHOF, DR. DUNCAN,

M.Q. MORITSCH, AC. DOLDE, DM. RAMIREZ, S.A.
ADAMS, and D. DELSORDO. 1983. Density and repro-
ductive performance of raptors in the Snake River
Birds of Prey Area. Pages 6-15 in Snake River Birds of
Prey Res. Proj. Ann. Rep, U.S. Dep. Inter., Bur. Land
Manage. Boise, ID.

KUCHLER, A.W. 1964. Potential natural vegetation of the
conterminous United States. Am. Geogr. Soc. Spec.
Publ. 36. Am. Geogr. Soc. New York, NY. 1 16pp.

LEIGHTON, FA., J.M. GERRARD, P. GERRARD, D.W.A.

WHITFIELD, and W.J. MAHER 1979. An aerial census
of bald eagles in Saskatchewan. J. Wildl. Manage.
43:61-69.

LEIN, MR. and G.A. WEBER. 1979. Habitat selection by
wintering snowy owls (Nyctea scandiaca). Can. Field-
Nat. 93:176-178.

MADER, W.J. 1976. Biology of Harris' hawks in southern
Arizona. Living Bird 14:59-85.

. 1978. A comparative study of red-tailed hawks and

Harris' hawks in southern Arizona. Auk 95:327-337.

MARION, W.R. and R.A. RYDER. 1975. Perch-site prefer-
ences of four diurnal raptors in northeastern Colo-
rado. Condor 77:350-352.

MARKS, J.S. 1984. Nest site characteristics, reproductive
success and food habits of long-eared owls in south-
western Idaho. M.S. Thesis, Univ. Montana, Missoula.
91pp.

and I.J. BALL. 1983- Burrowing owl {Athene cuni-

cularia). Pages 227-242 in Armbruster, J.S., ed. Im-
pacts of Coal Surface Mining on 25 Migratory Bird
Species of High Federal Interst. U.S. Dep. Inter., Fish
and Wildl. Serv. FWS/OBS-83/85. Fort Collins, CO.

MARSHALL, J.R., Jr. 1939. Territorial behavior of the flam-
mulated screech owl. Condor 41:71-78.

MARSHALL, J.T. 1967. Parallel variation of North and
Middle American screech owls. Monogr. 1 . West
Found. Vert. Zool. Los Angeles, CA. 72pp.

MARTI, CD. 1979. Status of barn owls in Utah. Pages 29-
35 in Schaeffer, P. and S.M. Ehlers, eds. Owls of the
West: Their Ecology and Conservation. Natl. Audubon
Soc. Tiburon, CA.

MARTIN, D.J. 1973. Selected aspects of burrowing owl
ecology and behavior. Condor 75:446-456.

MATRAY, P.F. 1974. Broad-winged hawk nesting and ecol-
ogy. Auk 91:307-324.

MAYFffiLD, E.G. 1961. Nest success calculated from expo-
sure. Wilson Bull. 73:255-261.

MEEHAN, R.H. and R.J. RITCHIE. 1982. Habitat require-
ments of boreal and hawk owls in interior Alaska.
Pages 188-196 in Ladd, W.N. and P.F. Schempf, eds.
Raptor Management and Biology in Alaska and West-
ern Canada. FWS/AK/Proc. 82. U.S. Dep. Inter., Fish
and Wildl. Serv., Anchorage, AK

MELQUIST, W.E. 1975. Eagles and osprey: abstracts and
discussion. Page 142 in Murphy, J. R., CM. White, and
B.E. Harrell, eds. Population status of raptors. Proc.
Conf. Raptor Conserv. Tech. Raptor Res. Rep. 3- Rap-
tor Res. Found. Vermillion, SD.

MIKOL, S.A. 1980. Field guidelines for using transects to
sample nongame bird populations. FWS/OBS-80/58.
U.S. Dep. Inter., Fish and Wildl. Serv. Fort Collins, CO.
26pp.

MILLSAP, B.A. 1981. Distributional status of falconiforms
in west central Arizona: with notes on ecology, repro-
ductive success and management. Tech. Note 355.
U.S. Dep. Inter., Bur. Land Manage. Serv. Cen. Denver,
CO. 102pp.

and S.L. VANA. 1984. Distribution of wintering

golden eagles in the eastern United States. Wilson
Bull. 96:692-701.

MINDELL, D.P. 1983. Nesting raptors in southwestern
Alaska: status, distribution, and aspects of biology.
Tech. Rep. 8. U.S. Dep. Inter., Bur. Land Manage.
Anchorage, AK 59pp.

MOSHER, J.A., K TITUS, and MR FULLER. 1985. Develop-
ing a practical model to predict nesting habitat of
woodland hawks in Verner, J., ML Morrison, and C.J.
Ralph, eds. Modeling Habitat Relationships of Terres-
trial Vertebrates. Univ. Wisconsin Press, Madison.

, MR FULLER, and M. KOPENY. (Unpubl. ms.). Sur-
veying woodland raptors: I. Responsiveness to broad-
cast of conspecific vocalizations. (To be submitted
to J. Wildl. Manage.).

-, and M. KOPENY. (Unpubl. ms). Surveying

woodland raptors: II. Density and distribution of
nesting hawks. (To be submitted to J. Wildl.
Manage. ).

McCRARY, M.D. 1981. Space and habitat utilization by
red-shouldered hawks {Buteo lineatns elegans) in
southern California. M.S. Thesis, California State Univ.,
Long Beach. 85pp.

McGOWAN, J.D. 1975. Distribution, density and productiv-
ity of goshawks in interior Alaska. Fed. Aid Wildl.
Restoration Proj. Rep. W 17-3, 4, 5, 6. Alaska Dep.
Fish & Game, Juneau. 55pp.

NERO, R.W. 1980. The great gray owl. Smithsonian Inst.
Press, Washington, DC. 167pp.

NEWTON, I. 1979. Population ecology of raptors. Buteo
Books. Vermillion, SD. 399pp.

Raptors

347

NICHOLLS, T.H. and D.W. WARNER. 1972. Barred owl
habitat use as determined by radiotelemetry. J. Wildl.
Manage. 36:213-224.

NOWICKI, T. 1974. A census of screech owls using tape-
recorded calls. Jack-Pine Warbler 52:98-101.

OLENDORFF, R.R. 1975. Population status of large raptors
in northeastern Colorado 1970-1972. Pages 185-295
in Murphy, J.R., CM. White, and BE. Harrell, eds.
Proc. Conf. Raptor Conserv. Tech. Raptor Res. Rep. 3.
Raptor Res. Found. Vermillion, SD.

PARKER, J.W. 1972. A mirror and pole device for examin-
ing high nests. Bird-Banding 43:216-218.

. 1975. Populations of the Mississippi kite in the

great plains. Pages 159-172 in Murphy, JR., CM.
White, and B.E. Harrell, eds. Proc. Conf. Raptor Con-
serv. Tech. Raptor Res. Rep. 3- Raptor Res. Found.
Vermillion, SD.

and J.C OGDEN. 1979. The recent history and

status of the Mississippi kite. Am. Birds 33:119-130.

PETERSEN, L. 1 979. Ecology of great horned owls and
red-tailed hawks in southeastern Wisconsin. Tech.
Bull. 111. Wisconsin Dep. Nat. Resour. Madison.
63pp.

PHILLIPS, R.L., T.P. McENEANEY, and A.E. BESKE. 1984.
Population densities of breeding golden eagles in
Wyoming. Wildl. Soc. Bull. 12:269-273.

PLATT, J.B. 1973- Habitat and time utilization of a pair of
nesting sharp-shinned hawks (Accipiter striatus
velox) — a telemetry study. M.S. Thesis, Brigham
Young Univ. Provo UT. 42pp.

POOLE, A. 1981. The effects of human disturbance on
osprey reproductive success. Colonial Waterbirds
4:20-27.

PORTER, R.D. and CM. WHITE. 1973. The peregrine

falcon in Utah, emphasizing ecology and competition
with the prairie falcon. Brigham Young Univ. Sci. Bull.
Biol. Ser. 18:1-74.

POSTOVIT, H.R. 1979. Population estimates of breeding
raptors in North Dakota Badlands. M.S. Thesis, North
Dakota State Univ., Fargo. 55pp.

POSTUPALSKY, S. 1974. Raptor reproductive success:
some problems with methods, criteria, and terminol-
ogy. Pages 21-31 in Hamerstrom, F.N. Jr., B.E. Harrell,
and R.R. Olendorff, eds. Management of Raptors. Rap-
tor Res. Found. Vermillion, SD.

REYNOLDS, RT. 1982. North American accipiter hawks.
Pages 288-289 in Davis, D.E., ed. Handbook of Census
Methods for Terrestrial Vertebrates. CRC Press. Boca
Raton, FL.

. 1983. Management of western coniferous forest

habitat for nesting accipiter hawks. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. RM-102. 7pp.

and B.D. LINKHART. 1984. Methods and materials

for capturing and monitoring flammulated owls. Great
Basin Nat. 44:49-51.

, EC. MESLOW, and H.M. WIGHT. 1982. Nesting

habitat of coexisting accipiters in Oregon. J. Wildl.
Manage. 46:124-138.

ROSENEAU, D.G. 1972. Summer distribution, numbers,
and food habits of the gyrfalcon (Falco rusticolus L. )
on the Seward Peninsula, Alaska. M.S. Thesis, Univ.
Alaska, Fairbanks. 124pp.

ROSENFIELD, R.M., J. BIELEFELDT, R.K ANDERSON, and
W.A. SMITH. 1985. Taped calls as an aid in locating
Cooper's Hawk nests. Wildl. Soc. Bull. 1 3:62-63.

SCHMUTZ, J.K, S.M. SCHMUTZ, and DA. BOAG. 1980.

Coexistence of three species of hawks (Buteo sp.) in

the prairie parkland ecotone. Can. J. Zool. 58:1075-

1089.
SCHNELL, G.D. 1967. Population fluctuations, spatial dis-
tribution, and food habits of rough-legged hawks in

Illinois. Kansas Ornithol. Soc. Bull. 18:21-28.
SCHNELL, J. 1979. Black hawk (Buteogallus anthraci-

nus). Tech. Note TN-329- U.S. Dep. Inter., Bur. Land

Manage. Serv. Cen. Denver, CO. 25pp.
SCOTT, V.E., KE. EVANS, DR. PATTON, and CP. STONE.

1977. Cavity nesting birds of North American forests.

Agric. Handb. 511. U.S. Dep. Agric, For. Serv. Wash-
ington, DC. 122pp.
SIMINSKI, DP. 1976. A study of great horned owl (Bubo

virginianus) population density with recorded calls

in northwestern Ohio. M.S. Thesis, Bowling Green

State Univ., Bowling Green, OH. 35pp.
SIMPSON, MB. Jr. 1972. Saw-whet owl population of

North Carolina's southern Great Balsam Mountains.

Chat. 36:47.
SMITH, CF. 1978. Distributional ecology of barred and

great horned owls in relation to human disturbance.

M.S. Thesis, Univ. Connecticut, Storrs. 104pp.
SNOW, C. 1973. Golden eagle. Tech. Note 239. U.S. Dep.

Inter., Bur. Land Manage. Serv. Cen. Denver, CO.

52pp.
SOKAL, R.R. and F.J. ROHLF. 1981. Biometry. The princi-
ples and practice of statistics in biological research,

2nd ed. W.H. Freeman, San Francisco, CA. 859pp.
SPRINGER, M.A. 1978. Foot surveys versus owl calling

surveys: a comparative study of two great horned owl

censusing techniques. Inland Bird-Banding News

50:83-92.
STALMASTER, M.V. 1983 An energetics simulation model

for managing wintering bald eagles. J. Wildl. Manage.

47:349-359.
STEENHOF, K 1976. The ecology of wintering bald eagles

in southeastern South Dakota. M.S. Thesis, Univ.

Missouri, Columbia. 147pp.
. 1 978. Management of wintering bald eagles. FWS/

OBS/78-79- U.S. Dep. Inter., Fish and Wildl. Serv.

Washington, DC. 59pp.
. 1986. Assessing raptor reproductive success and

productivity in Millsap, B.A. and KW. Kline, eds.

Raptor Management Techniques Manual. Vol. 1. Natl.

Wildl. Fed. Washington, DC.
, S.S. BERLINGER, and L.H. FREDRICKSON. 1980.

Habitat use by wintering bald eagles in South Dakota.

J. Wildl. Manage. 44:798-805.
and M.N. KOCHERT. 1982. An evaluation of meth-

ods used to estimate raptor nesting success. J. Wildl.
Manage. 46:885-893-

STUDER, CD. 1983. Where is condor habitat? Outdoor
California 44:17-18.

SYKES, P.W. Jr. 1979- Status of the Everglade kite in Flor-
ida 1968-1978. Wilson Bull. 91:495-511.

TAYLOR, A.L. Jr. and ED. FORSMAN. 1976. Recent range
extension of the barred owl in western North Amer-
ica, including the first records for Oregon. Condor
78:560-561.

THIOLLAY, J.M. 1978. Population structure and seasonal
fluctuations of the falconiforms in Uganda national
parks. East Afr. Wildl. J. 16:145-151.

348

Raptors

THOMPSON-HANSON, PA. 1984. Nesting ecology of
northern harriers on the Hanford site, south-central
Washington. M.S. Thesis, Washington State Univ.,
Pullman. 99pp.

TITUS, K. and J.A. MOSHER. 1981. Nest site habitat se-
lected by woodland hawks in the central Appalachi-
ans. Auk 98:270-281.

TRIMBLE, S.A. 1975. Habitat management series for

unique or endangered species, merlin (Falco colum-
barius). Tech. Note TN-271. U.S. Dep. Inter., Bur.
Land Manage. Serv. Cen. Denver, CO. 4 lpp.

U.S BUREAU OF LAND MANAGEMENT. 1979. Special
research report to the Secretary of the Interior. Bur.
Land Manage., Boise, ID. 14 lpp.

U.S. FISH AND WILDLIFE SERVICE. 1983- Northern
States bald eagle recovery plan. U.S. Dep. Inter., Fish
and Wildl. Serv. Denver, CO. 1 1 7pp.

VILLAGE, A. 1984. Problems in estimating kestrel breed-
ing density. Bird Study 31:121-125.

WAIAN, LB. and R.C. STENDALL. 1970. The white-tailed
kite in California with observations of the Santa Bar-
bara population. Calif. Fish Game. 56:188-189-

WAKELEY, J.S. 1978. Factors affecting the use of hunting
sites by ferruginous hawks. Condor 80:316-326.

WELLER, M.W., I.C. ADAMS, Jr., and B.J. ROSE. 1955. Win-
ter roosts of marsh hawks and short-eared owls in
central Missouri. Wilson Bull. 67:189-193-

WHITE, CM. and S. SHERROD. 1973. Advantages and
disadvantages of the use of rotor-winged aircraft in
raptor surveys. Raptor Res. 7:97- 1 04.

WIER, D.N. 1982. Cliff nesting raptors of the Kisaralik
River, western Alaska. Pages 138-152 in Ladd, W.N.
and P.F. Schempf, eds. Raptor Management and Biol-

ogy in Alaska and Western Canada. WS/AK Proc. 82.
U.S. Dep. Inter., Fish and Wildl. Serv. Anchorage, AK.

WflNANDTS, H. 1984. Ecological energetics of the long-
eared owl (Asio otus). Ardea 72:1-92.

WILBUR, S.R. 1978. The California Condor, 1966-76: a
look at its past and future. North Am. Fauna 72. U.S.
Dep. Inter., Fish and Wildl. Serv. Washington, DC.
136pp.

WILEY, J. W. 1975. The nesting and reproductive success
of red-tailed hawks and red-shouldered hawks in
Orange County, California, 1973- Condor 77:133-139.

WINTER, J. 1979- Status and distribution of the great gray
owl and the flammulated owl in California. Pages
60-85, in Schaffer, P.P. and S.M. Ehlers, eds. Owls of
the West: Their Ecology and Conservation. Nat. Audu-
bon Soc. Tiburon, CA.

WRAKESTRAW, G.F. 1972. 1972 Wyoming bald and

golden eagle survey. Job Completion Rep. W-50-4-21.
Wyoming Game Fish Comm., Cheyenne. 7pp.

. 1973- The 1973 Wyoming bald and golden eagle

survey. Am. Birds 27:716-718.

ZARN, M. 1974a. Spotted owl (Strix occidentalis). Tech.
Note TN-242. U.S. Dep. Inter., Bur. Land Manage. Serv.
Cen. Denver, CO. 22pp.

. 1974b. Burrowing owl {Speotyto cunicularia

hypugaea). Tech. Note TN-250. U.S. Dep. Inter., Bur.
Land Manage. Serv. Cen. Denver, CO. 25pp.

. 1974c. Osprey {Pandion haliaetus carolinensis).

Tech. Note TN-254. U.S. Dep. Inter., Bur. Land Man-
age. Serv. Cen. Denver, CO. 4 lpp.

. 1975. Rough-legged hawk (Buteo lagopus sancti-

johannis). Tech. Note TN-270. U.S. Dep. Inter., Bur.
Land Manage. Serv. Cen. Denver, CO. 23pp

Raptors

349

17

MARSH AND
SHOREBIRDS

Peter G. Connors

Bodega Research Associates

P.O. Box 247

Bodega Bay, CA 94923

Editor's Note: Marsh and shorebirds, because of the
habitat they occupy and the unique adaptations for
feeding and breeding, require inventory and moni-
toring techniques quite different from those used
for songbirds, for example. Although marsh and
shorebirds are not a homogeneous group in terms
of habitat, size, taxonomic affinity, or life history,
they nevertheless require methodology that differs
significantly from any other species group, includ-
ing waterfowl with which they often coexist. This
chapter describes these specialized techniques and
their application.

INTRODUCTION

Few general statements concerning habitat rela-
tionships will apply accurately to all the groups of
birds addressed in this chapter. Marsh and shorebirds
include two families in the order Gruiformes and
five families in the order Charadriiformes (Table 1).
Typical nesting habitats range from arctic tundra
to Colorado River marshes and from western range-
lands to the rocky California coastline. As if that
habitat diversity were not enough, species range
from permanent residents to some of the longest dis-
tance migrants in the world. Shorebirds are more
consistently and emphatically migratory than any
other taxonomic group of North American birds, and
the habitats occupied in different seasons and differ-
ent areas often contrast sharply. Finally, species
range in size from least sandpipers (Calidris minu-
tilla) to sandhill cranes (Grus canadensis) and in
behavior and accessibility, from visually prominent,
frequently flocking species seen in open landscapes
to some of the world's most secretive denizens of
densely vegetated marshes, the rails.

Table 1. Families of birds addressed in this chapter
occurring regularly in Alaska and western States.

Order
Family

Number of
Species

Gruiformes — marsh birds
Rallidae — rails and coots
Gruidae — cranes

7
2

Charadriiformes — shorebirds
Charadriidae — plovers
Haematopodidae —

oystercatchers
Recurvirostridae — stilts and

avocets
Scolopacidae — sandpipers,

phalaropes, and close

relatives

7
1

2

35

Marsh and Shorebirds

351

This great diversity of habitats, life cycles, and
behaviors requires a divided approach to this discus-
sion of habitat relationships and population estima-
tion techniques. In the interest of brevity, however,
not all species can be individually addressed. Instead
I consider four major categories of habitat, attempt-
ing to discuss the most general features of species
group habitat use within each category. This ap-
proach will unavoidably omit some species that do
not share characteristics with other species dis-
cussed below; some of these will be mentioned only
briefly, with appropriate references if available.
Others can best be investigated by methods dis-
cussed for other species groups with similar habitat
use, behavior, and size. For example, American coots
(Fulica americana) in breeding marshes can be
studied with grebes (Podiceps sp. ) and, in winter,
with small waterfowl; sandhill cranes can be partly
addressed with large arctic breeding waterfowl.

The four habitat categories addressed in this
chapter are tundra, rangeland, shorelines (coastal
and inland), and marshes. Table 2 outlines the spe-
cies groups treated under each habitat heading and
emphasizes the generalizations applied to efficiently
treat the diversity of this chapter's subject. These
combinations of habitats and species groups define
the sections and limit the scope of the discussion in
this chapter.

Table 2. Major habitats and associated species
groups.

Habitat

Species Group

Tundra —
Arctic and
subarctic

Breeding shorebirds and
cranes

Rangeland —
Upland
and
wetlands

Breeding shorebirds and
cranes

Shorelines —
Coastal
and
inland

Non-breeding shorebirds

Marshes —
Coastal
and
inland

Breeding and non-breeding
rails and coots

TUNDRA— BREEDING SHOREBIRDS

At most arctic and many subarctic sites, shore-
birds are a major, conspicuous element of the avi-
fauna. On the arctic coastal plain of Alaska near

Prudhoe Bay, Troy et al. ( 1983) found 13 of 24 nest-
ing species to be shorebirds, including 8 of the 10
species with nesting densities above 1 nest/knT
(0.4 mi.2).

Comparable results have been recorded at Bar-
row (Myers and Pitelka 1980) and Point Storkersen
(Bergman et al. 1977) as well as at other Alaska
sites. Total shorebird densities are usually higher on
tundra with numerous small wetlands than on well-
drained sites and, therefore, higher on most coastal
plain tundra than in inland elevated areas, although
densities of some species reverse this trend because
of habitat preferences. All tundra-breeding shorebirds
are migratory, arriving at nesting sites in May or
early June in Alaska and departing soon after nesting
duties are finished, from late June to September.
Timing of breeding seasons matches the hatching of
insects on which adults and newly hatched chicks
depend (Holmes and Pitelka 1968). In many areas,
shorebirds move from breeding tundra areas to
nearby coastal areas in late summer before beginning
southward migration (Connors et al. 1979).

Habitat Features

Many studies have related tundra habitat fea-
tures to shorebird populations in Alaska, but few
have dealt quantitatively with fine-scale variation in
habitat types or with basic features of habitat struc-
ture. The most detailed studies dealing principally
with shorebirds were conducted by Myers and Pi-
telka ( 1 980 ) at Barrow and Atkasook and by Troy et
al. (1983) at Prudhoe Bay. Many other studies de-
scribe habitat characteristics somewhat more gener-
ally, often relating shorebird densities to different
classes of habitat. These include Spindler (1978),
Martin and Moiteret ( 1981 ), Bergman et al. ( 1977),
Connors et al. (1979), Williamson et al. ( 1966),
and Holmes and Black ( 1973). Several other govern-
ment reports and a few graduate theses also address
this topic, and many papers deal with habitat prefer-
ences of particular species. The most quantitatively
derived and most generally applicable conclusions
from these sources will be discussed here.

Physical Features

Myers and Pitelka (1980) measured several
physical features of shorebird habitat use at Barrow
and Atkasook, Alaska. They measured average and
maximum microtopographic relief on transect units,
average slope of unit diagonals, average water depth,
and proportion of sampling points with different
degrees of saturation or surface water present. Com-
bining these variables with vegetation measurements,
they used multivariate factor analysis to identify ma-
jor gradients underlying the habitat variation at these
sites and found corresponding differences in shore-
bird densities across these gradients, as well as spe-
cies differences in habitat use. Other researchers

352

Marsh and Shorebirds

have used classification schemes based principally on
differences in topography, wetness, and vegetation
to identify differences among species and among
habitats (Bergman et al. 1977; Spindler 1978; Martin
and Moiteret 1981; Troy et al. 1983). Because both
these approaches relate bird use to integrated de-
scriptions of habitat, it is difficult to identify individ-
ual features of prime importance, but topographic
relief and the occurrence of surface water certainly
influence shorebird populations.

that provide a variety of nesting and foraging habi-
tats. However, this habitat mix is not well-repre-
sented at the arctic sites where detailed shorebird-
habitat relationships have been studied. Directly
censusing shorebird populations is, therefore, often
preferable to habitat measurements, except when
habitat measurements are obtained primarily by aer-
ial photography and are compared with bird-habitat
studies done in similar terrain within the same
region.

Individual species differ in habitat preferences,
but total shorebird densities are usually highest dur-
ing the breeding season on tundra of variable topog-
raphy (such as the polygons of ridges and troughs
produced by frost action), with numerous shallow
wet troughs or small ponds. This combination pro-
vides habitat for nesting and foraging for several
of the most common tundra shorebirds. Well-drained
sites without ponds, including highly polygonized
areas, support lower densities, although some species
such as golden plovers (Pluvialis dominica) and
buff-breasted sandpipers (Tryngites subruficollis)
may select these habitats. Areas of continuous marsh,
lacking elevated dry sites, are also less used during
the breeding season because of the scarcity of nest
sites, but species such as pectoral sandpiper (Cali-
dris melanotos) use these areas after chicks hatch.

These generations do not apply strictly to all
sites and to all seasons. However, on the Canning
River Delta, Martin and Moiteret ( 1981 ) measured
higher breeding densities on a mosaic of wet and dry
polygonized tundra than on sites of more homoge-
neous wet or dry habitat. Spindler ( 1978) found
similar results at the Okpilak Delta, although densi-
ties in low wetlands were also high. At Barrow, a
region of little topographic relief but extensive wet-
lands, Myers and Pitelka (1980) measured highest
densities of breeding season shorebirds in areas of
low to intermediate polygonization, with many small
ponds, and in areas of high polygonization, with
fewer ponds. Inland on the coastal plain at Atkasook,
where well-drained tundra is more widespread, they
found most shorebirds nesting in low, non-polygon-
ized areas having a wide range in density of ponds.
These patterns changed seasonally, however, as shor-
ebird use in all three studies became heavier in low
wetland sites after shorebirds finished nesting duties.

Thus assessment of habitat value for shorebirds
is not a simple matter of measuring features such
as density of ponds or height of polygons, for exam-
ple. Furthermore, regional differences in species
present or in gross geomorphology between sites
may override the influences of particular habitat
feature studies at one site. For example, the highest
shorebird nesting densities on the Yukon Delta oc-
cur in areas of heath tundra interspersed with marsh,
delta channels, and ponds (Holmes and Black 1973)

Vegetation

On most arctic tundra, vegetation structure is
relatively monotonous compared with temperate
ecosystems. Tundra ranges from almost unvegetated
soils to dense stands of grasses, sedges, and forbs,
with areas of low woody shrubs such as birches and
willows, often lower than the surrounding grasses.
Plant species composition and, therefore, vegetation
structure are to a great extent determined by the
physical features of the tundra within a region. Most
habitat classification schemes have defined tundra
types by combining these factors. For some species,
vegetation height or density may strongly influence
bird use, and gross differences in plant species com-
position (as between shrubby heath tundra and
graminoid tundra, for example) may separate some
species. In general, however, vegetation is not a sim-
ple predictor of total shorebird populations except
as it corresponds with the differences in physical
features discussed above.

Population Measurement Techniques

Direct measurement of tundra shorebird densi-
ties may often be preferred over indirect population
evaluation based on habitat measurements, primarily
because tundra shorebirds are easier to census than
most other non-colonial nesting birds. In areas of
tundra, such as the North Slope and arctic coastal
areas of Alaska, low vegetation reduces visibility
problems that plague most censuses of birds in habi-
tats with complex vegetation structure, such as
woodlands. Censusing shorebirds does not require
singing birds or estimates of distances to unseen
singing birds.

On arctic coastal plain tundra in Alaska, most
shorebird individuals, except some incubating sand-
pipers and secretive species such as long-billed dow-
itcher (Limnodromus scolopaceus) and common
snipe (Gallinago gallinago), can be located by an
experienced observer on a walking transect. The
relative tameness of most shorebirds allows a closer
approach by the census taker than is possible for
more elusive species such as waterfowl and cranes.
Reliable estimates of absolute density are often possi-
ble with an acceptable amount of effort. Measure-
ment of simple presence or absence of shorebirds is

Marsh and Shorebirds

353

therefore seldom justified, because relative density
estimates, at least, can usually be obtained with little
extra effort.

Sandhill cranes, breeding in some of the same
tundra habitats, require a very different approach
because of their lower densities and greater flushing
distances. Their large size, however, permits aerial
censusing for relative or absolute density measure-
ments (Riley 1982). This species is not addressed in
the discussions below of techniques designed to
assess densities of groups of shorebirds species.

Relative Density — Encounter Rate Method.

For a simple estimate of relative density of shore-
birds, by species, an observer walks through a site
recording all birds seen per distance or time trav-
eled. The census path can be straight or meandering,
provided that it does not incorporate any strong
bias toward habitat of a particular type that is not
equally abundant at all sites compared. A continuous,
roughly circular route from a point of departure
such as an airplane landing site may be most efficient
when comparing several remote sites. Distances are
best estimated from topographic maps; alternatively,
time elapsed at a reasonably constant hiking rate
can be used.

This method has relatively low accuracy and
precision, but it quickly and cheaply obtains a list of
most or all common species at a site and a rough
estimate of their relative abundances. Reliability of
this estimate depends on distance traveled (I suggest
4 to 12 km [2 to 7 mi.]); consistency; and experi-
ence of the observer, weather conditions, hiking
conditions, and season. For example, windy weather
reduces the number of birds heard and then ob-
served; hiking rates through tussocky or very marshy
tundra will be slower than in easier terrain; and
birds differ in visibility seasonally, depending on
their behaviors, such as displaying, incubating, or
mobbing. Furthermore, because species differ in
detectability within a season, relative densities of
certain species detected by this method incorporate
a somewhat predictable bias.

For best results, the census should be along
straight-line paths chosen from topographic maps,
during the same period of the breeding season, at all
sites or in all years. Early breeding season is best
for estimating relative breeding densities, about mid-
June at arctic locations.

This method requires only binoculars, notebook,
and hip boots, plus an experienced observer familiar
with displays, vocalizations, behavior, and identifica-
tion of the species that may be encountered. Cost
is low except for transportation to the sites. The
method should only be used when time and funds
are very limited; when many sites will be visited

briefly, usually only one time each; or when bird ob-
servations must be supplemental to other observer
duties. In these circumstances, it becomes a very
efficient method of obtaining relative density data of
limited accuracy. When more time is available or
greater accuracy is needed, the following methods
should be used.

Absolute Density.

Method 1 — Territory Spot Mapping. On a

marked study area, an observer records the location
of all birds, noting especially locations of territorial
displays and nests. Study plots and data sheets should
be correspondingly marked with grid lines indicating
50 m (165 ft) intervals. Censuses should take place
on several dates during the early and middle
breeding season. I suggest three to eight censuses
per study plot. Combining censuses for each species
using overlay maps, the observer plots territories
or nest sites. Densities can be calculated only from
fractional portions of territories that cross plot
boundaries within the study area.

This method is reasonably accurate and precise.
Successive censuses add to the information base,
increasing accuracy each time. Data from all cen-
suses are combined to estimate the actual resident
population. This process requires judgment in the
analysis and should be done by the field observer, re-
vising the map as required after each census. The
observer must know displays, nesting behavior, and
identification of all the local species, and must be
familiar with nest location techniques for shorebirds.
Study plots should be large enough to include sev-
eral territories of common species (typically 20 to
100 ha [50 to 250 a.]).

Little equipment is needed, as in the preceding
method. One observer can map 30 to 60 ha (75
to 150 a.) of study plot per day. Many of the inter-
fering factors discussed above apply also to this
method. However, the marked boundaries, measured
area, and repeated censuses greatly increase the ac-
curacy of the density estimate. In addition, densities
correspond to known or strongly suspected breeding
birds only, providing what is often the most useful
population measure. Multiple censuses, however,
require much greater personnel time and, therefore,
increase costs.

Method 2 — Rope-Drag Nest Mapping. The

rope-dragging technique can be used independently
or as a supplement to the previous method. Two
observers drag a rope (about 50 m [165 ft] long)
across the tundra, while a third observer follows
behind, noting locations of any birds that flush from
nests as the rope passes over or near them. This
simple method locates incubating birds that "sit

354

Marsh and Shorebirds

Rope-drag technique using vehicles.

close," not flushing, unless the observer almost steps
on them. These are the most difficult of shorebird
nests to locate by usual mapping techniques and are
frequent among some species such as the long-billed
dowitcher.

Method 3 — Marked Plot Strip Transect. To

measure changes in total bird populations (not just
nesting pairs) within a season, between seasons,
or between sites, many researchers have used perma-
nently marked transects treated as study plots (Con-
nors et al. 1979; Myers and Pitelka 1980; Troy et
al. 1983)- Numbered wooden stakes or similar mark-
ers are placed in a straight line at 50 m ( 165 ft)
intervals, indicating the center line of two adjacent
rows of 50 x 50 m (165 x 165 ft) plots. A single
observer walks a rectangular zigzag path across all
transect plots, walking within 25 or 18 m (82 or 59
ft) of all points on the plots. Alternatively, two ob-
servers walk the center lines of the parallel rows of
plots.

The plot markers help observers estimate lateral
distances, and additional stakes can be placed on
outer transect edges to aid the estimation. Time
required averages about 1 observer-hour/kilometer
(1/2 mile) of transect on arctic plain tundra. In areas
of taller shrubby vegetation, visual obstructions may
require increasing the amount of zigzagging and
the time required per transect. All birds are re-
corded, and the observer must note locations and
movements to avoid counting individuals more than
once. Censuses can be repeated at regular intervals

and averaged to obtain mean densities and changes
in bird use with time. Transects can be placed
within single habitats or randomly crossing all habi-
tats, depending on sampling design.

These plot transects work well because of the
relatively high visibility of tundra shorebirds, but
they are subject to errors caused by differential visi-
bility of species. Experienced observers are confident
that almost all shorebirds are recorded by this
method Q.P. Myers, F.A. Pitelka, D. Troy, pers. com-
mun. ). The most difficult application of the tech-
nique occurs after chicks hatch and some shorebird
parents engage in mobbing activities, flying across
several transect plots to display, near the observer, a
potential predator. The observer must note the
origins of these birds to avoid overcounting during
this period.

Precision and, to some extent, accuracy of this
technique can be assessed by replicate censusing
of control plots or of all plots. This might be done
by comparing different observers or a single ob-
server over time intervals of one hour, several hours,
or single days. This assessment greatly aids the inter-
pretation of data collected from single counts at
comparable sites. A similar approach of replicate
censuses is useful for all the techniques mentioned
in this chapter.

Method 4 — Unmarked Strip Transect. When
several sites must be evaluated, usually with single
visits, an unmarked strip transect of fixed width ( I

Marsh and Shorebirds

355

recommend 50 m [ 165 ft] total width) can be used.
This is an upgraded version of the Encounter Rate
Method for relative density. By controlling the width
of the census path, absolute densities can be calcu-
lated. Density accuracy depends on factors men-
tioned earlier, plus any error in estimating the 25 m
(82 ft) width on either side of the observer's path.
Accuracy also suffers compared with Marked Plot
Transects by the absence of replicate estimates, but
Method 4 has an obvious advantage by saving time
and money.

Method 5— Variable Width Line Transect.

Several variations of this method have been devel-
oped, principally for passerine censusing in wooded
habitats. Corrections have been applied to account
for differential detectability of species (Emlen 1971,
1977; Jarvinen and Vaisanen 1975; Ralph and Scott
1981). Briefly, all birds are recorded on either side
of a central transect line, along with the perpendicu-
lar distance from each bird to the transect line. Tun-
dra shorebirds do not pose the same visibility
problems as woodland songbirds, so the additional
complication of estimating horizontal distances to all
birds is probably not warranted.

Discussion

Tundra shorebird habitat relationships have
been investigated at several arctic sites. The closest
scrutiny focused on individual species preferences
and seasonal changes in habitat use. Nevertheless, no
clear definition of habitat features that can predict
shorebird abundances has yet emerged. The most
widely applicable generalization from these studies
suggests that total shorebird densities will usually be
highest in areas of numerous small wetlands with
moderate microtopographic relief, providing a fine-
grained mosaic of nest sites and foraging sites.

The low tundra vegetation, moderate conspicu-
ousness, and non-elusive character of most shorebird
species at most tundra sites make shorebird census-
ing on breeding grounds relatively easy. For this
reason, measures of simple presence or absence are
not recommended, because they are easily replaced
with estimates of relative or absolute density. All
these techniques require at least the ability to easily
identify all local species in a variety of plumages, and
all are more reliable when observers know the be-
havioral peculiarities and vocalizations of shorebird
species. Methods should be chosen on the basis of
available time and personnel, number of sites to be
monitored, duration of monitoring planned, and the
information needs of the monitoring program.

RANGELAND— BREEDING SHOREBIRDS

In contrast to their prominence on arctic tun-
dra, shorebirds are much less common and conspicu-
ous on western rangelands, but several species nest

there, and additional species are found throughout
the West during migration. A brief summary of habi-
tat affinities and census techniques follows.

Habitat Features

Western rangeland encompasses a greater vari-
ety of contrasting habitat types than can readily be
treated in this short section, from prairies and plains
to mountain slopes and valleys; from deserts to rich
grasslands, including marshes, streams, fresh and
saline lakes, over a wide range of elevations and win-
ter climates. Most nesting shorebird species can be
assigned to groups associated with prairie grasslands,
marshes, or stream and lake shores, although these
categories overlap for some species.

Grassland-nesting shorebird species include
mountain plover (Charadrius montanus), a long-
billed curlew (Numenius americanus), upland sand-
piper (Bartramia longicauda), and sometimes kill-
deer (Charadrius vociferus). Physical characteristics
of different grasslands can be described by measur-
ing such factors as mean vegetation height; percent-
age vegetation cover types (grass, forb, shrub,
cactus); slope and topographic diversity; and dis-
tance to water.

Intensive studies of long-billed curlews in Idaho
(Jenni et al. 1982; Redmond et al. 1981; Redmond
1984) have demonstrated that curlew densities in
that area are highest in areas of low topographic
slope, low vegetation height, and low vertical vegeta-
tive cover. Mountain plovers in Colorado also se-
lected low slope, shortgrass habitats (Graul 1975).
Upland plovers occur in tallgrass or mixed grass
prairie habitats (Wiens 1973). Since these species
are conspicuous enough in these habitats, direct
censusing of birds during the breeding season is
more effective than attempts to assess habitat
suitability.

Marsh and pond nesting shorebirds occurring
within western rangelands are American avocet (Re-
curvirostra americana), black-necked stilt (Himan-
topus mexicanus), willet (Catoptrophorus
semipalmatus), marbled godwit (Limosa fedoa),
common snipe, and Wilson's phalarope (Phalaropus
tricolor). Distributions and habitats of these species
vary widely and have not all been the subjects of
intensive habitat studies. Furthermore, habitat use by
a single species may change regionally as wetland
types change over large distances of western North
America. Wetlands in many of these areas are sea-
sonal or semipermanent and have been classified
according to permanence by Stewart and Kantrud
( 1971 ) for the northern prairie region.

Black-necked stilts and American avocets nest in
similar habitats, usually along shorelines and grassy
or unvegetated flats near fresh water, brackish or

356

Marsh and Shorebirds

Black-necked stilt turning eggs.

alkaline marshes, and ponds. Foraging habitats, usu-
ally shallow water areas or wet mud flats, may be
continuous with or separated from nest territories
(Gibson 1971; Hamilton 1975). In prime areas, nest
densities may be very high, clustered in semi-colon-
ial situations. This distribution of nests and the usu-
ally low vegetation facilitate direct spot-mapping
techniques for nesting birds.

Willets and marbled godwits nest near prairie
marshes and ponds surrounded by grasslands. Both
species prefer permanent or semipermanent pond
types, either fresh or brackish, and nest in adjacent
grassy areas (Ryan 1985). Godwits use surrounding
uplands in addition to marshes for foraging and pre-
fer short to medium grass heights for nesting. Willets
use taller grass areas (Ryan, unpubl. ms). Wilson's
phalaropes nest in ponds and marshy wetlands of
prairie and basin areas, both fresh and saline, and
common snipes nest in marshes of the northern
States.

Shoreline nesting species include killdeer,
snowy plover (Charadrius alexandriniis), spotted
sandpiper (Actitis macularia), and in some northern
prairie regions, piping plover (Charadrius melodus).
These are usually associated with lakes, ponds, or
streams (except that killdeer may nest well away
from water) and nest in open areas or areas of sparse
or short vegetation (Bent 1929; Renaud 1979; Page
et al. 1983; Oring et al. 1983; Weseloh and Weseloh
1983).

Sandhill cranes also nest in western rangeland
areas, in or near marshes or ponds (Bent 1926; Lit-
tlefield and Ryder 1968; Drewien 1973).

As with shorebirds breeding on tundra, many
rangeland species are relatively conspicuous in the
sparsely vegetated or shortgrass habitats they occupy
in western rangelands. This circumstance facilitates
direct measurement of shorebird or crane densities.
For species associated with wetlands, however, locat-
ing and mapping areas of wetlands by wetland type
may provide a large-scale estimate of relative abun-
dance of these species. This approach requires cali-
bration from bird studies done within the same
region. Information relating presence, absence, or
relative or absolute densities of species to wetland
type or wetland size class can then be applied to
areawide inventories of wetlands. Weber et al.
(1982) provided an example of the kind of calibra-
tion information required on northern prairies.

Population Measurement Techniques

Census methods must be chosen to match the
species or habitat characteristics within this diverse
group of species. Many of the methods listed for
tundra shorebirds are also applicable to rangeland
species and are listed here briefly with the habitats
or species of concern. A few additional approaches
are described more fully.

Relative Density — Encounter Rate Method.

The encounter rate method is generally less useful in
rangeland shorebird habitats than in tundra, where
the nesting habitats of large numbers of species oc-
cur in a fine-grained mosaic. Nevertheless, walking
around wetlands or along shorelines during the
breeding season, or traversing grasslands on foot or
horseback, can provide easy but limited information
on relative density of nesting birds. Large areas can
be covered by aircraft searching for cranes, but the
effort involved argues for attempting absolute den-
sity estimates.

Absolute Density.

Method 1 — Nest Spot Mapping. Most studies
of shorebird populations breeding in the western
U.S. have been concentrated on locating all nests

Marsh and Shorebirds

357

within a study area. Techniques depend, to some
extent, on the species involved and habitat features,
but in all instances, actual ground searching is
required. This is usually initiated by observing
locations of birds flushing from or returning to
brooding activities, from a vantage point inside or
outside the studied habitat. For large birds (sandhill
crane, long-billed curlew, marbled godwit, avocet),
initial observations may be made from a road, dike,
or even an airplane. For smaller birds, initial
observations may require field work within the
habitat. For many species, thorough hiking may be
required to flush brooding birds or to elicit
antipredator responses. For species nesting in high
densities within limited habitats, such as the loosely
colonial breeders, American avocet and black-necked
stilt, spot-mapping techniques to locate all nests or
territories can be fairly efficient. In other situations,
this would be time-consuming.

Method 2 — Rope-Drag Nest Mapping. This
method is usually used only as a supplement to
Method 1 and is most useful for species that nest in
dense vegetation and those that may not flush with-
out a close approach, such as the marbled godwit
(Ryan 1985). The technique is described in the tun-
dra section.

Method 3 — Strip Transect. This method,
which requires locating birds within a linear transect
of a fixed width, can be applied as a walking transect
through grasslands, around wetlands, or near shore-
lines in a manner similar to the tundra transect de-
scribed in the preceding section. Alternatively, to
deal with the low densities of some rangeland spe-
cies and the patchy distribution of their habitats, the
method can be ( 1 ) a driving transect along a road,
counting all individuals of large species within a
standard perpendicular distance; (2) driving along a
road and stopping at predetermined stations for cen-
susing; or (3) a road transect linking wetland census-
ing stations. An aircraft flying transect lines with
observers censusing a specified width along the tran-
sect line is a useful variant for cranes.

Method 4 — Variable Width Line Transect.

Variations of this approach, described for tundra, can
also be applied to solve the particular problems of
rangeland species and habitats. Unlike most tundra
and shoreline shorebirds, long-billed curlews present
special difficulties, in both size and breeding habitat,
which affect census techniques. Redmond et al.
( 1981 ) compared three methods of censusing cur-
lew breeding populations in Idaho and recom-
mended a modified Finnish Line Transect method.
Observers drive a route along existing roads, record-
ing all curlews sighted within 500 m ( 1,650 ft) of
the road, and estimate perpendicular distances to all
birds sighted. The census is repeated five times on
different days, and pooled data are corrected for

Avocet turning eggs.

distance effects of detectability (Emlen 1971; Jarvi-
nen and Vaisanen 1975) before calculating densities.
Censuses should be done soon after curlews arrive
on breeding grounds (late March or early April).
Results of this and the strip transect method were
fairly accurate when compared with results of terri-
tory mapping studies done on the same area (Red-
mond et al. 1981). As with tundra shorebird
censusing, the change in behavior during the brood-
rearing period, with adults flying long distances to
mob potential predators, introduces errors. Census-
ing early in the breeding season avoids this problem.

Discussion

Western rangeland, including such diverse habi-
tats as grasslands, shrub deserts, mountain slopes,
marshes, and alkali lakes, to name a few, poses
greater problems for generalizing conclusions than
other major habitat areas addressed in this chapter.
Most rangeland nesting shorebirds, however, fall into
three habitat categories, as described above, and
census techniques applicable to these categories and
species are similar to those for comparable situations
elsewhere. The major modifications necessary arise
from the large size and low density of grassland
shorebirds or cranes, compared with most tundra
shorebirds, and the patchy distribution of western
ponds, marshes, and lakeshores. On a large-scale
area, the patchiness of gross habitat types, such as
marshy wetlands, permits successful assessment of
probable shorebird locations, but assessments of fine
habitat distinctions will usually be less useful than
direct censuses of birds.

358

Marsh and Shorebirds

SHORELINES— MIGRANT SHOREBIRDS

During much of the year, from late summer to
late spring, shorebirds are most common in or near
shoreline habitats, predominantly coastal but includ-
ing inland shores of lakes and marshes. This descrip-
tion encompasses a great variety of habitats ranging
from muddy freshwater ponds and flooded fields,
to sandy coastal beaches; from tidal mudflats and
saltmarshes, to exposed rocky coastlines; and from
arctic Alaska, to southern California and beyond.
Naturally, shorebird use, habitat features, and census
methods vary over this range, but some habitat char-
acteristics are common and these guide the choice
of census techniques.

First, most shoreline habitats of importance to
shorebirds are open and free of much vegetation
structure or topographic relief. Shorebirds in these
circumstances are even more readily located than on
tundra or rangelands. Second, most of these habitats
occur where land meets water, in areas where acces-
sible invertebrate prey in high densities cause forag-
ing shorebirds to concentrate locally. This factor also
facilitates shorebird censusing. However, the tidal
nature of most coastal shoreline habitats and the
flocking habits of most migrant and wintering shore-
birds inject elements of instability into shorebird
density measurements.

Migrant marsh and shorebirds move frequently
to exploit food sources, escape predators, or con-
tinue migrations. This instability can usually be ac-
commodated in at least one of several ways. For
example, biologists might census when shorebirds
are most predictably concentrated and stable, as
at high tide roosts in coastal areas. They might al-
ways census under the same set of conditions that
relate to movements, such as tide height, time of
day, and calendar date. They might census regularly
and repeatedly throughout a period of migration
or simultaneously census an entire local population,
using multiple observers. The following discussion
addresses the most generally important habitat fea-
tures and most widely applicable census methods for
inland and coastal shoreline habitats throughout
Alaska and the western States.

Habitat Features

Most migrating and wintering shorebirds do not
require vegetative cover for nests or chicks or for
their own protection. This frees them to use habitats
where their food, mainly small invertebrates of many
orders, are most abundant. In coastal locations, these
habitats are usually bare or sparsely vegetated, the
exceptions being mainly algae-covered rocks and
mudflats, and saltmarshes with low or sparse
vegetation.

Greater sandhill cranes, a western rangeland nester.

Marsh and Shorebirds

359

The primary features of habitat that relate to
shorebird densities must usually be those features
that result in the highest densities of available shore-
bird prey in an area, because shorebirds in winter
are generally distributed in relation to availability of
prey (Goss-Custard 1970; Bryant 1979; Myers et al.
1979; Rands and Barkham 1981). In arctic Alaska,
these conditions prevail in two separate situations in
late summer after shorebird breeding has been com-
pleted on the tundra. Densities of some species (red
phalaropes [Phalaropus fulicaria], red-necked phal-
aropes [P. lobatus], and sanderlings [Calidris alba])
are highest along beaches or spits, bars, and barrier
islands. Numbers of several other species (especially
dunlin [Calidris alpina], semipalmated [C. pusilla],
western [C. mauri] and pectoral sandpipers, and
long-billed dowitchers) are highest on mudflats, at
edges of sloughs, and around shallow pools in short
vegetation saltmarshes (Connors et al. 1979; Con-
nors 1984). These habitats usually support higher
shorebird use than other arctic shorelines, but densi-
ties are highly variable from one site to another and
cannot be predicted with accuracy.

On the Bering Sea coast, similar habitats, espe-
cially river deltas, lagoon mudflats, and saltmarsh
margins, attract highest shorebird densities during
post-breeding movements, from July through Octo-
ber (Shields and Peyton 1979; Gill et al 1981; Gill
and Handel 1981). From a large-scale view, these
coastal habitats in arctic and subarctic Alaska are lo-
calized and discrete, and their occurrences correlate
well with post-breeding shorebird densities through-
out the region. Field recognition of these habitats is
straightforward and is frequently possible from maps
and aerial photography. Quantitative characterization
of these gross habitat features is easily accomplished
by referencing properties of water depth, slope, dis-
tance, or extent from the water's edge; substrate
sediment size; substrate penetrability; vegetation
density and height; plant species composition; salin-
ity or saltwater influence; and surrounding landform
type.

On a finer scale, little quantitative work has
been done to isolate those features within mudflats
or saltmarsh, for example, that correlate with bird
density distributions within the habitat. As in other
coastal mudflat environments, density of available
invertebrate prey is probably the best predictive
feature to relate to shorebird densities. This is not
always easily measured, and habitat correlates of
prey densities are unstudied in these areas. For most
management purposes, the simple presence of more
than trace amounts of wet mudflats, shallow muddy
pools, and sparsely vegetated saltmarsh with wet
mud and shallow pools is sufficient to predict ele-
vated levels of seasonal shorebird use.

A few physical and vegetative features can be
used to identify and separate arctic saltmarshes from

freshwater tundra. First, because lunar tidal fluctua-
tions are so slight (about 15 cm [6 in.] at Barrow,
Alaska, for example), saltmarsh areas are maintained
by occasional flooding during storms in late summer
and fall. Therefore, saltmarsh is always located in
areas with access routes for storm-flood saltwater
from oceans or lagoons. The importance of storm
flooding leaves another clue in the form of driftwood
lines on coastal land, inland from any saltmarshes.
Depending on the frequency of storm flooding repre-
sented by different driftwood lines, vegetation shore-
ward of the driftwood may be very salt-tolerant,
mildly salt-tolerant, or fresh tundra pioneers invading
after the previous salt-killing of tundra vegetation.

In the permanent arctic saltmarsh zone of most
interest, two salt-tolerant plants are characteristic
indicators: Puccinellia phryganodes, a prostrate
grass, and Carex subspathacea, a very short sedge.
Both have a pronounced reddish color in some life
stages and, together with the reddish iron oxides
in saltmarsh mud, create a final color attribute that is
helpful in identifying and delineating arctic salt-
marshes. Within these vegetated areas, however, the
shallow pools, wet mudflats, and muddy margins
show the heaviest shorebird use.

For phalaropes, sanderlings, and lesser numbers
of several other species feeding primarily on zoo-
plankton along arctic shorelines, the habitat feature
of most importance seems to be the landform on
which arctic beaches occur. Spits and islands gener-
ally have higher densities of these species than main-
land beaches (Johnson and Richardson 1981;
Connors 1984). Vegetation and substrate characteris-
tics are less important than landform in predicting
these bird densities.

Farther south in Alaska, coastal shorebird densi-
ties are usually highest in areas of extensive mudflats
and, for some species, on rocky intertidal beaches.
The Copper River Delta is an important habitat area
with such large numbers of migrating shorebirds,
especially during spring, that total population num-
bers can only be estimated rather than directly
counted (Isleib 1979; Mickelson et al. 1980; Senner
1 979 ). The ample prey base, principally of small
molluscs on the extensive mudflats and sandflats of
that area, is responsible for the massive shorebird
concentrations (Senner 1979). In Washington, Ore-
gon, and California, the largest numbers of shore-
birds occur on mudflats and sandflats of lagoons,
estuaries, and fresh- and saltwater marshes; on
coastal beaches; and on rocky intertidal shorelines,
with different groups of species in different habitats.

Several studies have addressed species habitat
preferences among coastal wetland habitat types (for
example, Recher 1966; Gerstenberg 1979; Page et
al. 1979). These do not, however, permit assessment
of wildlife values of different examples of a single

360

Marsh and Shorebirds

wetland type, except on the basis of the extent of
habitat. Indeed, the population densities of some
wetland habitat areas seem to depend on the loca-
tions of nearby wetlands of other types, rather than
on strictly intrinsic properties (Page et al. 1979;
Connors et al. 1 98 1 ). Thus, one cannot yet profitably
judge shorebird wetland habitats by measuring par-
ticular habitat variables, except in a few specialized
instances. Sanderling densities on beaches in central
California, for example, are sometimes correlated
with beach slope and sediment size, because these
are correlated with prey densities (Myers et al. 1979;
Connors et al. 1981). On a larger scale, however,
one can assess areas by the presence and extent of
wetland habitats that provide foraging areas of the
types selected by particular species.

The situation for inland areas is similar. Wet-
lands providing appropriate species foraging areas
will attract shorebirds, but measurements of physical
or vegetational features within those habitats do
not permit assessment of relative value. This is not
generally a handicap, however, because direct mea-
surement of the shorebird populations in different
areas gives the best assessment of their relative val-
ues, and this measurement is usually not prohibi-
tively difficult to obtain.

On a large scale, total area of different wetland
types sets limits on migrant marsh or shorebird pop-
ulations of most species at inland sites. For example,
most migrant sandpipers will occur only on mudflats
and flooded fields or the beaches and shorelines of
ponds and lakes; Wilson's phalaropes concentrate in
large saline lakes during fall migration. Sandhill
cranes roost on river bars and forage in croplands.
The most efficient way of assessing the value of habi-
tat to shorebirds of different habitats, however, re-
mains the direct censusing of bird populations found
there. For example, recent efforts to relate numbers
of roosting sandhill cranes on Nebraska river bars
to habitat features such as water depth and water ve-
locity have established significant relationships.
These do not permit model predictions of crane
numbers at other river locations away from the stud-
ied sites (Latka and Yahnke, unpubl. ms), however.

Except for the influence of the tidal period on
bird movements and habitat availability, most aspects
of the census techniques discussed below apply
equally well to coastal and inland habitats. In the
western U.S., however, most inland and coastal wet-
lands differ in annual seasonality of bird use, because
inland wetlands away from the coast or away from
low elevations in the Southwest are usually frozen
during winter. These areas are used by shorebirds
only during spring and fall migration, mainly April
through May and July through September. During
these periods, however, densities of migrant shore-
birds may be very high, even in temporary, seasonal
wetlands. Flooded fields in springtime, in particular.

may attract large numbers of northbound migrant
shorebirds, perhaps because of elevated densities of
invertebrate prey (Krapu 1974). The temporary na-
ture of this shorebird habitat, of course, requires that
any habitat evaluation program, like shorebird census
studies, focus on conditions during the brief migra-
tion period.

Population Measurement Techniques

Presence — Aerial Survey Method. Flying in a
helicopter or fixed-wing aircraft, one or two
observers record, on tape or on maps, the presence
of flocks of shorebirds in different wetland locations
over large surveyed areas. Flight elevation should
be 25 to 60 m (80 to 200 ft) to promote flushing of
roosting or foraging birds. Observers can record
flock size or identify species or species groups
whenever possible.

This method is useful only in areas where al-
most nothing is known of shorebird distributions for
the time period surveyed. Except in unusual circum-
stances, aerial surveys cannot provide accurate
estimates of population size because of the low relia-
bility with which birds are sighted or counted from
the air. Compounding this problem, only well-
marked species that fly under the aircraft (dowitch-
ers, willets, black-bellied plovers [Pluvialis squata-
rola\ or phalaropes sitting on calm water, etc.) can
be identified consistently. Nevertheless, if large inac-
cessible areas or large numbers of inaccessible, small
wetlands must be surveyed for presence of migrant
shorebird flocks, the high cost of aircraft may be
justified.

Relative Density. Most techniques of ground
censusing for migrant shorebirds can be targeted to
either relative or absolute density estimates,
depending on the physical arrangement of local
habitat areas and the schedule and range of
shorebird movements among different areas. In
general, relative density estimates are obtained when
shorebirds are patchily distributed over the available
habitat of interest, when counts are too limited to
sample enough of the patchiness, or when parts of
local populations are counted in special circum-
stances, such as at roosting sites. Actual censusing
methods are discussed only under Absolute Density
below, but the applicability of census results must be
judged by biologists on the basis of factors just
mentioned. Although actual numbers of birds in an
area are counted in either situation, local population
distributions and movements and local habitat
distributions may require interpreting census
numbers as relative rather than absolute densities.

Marsh and Shorebirds

361

Absolute Density. For most of the methods
described here, important choices must be made
beforehand, to set some conditions for the census. In
coastal areas, the most important decision concerns
tidal conditions. Birds can be counted at high or low
tides, at specified time intervals before or after high
or low tides, or at particular tide levels on either
rising or falling tides. Setting these conditions is
central to intertidal shorebird censusing because
shorebirds move between habitat areas on a tidal
schedule, and the area of available habitat changes
on a tidal schedule (Connors et al. 1981).

The choice of conditions should be coordinated
with the census method chosen. For foraging birds
on mudflats, a falling tide is generally best, because
birds usually forage most intensively on falling tides,
having been prohibited from access to feeding areas
during the previous high tide. Selecting a particular
interval of tide levels permits the observer to locate
census plots conveniently. If very low tide levels
expose extensive mudflats, birds may forage too far
from convenient observation points along shorelines;
a mid-tide level may be more practical (Page et al.
1979). For censusing birds on beaches or at roosting
sites, a very high tide level may be best, because
this forces birds out of foraging areas on nearby
mudflats (Connors et al. 1981). In nontidal areas,
time of day assumes greater importance, and foraging
birds should usually be censused in early morning
or late afternoon when most are active.

Frequency and duration of censusing also must
be determined, because migrant shorebird popula-
tions may change quickly. To determine peak migra-
tion densities or timing of migration, multiple counts
are needed, spaced every 1 to 5 days. Single counts
or longer intervals between counts will be useful,
but could miss peak movements of short duration.

Method 1 — Sample Plot Count. In the sample
plot technique, an observer counts all birds within
a specified area. The observer can select areas with
natural topographic boundaries, mark plots with
stakes, or count a circular plot with a specified ra-
dius around the observing station. This last scheme,
also called a fixed-distance point count, is most use-
ful when an area will be censused only once, be-
cause it permits sampling many points without
mapping or staking. However, it requires the ob-
server to walk through the open shorebird habitat,
potentially disturbing the birds, and its accuracy
depends on the observer's ability to estimate the
specified radial distance.

Choosing sample plots with permanently
marked boundaries avoids this last problem and, in
many instances, the observer can count the entire
plot from one observing station, frequently posi-
tioned on a plot boundary (which may be the shore-
line ). Because of the open character and lack of
vegetation or topographic diversity in many migrant

shorebird habitats, an observer with binoculars or
a telescope can count and identify birds at great
distances (to 200 m [660 ft]) under good viewing
conditions. This permits large sample plot areas to
be covered from a single station or from a series of
stations along a shoreline. This method is well-suited
to regular repetitive censusing. Densities are deter-
mined by field measurement or map estimation of
sample plot areas.

Accuracy of the sample plot method is generally
high, limited mainly by difficulties in counting birds
that may move appreciably during the brief count
period. Poor visibility, estimates of large flocks, diffi-
culties with identifications at long distances, and
parallax problems of estimating plot boundary loca-
tions may also affect accuracy. As with all methods
listed here, equipment is minimal, training is easy if
observers have mastered shorebird identification, and
cost depends mainly on the number and frequency
of sample plot counts.

Method 2 — Marked Plot Strip Transect. This
technique is essentially identical to Method 3 for
tundra-breeding shorebirds. It is especially well-
suited to marshy areas or wetlands with open pools
and mudflats, interspersed with marsh, because the
observer moves through the transect close enough
to all points to discover birds partially obscured
by vegetation. The division of the transect into small
plots permits association of birds with habitat infor-
mation for each plot, an advantage in some studies.
In open mudflat habitats, the excellent visibility al-
lows the observer to walk directly along the center
line of a 100 m (330 ft) wide transect, accelerating
the census. As with the preceding method, accuracy
is high except when birds move on or off the tran-
sect during the count. The time required to mark
the transect is a good investment only when a series
of censuses are planned at the same site, such as
censuses throughout a migration or winter period.

Method 3 — Unmarked Strip Transect. In this
variant of Method 2, with no permanent transect
markers, the observer must estimate a census width
to either side of the transect line, a factor that re-
duces the accuracy and precision of the census.
However, the technique requires no setup time, and
is useful for surveying areas when only single cen-
suses are possible.

Method 4 — Linear Density Transects. Several
variations of this technique all obtain a linear density
estimate along shorelines, rather than the conven-
tional density measured over two-dimensional areas.
The observer moves parallel to the shoreline, walk-
ing or riding a vehicle (three-wheeled motorcycles
are effective on arctic beaches), counting birds con-
tinually. The counting width may be constant, cen-
tered on the water's edge or landward from the
water's edge, or it may be variable, including for

562

Marsh and Shorebirds

example, all of a beach from water to dunes or all of
a specific habitat, such as the wave-washed zone of
a sandy beach or the rocky intertidal zone below
a cliff. Densities are typically expressed as birds per
kilometer (0.6 mile). Linear density transects share
the same properties of accuracy, precision, and re-
quired training as comparable area transect methods.

Method 5— Roosting Flock Counts. If con-
sistent roosting areas are known from prior observa-
tions, specific areas can be treated as sample plots,
or general areas can be searched for the daily loca-
tion of total roosting flocks. This method clearly
depends on properly choosing sampling times, usu-
ally dusk or midday for inland areas and high tide for
coastal areas. It may provide an absolute density for
a roosting area, but the real values of such a count
are usually ( 1 ) as an index to the population using
nearby areas for foraging or ( 2 ) in some circumstan-
ces, as an estimate of the total local population size.
In counting roosting flocks, accuracy and precision
are sometimes reduced by the inability to visually
separate and count all individuals in very large flocks
of small birds. A vantage point providing some eleva-
tion above the roosting flock is helpful in these
situations.

Method 6 — Total Population Counts. Be-
cause migrating shorebirds are so concentrated in
specific, limited habitats, total population size esti-
mates for a local area may be obtainable. This is
unusual compared with most avian censusing situa-
tions and is often the most useful of population vari-
ables. This technique also requires prior knowledge
of the local distributions of shorebird habitats or the
behavior of shorebird species, and usually more than
one observer censusing different areas simultane-
ously. It has been used to estimate total populations
of a single estuary (Page et al. 1979; Connors et al.
1981 ) or of hundreds of kilometers of a coastline
(J. P. Myers, pers. commun.). This technique is costly
in personnel and its accuracy depends on the physi-
cal layout of the area censused and on movements of
birds during the census period. However, the num-
bers obtained are frequently the most useful in man-
agement studies.

Method 7 — Aerial Counts. As discussed
above, transect counts from aircraft are not usually
accurate for shorebirds because of shorebirds' small
size and their frequent occurrence in large flocks
of several species. For cranes, however, the large size
of the birds makes this approach feasible and effi-
cient when large areas must be sampled.

Method 8 — Photographic Counts. For very
large numbers of roosting birds, especially shore-
birds or sandhill cranes in migration, photographs
taken from aircraft or from an elevated fixed point
near a roost may provide the best means of obtain-
ing an accurate count. A photograph can be taken

quickly, before birds flush or while they are flying
but still concentrated. It can then be analyzed me-
thodically and carefully at a later time.

Discussion

Shorebirds in migration occupy such a wide
range of habitats that profitable generalizations are
difficult. However, in most migrant shorebird situa-
tions, the open nature of the habitat and the conspic-
uousness and high densities of the birds make
shorebird censusing relatively simple. In these situa-
tions, monitoring shorebird densities directly is usu-
ally preferable to measuring habitat features as a
means of evaluating probable shorebird use. For the
same reasons, counting actual shorebird totals in
monitored areas can often give absolute density esti-
mates at little extra cost over simple presence/ab-
sence or relative density estimates. However,
interpreting census results, in light of information on
local habitat distributions and bird movements, is
more centrally important than for most other bird-
habitat groups, as is the proper choice of census
conditions, especially with tidal fluctuations. Because
of the disjunct distribution of most migrant and win-
tering shorebird habitats, estimates of total size of
temporarily discrete local populations are frequently
possible and may be more meaningful than estimates
of bird densities in particular habitats under particu-
lar conditions.

Finally, dealing with populations in migration at
a site carries certain complications that differ from
breeding population studies. Most notable is the
inherent instability of the population, which requires
frequent regular censuses to estimate peak or mean
densities during a migration period. Furthermore,
unless rates of individual turnover at a site are
known, single census results or even continuous
census results do not provide estimates of total pop-
ulation numbers using the site — a relationship that
is markedly different from that which applies to
breeding bird studies.

MARSHES— RAILS AND COOTS

Five of the six North American rail species
breed and winter in the western U.S. where they
constitute a characteristic segment of the avifauna of
fresh, brackish, and saltmarshes. They are secretive
and inconspicuous, spending most of their time in
wetlands surrounded by dense vegetative cover,
a situation that has led to the development of spe-
cialized census techniques. Coots and gallinules nest
in some of the same marshes, and coots in winter
are widespread in open water situations, both fresh-
and saltwater, where waterfowl are common.

Marsh and Shorebirds

363

Habitat Features

Rails are among the strictest of all bird groups
in their consistency of habitat use. They seldom ven-
ture out of marshes or the muddy borders of
marshes. This basic requirement of vegetated wet-
lands provides the best, albeit rough, guide to assess-
ing rail habitat. Within this general description,
however, suitable marshlands vary from freshwater
to brackish to saltwater; from edges of ponds, lakes,
or rivers to large estuaries; and from moderately low
vegetation, such as pickleweed (Salicornia sp.), to
tall cattails (Typha sp.), sedges (Scirpus sp.), and
rushes (/uncus sp.). Species habitat preferences and
seasonal habitat use patterns differ (Repking and
Ohmart 1977; Gill 1979; Glahn 1974; Griese et al.
1980; Sayre and Rundle 1984). Even within a single
species and a single region, questions remain con-
cerning the differences in rail population densities
between superficially similar marshes, as with popu-
lations of clapper rail (Rallus longirostris) in San
Francisco Bay (Gill 1979; T. Harvey, pers. commun.).

Coots are abundant and fairly conspicuous at
wetland nesting sites in western North America,
where they use a wide variety of marsh and pond

types. Habitat requirements, in terms of pond size,
permanency, and characteristics of emergent vegeta-
tion at nest sites, have been studied by Sugden
(1979) and Nudds ( 1982).

Population Measurement Techniques

Presence. Because of the difficulty of observing
secretive rails in dense cover, presence and absence
is sometimes the only information one can obtain.
There are no special techniques for determining
presence, however, other than direct observation of
individuals, usually during foraging periods, or calls
and observations recorded by the methods listed
under Relative or Absolute Density.

Relative Density.

Method 1 — Breeding Vocalization. The

breeding vocalization technique relies on naturally
occurring vocalizations of territorial breeding rails
during the early breeding season (April, May, and
June in most areas). An observer familiar with rail
calls listens at selected points within or on the
periphery of a marsh. Listening periods should be of

Coots in an open-water habitat.

364

Marsh and Shorebirds

the same duration and similar in time of day, date,
and weather conditions at all sites to be compared.
Relative density can be expressed as calling birds per
time of observation or per listening station.
Recordings or rail vocalizations are available for
training purposes. Precision and accuracy of this
technique are unknown, but may be low. Costs are
also low.

recorded within the local region of the species'
range, is recommended (Johnson et al. 1981 ). A con-
sistent schedule of tape playback duration and listen-
ing duration must be maintained, on the order of 1
min. of playback followed by 5 min. of listening,
possibly repeated, at each station. Training and data
treatment are similar to Method 1.

Method 2 — Tape Playback Response. The

tape playback technique is more widely used, re-
quires little extra effort, and increases the chance of
hearing a rail that is present during the station
count. Glahn (1974) located 71% more territories of
rails in Colorado using playback techniques, com-
pared with territory mapping without playback. Ac-
curacy is therefore increased but may still be low,
depending mainly on the consistency with which
rails respond.

Although taped response techniques are most
effective during the breeding season, they also have
been used with appropriate taped calls to assess
winter populations (Tomlinson and Todd 1973; Mar-
ion et al. 1981 ). Use of a variety of calls, preferably

Method 3 — Flood Tide Counts. In tidal areas,
some rails can be located visually during extremely
high tides, when almost all marsh vegetation is inun-
dated. In San Francisco Bay, clapper rails are forced
onto small patches of high ground or floating debris
during the highest tides of the year, which occur
during daylight only in winter. They can be located
from vantage points surrounding the marsh or in
extensive marsh areas from a canoe or an airboat
traveling through the marsh. Detection of birds can
be increased if the observer works with a competent
hunting dog (T. Harvey, pers. commun. ).

With sufficiently high tides and scarce, accessi-
ble refuges, this method can be accurate in terms of
total birds sighted, but the distance over which birds
have traveled to a refuge is generally unknown. For

Gallinule in a ryp/ca/ marsh habitat.

Marsh and Shorebirds

365

this reason it is listed first as a relative density tech-
nique, but this technique and the two previous tech-
niques can all be adapted to absolute density
estimation.

Method 4 — Open Water Counts (Coots).

American coots are the only species in this group to
regularly forage in open water away from emergent
vegetation. Direct visual counts, made from shore or
in boats, can provide an index of numbers of coots
nesting in wetlands. The fraction of the population
missed will remain unknown and may vary in differ-
ent types of wetlands. All counts of comparative
purposes should be made .at the same time of day
and over as short a calendar interval as practical.

Absolute Density.

Method 1 — Breeding Vocalization. Combin-
ing the breeding vocalization technique with a terri-
tory mapping effort, a variable width line transect
method, or a variable circular plot method permits
estimation of absolute densities. Accuracy is proba-
bly low in most situations, however, because an un-
known percentage of rails present will not call
during the census period. All three approaches re-
quire the observer to estimate distances to unseen
vocalizing rails, a difficult task.

Method 2 — Tape Playback Response. Simi-
larly, the tape playback method is the combination
of the technique discussed under Relative Density
and the area estimating techniques of mapping, varia-
ble width line transect, or variable circular plot
method. It suffers the same limitations of accuracy
and precision as Method 1 in this section, but the
tape playback elicits vocalizations at a higher fre-
quency than the natural calling rates, so fewer birds
are missed.

Method 3 — Flood Tide Counts. Methods for
this technique are described under Relative Density.
However, when discrete patches of saltmarsh or
extensive areas of continuous saltmarsh can be
searched thoroughly, absolute density measurements
result. Local movements of birds to refuges within
their population area do not distort the density esti-
mate if the entire population area is censused. (Tran-
sect sampling, under these circumstances, may
produce large errors. ) If all potential refuges can be
carefully searched, accuracy and precision of these
estimates should be high. Costs may also be high,
however, because of logistics, especially with the use
of airboats at high tides. This method is only useful
in winter, when high tides occur during daylight
hours.

Method 4 — Rope-Drag Nest Mapping. As

described for Tundra — Breeding Shorebirds, two ob-
servers walk through potential nesting habitat drag-
ging a rope (about 50 m [165 ft] of 1/4-in. rope)
between them; a third observer walks behind, re-
cording locations of nests and adults when birds
flush. The technique greatly increases the numbers
of nests located and is especially valuable when used
in conjunction with Methods 1 and 2, because it
provides an independent estimate of breeding den-
sity. Furthermore, information from one method can
sometimes be added to information from the other
when clear discrepancies occur. The combined cen-
sus may be more accurate than either census alone.

Method 5 — Nest Searches (Coots). Direct
nest searches in marshes can provide reliable esti-
mates of coot nest densities (Sugden 1979). This can
be combined with observations of bird-nesting be-
havior and territoriality to narrow the search areas,
but will usually require a time-consuming field effort.

Discussion

Rails in marshes share one characteristic with
shorebirds in migration. Their habitats are often
clearly distinguished from surrounding areas (marsh
or mudflat versus uplands, for example) and are eas-
ily recognized and mapped. Habitat evaluation at this
very general level is straightforward. At the finer
scale, which requires distinguishing better from
poorer marshlands, knowledge is limited.

The two bird groups contrast markedly in ease
and accuracy of censusing populations for absolute
density estimates. For all the reasons that shorebirds
are extremely tractable, rails are difficult to treat.
Their secretiveness and the densely vegetated, rela-
tively inaccessible habitats they favor impose stum-
bling blocks in any rail population studies. The
methods just described have limited accuracy and
precision, but provide useful estimates of population
sizes in most situations.

CONCLUSIONS

The 54 species of marsh and shorebirds that
occur in Alaska and the western States cover such a
wide range of sizes, habitats, and natural histories
that the difficulty of any neat generalizations of bird-
habitat relationships or of preferred census tech-
niques is not surprising. Most of these species are
relatively easy to assign to large-scale habitat types
(marsh versus forest, for example) and some can be
partially defined by minor habitat differences in habi-
tat features, such as vegetation height or water
depth. In almost no situations beyond a local study

366

Marsh and Shorebirds

area, however, can one predict densities of bird use
based on habitat measurements with a sufficient de-
gree of accuracy to justify such an approach. For
most shorebirds and cranes, this is not a serious lack,
because these species are usually so visible in their
breeding and non-breeding habitats that a direct
censusing method can be efficient and acceptably
accurate. Rails, by their secretive habits and the
dense vegetative cover in their habitats, present a
much more difficult censusing challenge.

For these reasons, presence and absence tech-
niques are usually warranted only for rails and gallin-
ules. Relative density census techniques have much
wider application for many shorebirds, rails, coots,

and cranes. With an increase in effort and minor
modifications, many of these techniques provide ab-
solute density measurements. The range of absolute
density census techniques is especially large for
shorebirds nesting on tundra or migrating along
shorelines, because in these situations a large mix-
ture of species often occurs at high densities under
conditions of good detectability. Even these generali-
zations, however, gloss over individual species differ-
ences that may be of major importance in selecting
research and management techniques. The Literature
Cited section in this chapter should be consulted
directly for additional details concerning particular
species, areas, or habitats before field work is
initiated.

Marsh and Shorebirds

567

LITERATURE CITED

BENT, A.C. 1926. Life histories of North American marsh
birds. U.S. Natl. Mus. Bull. 135.

. 1929. Life histories of North American shorebirds.

U.S. Natl. Mus. Bull. 146.

BERGMAN, R.D., R. HOWARD, K ABRAHAM, and M.
WELLER. 1977. Water birds and their wetland re-
sources in relation to oil development at Storkersen
Point, Alaska. U.S. Dep. Inter., Fish and Wildl. Serv.
Resour. Publ. 129. 39pp.

BRYANT, D.M. 1979. Effects of prey density and site char-
acter on estuary usage by overwintering waders
(Charadrii). Estuarine and Coastal Marine Science 9:
369-384.

CONNORS, P.G. 1984. Ecology of shorebirds in the Alas-
kan Beaufort littoral zone. Pages 403-416 in Barnes,
P., E. Reimnitz, and D. Schell, eds. The Alaskan Beau-
fort Sea: Ecosystems and Environments. Academic
Press.

, J.P. MYERS, C.S.W. CONNORS, and FA. PITELKA.

1981. Interhabitat movements by sanderlings in rela-
tion to foraging profitability and the tidal cycle. Auk
98:49-64.

, , and FA. PITELKA. 1979. Seasonal habitat

use by arctic Alaskan shorebirds. Studies Avian Biol.
2:101-111.

DREWIEN, R.C. 1973. Ecology of Rocky Mountain greater
sandhill cranes. Ph.D. dissertation, Univ. Idaho, Mos-
cow. 152pp.

EMLEN, J.T. 1971. Population densities of birds derived
from transect counts. Auk 88:323-342.

. 1977. Estimating breeding season bird densities

from transect counts. Auk 94:455-468.

GERSTENBERG, R.H. 1979. Habitat utilization by winter-
ing and migrating shorebirds on Humboldt Bay, Cali-
fornia. Studies Avian Biol. 2:33-40.

GIBSON, F. 1971. The breeding biology of the American
avocet (Recurvirostra americana) in central Oregon.
Condor 73:444-454.

GILL, R.E., Jr. 1979. Status and distribution of the Califor-
nia clapper rail (Rallus longirostris obsoletiis). Cali-
fornia Fish and Game 65:36-49-

and CM. HANDEL. 1981. Shorebirds of the eastern

Bering Sea, Pages 719-738 in Hood, D.W. and J.A.
Calder, eds. Eastern Bering Sea Shelf: Oceanography
and Resources, Vol. 2. Univ. Washington Press, Seattle.
-, MR. PETERSEN, and P.D. JORGENSEN. 1981. Birds

of the north Alaska Peninsula, 1976-1980. Arctic

34:286-306.
GLAHN, J.F. 1974. Study of breeding rails with recorded

calls in north-central Colorado. Wilson Bull. 86:206-

214.
GOSS-CUSTARD, J.D. 1970. The response of redshank

(Tringa totanus L.) to spatial variations in the density

of their prey. J. Animal Ecol. 39:91-1 13.
GRAUL, WD. 1975. Breeding biology of the mountain

plover. Wilson Bull. 87:6-31.
GRIESE, H.J., R.A. RYDER, and C.E. BRAUN. 1980. Spatial

and temporal distribution of rails in Colorado. Wilson

Bull. 92:96-102.
HAMILTON, R.C. 1975. Comparative behavior of the

American avocet and the black-necked stilt (Recurvi-

rostridae). A.O.U. Monographs, 17.

HOLMES, R.T. and C.P. BLACK. 1973. Ecological distribu-
tion of birds in the Kolomak River-Askinuk Mountain
Region, Yukon-Kuskokwim Delta, Alaska. Condor
75:150-163.

and FA. PITELKA. 1 968. Food overlap among coex-
isting sandpipers on northern Alaska tundra. System-
atic Zoology 17:305-318.

ISLEIB, M.E.P. 1979. Migratory shorebird populations on
the Copper River Delta and eastern Prince William
Sound, Alaska. Studies Avian Biol. 2:125-130.
JARVINEN, O. and R.A. VAISANEN. 1975. Estimating rela-
tive densities of breeding birds by the line transect
method. Oikos 26:316-322.
JENNI, DA., R.L. REDMOND, and IK. BICAK. 1982. Be-
havioral ecology and habitat relationships of long-
billed curlews in western Idaho. Unpubl. Rep. to U.S.
Dep. Inter., Bur. Land Manage., Boise, ID. 234pp.
JOHNSON, R.R., B.T. BROWN, L.T. HAIGHT, and J.M.
SIMPSON. 1981. Playback recordings as a special
avian censusing technique. Studies Avian Biol. 6:68-
75.
JOHNSON, S.R. and W.J. RICHARDSON. 1981. Beaufort
Sea — Barrier Island — lagoon ecological processes
studies: Final report, Simpson Lagoon: birds. Pages
109-383 in Environmental Assessment Alaskan Conti-
nental Shelf. Vol. 7.

KRAPU, G.L. 1974. Feeding ecology of pintail hens during
reproduction. Auk 91:278-290.

LATKA, DC. and J.W. YAHNKE. 1984. Simulating roosting
habitat of sandhill cranes and validation of suitability -
of-use indices. Unpubl. abstract of poster presented
at Wildlife 2000 Conference, Fallen Leaf Lake, CA.

LITTLEFIELD, CD. and R.A. RYDER. 1968. Breeding biol-
ogy of the greater sandhill crane on Malheur National
Wildlife Refuge, Oregon. Trans. North Am. Wildl.
Nat. Resour. Conf. 33:444-454.

MARION, W.R., T.E. O'MEARA, and D.S. MAEHR. 1981. Use
of playback recordings in sampling elusive or secre-
tive birds. Studies Avian Biol. 6:81-85.

MARTIN, P.D. and C.S. MOITERET. 1981. Bird populations
and habitat use, Canning River Delta, Alaska. Unpubl.
Rep. to U.S. Dep. Inter., Fish and Wildl. Serv., Fair-
banks, AK 196pp.

MICKELSON, PS., J.S. HAWKINS, DR. HERTER, and S.M.
MURPHY. 1980. Habitat use by birds and other wild-
life on the Eastern Copper River Delta, AK. Unpub-
lished Report. Alaska Coop. Wildl. Res. Unit, Univ.
Alaska, Fairbanks. 189pp.

MYERS, J P., P.G. CONNORS, and FA. PITELKA. 1979.

Territory size in wintering sanderlings: The efifects of
prey abundance and intruder density. Auk 96:551-
561.

and FA. PITELKA. 1980. Effect of habitat condi-
tions on spatial parameters of shorebird populations.
Rep. to the U.S. Dep. Energy. 82pp.

NUDDS, T.D. 1982. Ecological separation of grebes and

coots: Interference competition or microhabitat selec-
tion? Wilson Bull. 94:505-514.

ORING, L.W., D.B. LANK and S.J. MAXSON. 1983. Popula-
tion studies of the polyandrous spotted sandpiper.
Auk 100:272-285.

PAGE, G.W., L.E. STENZEL, D.W. WINKLER, and C.W.

SWARTH. 1983. Spacing out at Mono Lake: Breeding
success, nest density, and predation in the snowy
plover. Auk 100:13-24.

, , and CM. WOLFE. 1979. Aspects of the

occurrence of shorebirds on a central California estu-

368

Marsh and Shorebirds

ary. Studies Avian Biol. 2:15-32.

RALPH, C.J. and J.M. SCOTT, eds. 1981. Estimating num-
bers of terrestrial birds. Studies Avian Biol. 630pp.

RANDS, M.R.W. and J.P. BARKHAM. 1981. Factors control-
ling within-flock feeding densities in three species of
wading bird. Ornis Scandinavica 12:28-36.

RECHER, H.F. 1966. Some aspects of the ecology of mi-
grant shorebirds. Ecology 47:393-407.

REDMOND, R.L. 1984. The behavioral ecology of long-
billed curlews (Numenius americanm) breeding in
western Idaho. Ph.D. dissertation, Univ. Montana,
Missoula.

, T K BICAK, and DA. JENNI. 1981. An evaluation

of breeding season census techniques for long-billed
curlews (Numenius atnericanus). Studies Avian Biol.
6:197-201.

RENAUD, WE. 1979. The piping plover in Saskatchewan.
Blue Jay 37:90-103.

REPKING, C.F. and R.D. OHMART. 1977. Distribution and
density of black rail populations along the lower
Colorado River. Condor 79:486-489.

RILEY, J. L. 1982. Habitats of sandhill cranes in the south-
ern Hudson Bay lowland, Ontario. Canada Field-Nat.
96:51-55.

RYAN, MR. 1985. Marbled godwit habitat selection in the
northern prairie region. J. Wildl. Manage., in press.

SAYRE, M.W. and WD. RUNDLE. 1984. Comparison of
habitat use by migrant soras and Virginia rails. J.
Wildl. Manage. 48:599-605.

SENNER, S.E. 1979. An evaluation of the Copper River
Delta as a critical habitat for migrating shorebirds.
Studies Avian Biol. 2:1 31- 146.

SHD2LDS, G.F. and L.J. PEYTON. 1979. Avian community
ecology of the Akulik-Inglutalik River Delta, Norton

Bay-Alaska. Environmental Assessment, Alaskan Conti-
nental Shelf, Ann. Reps. 5:608-710. NOAA, Boulder,
CO.

SPLNDLER, MA. 1978. Bird populations and habitat use in
the Okpilak River Delta area, Arctic National Wildlife
Range, Alaska, 1978. Unpubl. Rep. to U.S. Dep. Inter.,
Fish and Wildl. Serv., Fairbanks, AK 86pp.

STEWART, RE. and H.A. KANTRUD. 1971. Classification of
natural ponds and lakes in the glaciated prairie re-
gion. U.S. Dep. Inter., Fish and Wildl. Serv. Resour.
Publ. 92. 57pp.

SUGDEN, L.G. 1979. Habitat use by nesting American

coots in Saskatchewan parklands. Wilson Bull. 91:599-
607.

TOMLLNSON, RE. and R.L. TODD. 1973- Distribution of
two western clapper rail races as determined by
responses to taped calls. Condor 75:177-183-

TROY, DM., D. HERTER, and R. BURGESS. 1983. Prudhoe
Bay Waterflood Project tundra bird monitoring pro-
gram. Unpubl. Rep. to Dep. Army, Corps of Engineers,
Anchorage, AK. 98pp.

WEBER, M.J., PA. VOHS, Jr., and L.D. FLAKE. 1982. Use of
prairie wetlands by selected bird species in South
Dakota. Wilson Bull. 94: 550-554.

WESELOH, D.V. and L.M. WESELOH. 1983 Numbers and
nest site characteristics of the piping plover in central
Alberta, 1974-1977. Blue Jay 41:155-161.

WIENS, J.A. 1973. Pattern and process in grassland bird
communities. Ecol. Monogr. 43:237-270.

WILLIAMSON, F.S.L., M.C. THOMPSON, and J.Q. HINES.
1966. Avifaunal investigations, Pages 437-480 in Wili-
movsky, N. J., ed. Environment of the Cape Thompson
Region, Alaska. U.S. Atomic Energy Commission, Oak
Ridge, TN.

Marsh and Shorebirds

369

18

WATERFOWL

Robert L. Eng

Montana State University
Bozeman, MT 59717-0001

Editor's Note: Waterfowl, because of their economic
and recreational value have received much atten-
tion for many years. And because of their obligate
association with wetlands, habitat has long been a
concern with these species. Thus, inventory, moni-
toring and management of waterfowl habitat has
a long history compared to species groups whose
management has focused on population manage-
ment. However, waterfowl biology has become a
specialized discipline and has tended to focus on
major wetland areas. Yet large numbers of water-
fowl depend on smaller wetlands scattered through-
out the West. These areas, if managed at all, are
probably managed by biologists with no special-
ized expertise in waterfowl. This chapter provides
such biologists an introduction to inventory and
monitoring of waterfowl and waterfowl habitat.

INTRODUCTION

The term waterfowl includes ducks, geese, and
swans. In the most recently published classification, a
single family, the Anatidae, is separated into two
subfamilies, the Anserinae (whistling-ducks, swans,
and geese) and the Anatinae (ducks; Check- List of
North American Birds, American Ornithologists
Union, 1983). Both subfamilies are further divided
into species, groups, or tribes — the Anserinae, into
three and the Anatinae, into five.

Most species of waterfowl are migratory to
some extent, with many participating in lengthy sea-
sonal movements. Thus habitat features for different
species or species groups will vary considerably
throughout the year, both as a result of geographic
availability and by demands placed upon the species
as a result of seasonal behavior.

As with upland game birds, most early water-
fowl investigations emphasized the reproductive
season (Chabreck 1979). This is somewhat under-
standable because it covers the period when off-
spring are produced. The nesting season also
embraces the time of the year when waterfowl habi-
tat is most important, because self-imposed isolation
of pairs results in a density far below that which may
be found during other seasons. Also, nesting habitat
seemingly was the most threatened by land uses
by man and consequently was thrust into the
limelight.

In recent years, attention has increased on win-
tering and migration habitat. Even though the gregar-
ious nature of waterfowl during these periods
permits rather large numbers in relatively small
areas, winter habitats in parts of the Atlantic Flyway
have deteriorated to the extent that they are no
longer suitable for ducks (Addy 1964). As was
pointed out by Chabreck (1979), dabbling ducks
may spend two-thirds of the year on the winter habi-

Waterfowl

371

tat, and the quality of that habitat may greatly influ-
ence their reproductive performance in the spring.

The habitat features for the various tribes of
waterfowl appear more distinguishable for the breed-
ing season than for other times of the year. This
may in part be a result of the greater investigative
effort during the breeding season. It may also be
a result of less diverse activities by all species during
winter (largely feeding and loafing), permitting a
more intense use of a resource by many members of
the same or different species. Thus, in this chapter,
the habitat features for the nesting season are dis-
cussed by species groups for the more widely dis-
tributed or numerically important tribes with
references to individual species when deemed perti-
nent. Winter habitat is discussed in a more general
fashion for the entire family.

Trumpeter swans, adults, and cygnet.

HABITAT FEATURES CORRELATED WITH
WATERFOWL

Swans (Cygnini)

Two species of swans are native to North Amer-
ica. The tundra swan (Cygnns columbianus), until
recently called the whistling swan, is the most abun-
dant; about 150,000 were reported in 1971 (Bellrose
1978). The trumpeter swan (C buccinator), once
believed to number less than 100 birds in the U.S.
and Canada, now constitutes a population of about
1,500 plus about 9,000 in Alaska.

Physical Features. As the name implies, tundra
swans nest in the arctic from north of Hudson Bay to
the coast of Alaska (Bellrose 1978). Lensink (in
Bellrose 1978) reported nest site locations as
follows: 50% on the shore of a lake or pond within
18 m (20 yd) of water, 30% on small islands or
points, and the remainder in a variety of situations.
In most instances the nest was elevated
aboveground.

Trumpeter swans in Alaska showed a nest pref-
erence for small beaver impoundments over other
larger bodies of water; one-half of the 40 nests were
located in beaver ponds of less than 5.6 ha ( 14 a.).
This preference appeared to be related to a stable
water level and little wave action, which promoted
the development and maintenance of extensive
stands of emerging and floating vegetation in which
and from which the nests were constructed (Hansen
et al. 1971). Trumpeter swans in the Targhee Forest
in Idaho and Wyoming use older, more eutrophic
lakes, which are relatively shallow (average 1.2 m [4
ft]), have a large part of the surface covered with
vegetation (80% ), and have at least 25% of their to-
tal area less than 1 m deep. Aquatic vegetation is
used for the nest site and nest material (Maj 1983).

Vegetation.

Structure. Whereas the trumpeter swan nests
almost exclusively over water in emergent aquatic
vegetation, the tundra swan seems to prefer more
upland sites not associated with emergent vegetation
(Hansen et al. 1971), but close enough to water
that water was often exposed when vegetation was
removed from a circle around the nest (Lensink
in Bellrose 1978). Both species also feed largely on
aquatic plants, both submergents and emergents,
although the tundra swan will feed on upland
sources when marshes freeze (Nagel 1965). Upland
feeding in cornfields by tundra swans has become
relatively common along Chesapeake Bay and can be
readily observed during spring migration through
Montana.

Species Composition. Tundra swans, as inhab-
itants of the open tundra and somewhat more terres-
trial than trumpeter swans, construct nests from
material which may include mosses, grasses, and
sedges (Carex sp.). In Alaska, nests of trumpeter
swans were usually constructed with the stems and
rhizomes of Carex sp., Equisetum fluviatile, or Po-
tentilla sp. (Hanson et al. 1971 ). In Idaho and Wyo-
ming, nests were frequently in and constructed of
Typha sp., Scirpus sp., or Carex sp. Occasional nests
were found on islands or beaver (Castor canaden-
sis) lodges, near the water's edge, and constructed
largely of aquatic vegetation (Maj 1983).

Both species feed largely on aquatic plants, in-
cluding the seeds, stems, and tubers of several spe-
cies. Sago pondweed (Potamogeton pectinatus) is a
prominent plant in the diet of tundra (Sherwood
I960; Nagel 1965) and trumpeter swans (Bellrose
1978). Other plants frequently mentioned as food
are widgeon grass (Ruppia sp. ), duck potato (Sagit-
taria sp. ), other pondweeds (Potamogeton sp. ),
and water milfoil (Myriophyllum sp. ).

372

Waterfowl

Geese (Anserini)

Of the several species of geese that winter in
the U.S., only one, the Canada goose (Branta cana-
densis), nests within the U.S. south of Canada. This
species has great morphological and ecological
breadth, with sizes ranging from a large duck to a
small swan, and nesting distribution from farm ponds
to forested lakes to arctic tundra. This diversity has
led to considerable discussion as to the number of
identifiable races. Bellrose (1978) illustrated the
ranges of 1 2 populations of this species, established
largely on the basis of the close alliances shown
during the winter. Two races nest south of Canada,
the giant Canada goose (B. c. maxima) and the
western Canada goose (B. c. moffltti), whereas at
least nine other races nest across the breadth of the
arctic or northern Canada (Bellrose 1978).

Physical Features. Nesting habitat for Canada
geese is extremely diverse, particularly for the two
races nesting south of Canada. This is probably
because of the diverse habitats into which the two
races have been introduced through management
(transplant) programs. Canada geese have been
observed to nest in the vicinity of lakes, ponds,
marsh, and river habitat; on the ground; on muskrat
houses; in trees in unused raptor, heron, and magpie
(Pica pica) nests; and on cliffs. However, in spite
of the great variety of reported nest sites, the
preference for islands as nest sites appears to
dominate (Hammond and Mann 1956; Ewaschuk and
Boag 1972; Giroux 1981; Giroux et al. 1983). This
preference is also shown by some of the smaller
arctic nesting Canada geese (Maclnnes 1962), but
less so by the colonial nesting snow geese {Chen
caerulescens; Lemieux 1959) and Ross' goose (C
rossii; Ryder 1967).

Terrestrial nesting Canada geese.

Geese are very terrestrial compared with most
ducks and swans. Consequently, secure brood-rear-
ing areas are an integral part of breeding habitat. The
habitat usually consists of a meadow or flat area pro-
viding a grazing area with good visibility and access
to water. Such areas are frequently near the nest
sites but some birds nesting on western rivers may
move several miles to congregate on desirable
brood-rearing areas.

Another habitat requirement of geese used dur-
ing the breeding season, but usually distinct from
breeding areas, is the molting area. Nonproductive
geese, primarily yearlings (65% ), 2-year-olds (20% ),
and older birds ( 1 5% ), leave the breeding areas in
late spring for traditional molting areas (Krohn and
Bizeau 1979). The migration is usually in a northerly
direction. Although several molting areas are present
in the western U.S., many of the Canada geese from
this area migrate to the Northwest Territories to
molt (Krohn and Bizeau 1979). Key physical charac-
teristics of a molting area are a large body of water
for security, a good food supply, and little human
disturbance. One such area in southwestern Mon-
tana, under the jurisdiction of the U.S. Bureau of
Land Management, annually harbors 8,000 to 10,000
geese. This site has received special consideration
to maintain its integrity as an important molting area
(Hildebrand 1979).

Vegetation.

Structure. Protective cover near goose nests
seems optimal when it at least breaks the outline of
the incubating bird, but is not so high as to prevent
visibility from the nest. Thus, in areas where
vegetation was relatively short and nesting densities
high, geese selected islands that had greater
coverage of forbs and grass (Giroux 1981).
Conversely, where vegetation was relatively high and
nesting densities low, geese selected islands with
less dense vegetation (Kaminski and Prince 1977). In
areas where high densities of geese are nesting,
visual barriers between nest sites (vegetation or
topography) permit pairs to nest closer to one
another with less aggressive encounters (Giroux
1981).

Structure of vegetation in brood-rearing areas
should be short enough to permit visibility, at least
by the adults. Grazing geese generally avoid dense or
high vegetation because both adults and goslings
are often flightless.

Species Composition. Where nesting cover is
concerned, vegetation structure is more critical than
species composition. Migrating and wintering geese
make extensive use of agricultural crops, including
cereal grains and green forage. However, on brood-
rearing areas where flightless birds must graze adja-

Waterfowl

373

cent to the water, a wide variety of green forage
is utilized. At Canyon Ferry Wildlife Management
Area near Townsend, Montana, where about 1,000
goslings are produced annually, I have observed
Canada geese grazing heavily on wheatgrass (Agropy-
ron sp. ), spikerushes (Eleocharis sp. ), new shoots of
bulrush (Scirpus sp. ), alfalfa (Medicago sativa), and
reed canarygrass (Phalaris arundinacea), while
virtually ignoring abundant plants of sweet clover
(Melilotus officinalis).

Dabbling Ducks (Anatini) and Wood Ducks
(Cairinini)

Under the subfamily Anatinae (ducks), more
than one tribe may be treated as a single species
group because they will frequently share some com-
mon habitat feature. Although many ducks have
breeding ranges extending far into Canada, unlike
most geese, many also nest south into the U.S. Some
have very extensive breeding ranges, being abundant
on both the prairies and the tundra (e.g., pintail
[Anas acuta] whereas the black duck (A rubripes)
and the mottled duck (A fulvigula) are consider-
ably more restricted geographically. Many dabbling
ducks and some diving ducks reach their highest
density of breeding birds in the north-central U.S.
and south-central Canada in an area called the Prairie
Pothole region. Among the dabblers in this group
are the mallard (A platyrhynchos), pintail, gadwall
(A strepera), blue-winged teal (A discors), and
shoveler (A clypeata).

Physical Features. Most species of this group are
birds that breed primarily in the prairies, parklands,
or tundra, all of which provide relatively open
shallow marshes. Notable exceptions include the
black duck, which is found in fresh and marine
marshes, swamps, and lakes (and prefers wooded
marshes for breeding), and the wood duck (Aix
sponsa), which prefers small, quiet inland streams
and ponds near woodland.

Because all of these ducks are territorial to
some degree, breeding habitat is enhanced with an
abundance of water areas of varying sizes, depths,
and configurations that provide a diversity of aquatic
vegetation structure. Patterson (1976) found a sea-
sonal shift in habitat requirements by prairie-dwell-
ing ducks, where the breeding pairs disperse to all
bodies of water, irrespective of the productivity of
the waters. As the breeding chronology progressed,
the broods and fledged ducks (which had greater
demands for nutrition but less demands for isolation)
moved selectively to the more productive waters,
thus allowing full utilization of a spatially heteroge-
neous aquatic habitat.

Stock ponds often provide valuable habitats
(Bue et al. 1964). Although in comparatively low

densities (seldom exceeding 1/km [1/247 a."] on arid
grazing lands in the West), these ponds are often
more stable (Brewster et al. 1976) and more pro-
ductive over the years than many natural wetlands.
The merits of different types of stock ponds as wa-
terfowl habitat were discussed by Eng et al. ( 1979).

Stock pond in Great Basin Desert used by waterfowl.

As with geese, several species of ducks are at-
tracted to island habitats for nesting. From this spe-
cies group, mallards and gadwalls are very likely
to nest on islands. In the absence of mammalian pre-
dation, a high nest success combines with a high
degree of homing in ensuing years by surviving
young, resulting in some rather high nest densities
(Duebbert 1982; Duebbert et al. 1983).

Vegetation.

Structure. Most dabbling ducks nest on the
ground. Black ducks nest in tree cavities and stumps
(Cowardin et al. 1967), whereas wood ducks are
obligate cavity nesters (Bellrose et al. 1964). The
mallard, although typically a ground-nesting bird,
will nest in a variety of elevated sites, natural and
artificial (Cowardin et al. 1967; Bishop and Barratt
1970).

Of the ground-nesting ducks in this group, the
pintail appears more prone to nest in the least cover
situation than any of the others. In a study of duck
nesting in intensively farmed areas, pintail nests were
in near-equal densities in summer fallow, mulched
stubble, standing stubble, and untitled uplands,
whereas 72% or more of other duck nests were in
the unfilled uplands (Higgins 1977).

Duck-nesting density in idle grasslands was stud-
ied in North Dakota for 3 years; the 3-year average
nesting density was 76.6 nests/km (30 nests/mi. ;

374

Waterfowl

Duebbert and Lokemoen 1976). Nesting density on
the idle acres for one of the years was four times
higher than that found on nearby agricultural lands
(farmed and grazed) during the same year. Vegeta-
tion measurements on the idle lands in May showed
an average of 90 dead stems/m (108/yd") and
heights of 71, 47, and 28 cm (28, 19, and 11 in.) on
dead stems, dead leaves, and live leaves, respectively.
The preference that nesting hens had for this vegeta-
tion density and structure was shown by flights up
to 1.6 km (1 mi.) from wetlands to nest in the idle
lands.

Species Composition. As with geese, when
vegetation is used as cover, species composition
appears to take a secondary role to structure. Several
species used heavily for nesting cover by dabbling
ducks in Alberta (Calamovilfa longifolia, Hordeum
jubatum, Spartina gracilis) had one common de-
nominator— they were unpalatable to cattle and con-
sequently provided the structure for cover (Keith
1961 ). The same could probably be said for dense
thickets of snowberry (Symphoricarpos sp. ) and rose
{Rosa sp.), which are frequently found on the prairie
and have had high usage by nesting ducks (Hines
and Mitchell 1983; Lokemoen et al. 1984).

Successful brood-rearing involves a proper com-
bination of cover and food. Aquatic cover can be
provided by a variety of emergent plants, some of
which provide cover only (Cattails [Typha sp.])
whereas others (bulrush, spikerush, arrowhead [Sag-
ittaria sp.], smartweeds [Polygonum sp.]) provide
food from tubers, foliage, or fruiting bodies, as well
as cover. Cattail, although often used as cover for
nesting ducks, can detract from the attractiveness of
a pond, at least to certain species. Keith (1961) re-
ported an increase in waterfowl usage of ponds
when shoreline cattail was reduced.

Submerged aquatics, such as water milfoil and
pondweeds combined with seeds of spikerush, were
used heavily by adult dabblers in the spring (Keith
1961). Ducklings showed a heavy use of spikerush
seeds suggesting a feeding pattern along the shore
where this plant was prevalent. Flying juveniles fed
more heavily on submergents, indicating more in-
tense feeding in the open water.

Major foods found in stock pond habitat in Mon-
tana include pondweeds, water milfoil, common
hornwort (Ceratophyllum demersum), Canadian
waterweed (Elodea canadensis), watercrowfoot but-
tercup (Ranunculus aquatilis), and longspike spike-
rush ( Eleocharis macrostachya). At least 5 years
must pass for newly constructed stock ponds to de-
velop aquatic vegetation to provide adequate food
and cover (Hudson 1983).

Bay Ducks (Aythyini) and Ruddy Ducks
(Oxyurini)

Five species of the tribe Aythyini, all of a single
genus, are common to North America. These five are
the canvasback (Aythya valisineria), redhead (A
americana), ring-necked duck (A collaris), greater
scaup (A marila), and lesser scaup (A affinis). The
ruddy duck (Oxyura jamaicensis), although a mem-
ber of a different tribe (stiff-tailed ducks), is included
here because it has some similar habitat require-
ments during the breeding season. These two tribes
and a third, Mergini, are often included under the
descriptive term diving ducks as opposed to the
dabbling ducks discussed earlier.

Although many diving and dabbling ducks may
be observed on the same bodies of water, particu-
larly during the breeding season, the morphological
differences between the two groups are reflected
in many of their activities and often in the segments
of a pond that each uses. Divers have lobed hind
toes and generally much larger feet than the dab-
bling ducks, which aid in their underwater swim-
ming capabilities. The legs on divers are located
farther back on the body which enhances their div-
ing ability, but reduces their walking capabilities —
resulting in little or no utilization of terrestrial
habitats.

Body contour, method for taking flight, and
feeding methods also separate divers from dabblers.
On the water, divers provide a much lower profile
than dabbling ducks; their shorter tail feathers pro-
vide an appearance of the back sloping down to the
water's surface. In taking flight, divers patter along
the water before lift-off, whereas the dabblers leap
immediately into the air. Divers commonly go well
below the surface to feed, whereas the dabbling
ducks simply tip to reach below the surface for
aquatic foods. Primarily because of the divers' inabil-
ity to maneuver on land, ducks observed feeding in
grain fields will be dabblers.

Physical Features. All except two of the species
within this group normally inhabit the open prairie
marshes. One of these exceptions, the ring-necked
duck, seems to show preference for marshes
bordered by woodland. The other, the greater scaup,
is primarily an inhabitant of marshy tundra. The
lesser scaup, although commonly found associated
with prairie marshes, also extends its breeding range
into boreal lakes and tundra habitat.

When nesting, canvasbacks, redheads, and ruddy
ducks all inhabit a variety of prairie marshes, sloughs,
and potholes with stands of emergent vegetation. Of
the three species, canvasbacks seem most prone to
nest in the small ponds; ruddy ducks and especially
redheads seem to favor the larger, deeper bodies
of water. This preference, in conjunction with its

Waterfowl

375

Ring-necked duck.

gregarious nesting habits, makes the redhead an im-
portant nesting species in many marshes of the Inter-
mountain West — 20-25 times more abundant than
canvasbacks. Although not evenly distributed
throughout the Intermountain West — nor the marsh
habitat — redheads in particular, but also ruddy
ducks, lesser scaup, and canvasbacks reach rather
high breeding densities in western marsh habitats.
Some Nevada, Utah, and Montana marshes attain very
high breeding densities (919 redheads/km [355/
mi.2] of wetlands on Salt Lake marshes; Bellrose
1978).

All of the species in this group are inclined to
nest closer to the water than most dabbling ducks, a
reflection of their close tie to the aquatic environ-
ment. Canvasbacks and ruddy ducks invariably nest
over water. Redheads nest over water, but also on
land (Keith 1961) and often on islands (Vermeer
1970; Carlsen 1984). Ring-necked ducks usually nest
within marshes; over 85% nest on small clumps of
floating vegetation or within clumps of marsh vegeta-
tion (Mendall 1958). Although both the greater and
the lesser scaup will nest on islands, the latter seems
to show a greater preference for such sites (Vermeer
et al. 1972; Vermeer 1970).

Vegetation.

Structure. The canvasback, redhead, and ruddy
duck, all of which commonly nest over the water,
are highly selective to nesting ponds that provide
emergent vegetation and often band around the
shoreline of small ponds. Redheads tend to avoid
large stands of emergents without breaks of open
water (Low 1945). Ring-necked ducks and scaup all
prefer to nest close to water and consequently use
a variety of structural cover over water (ring-necks)
or near the water (scaup).

Canvasback broods mostly used permanent type
wetlands, larger and deeper wetlands (0.4 ha [1 a.]),

and wetlands with less than one-third of the water
area covered by emergent vegetation (Stoudt 1982).
Broods of redheads, ruddy ducks ( Evans and Black
1956), and lesser scaup (Sugden 1973) showed a
propensity for feeding in the deeper, more open
water segments of the pond. Although amphipods, a
favored food of young scaup, appeared to be more
numerous in larger and deeper ponds, Sugden
(1973) believed that the movement to larger ponds
by older scaup was related to security and not food.

Species Composition. A variety of species of
emergent vegetation have been noted as nesting
cover for canvasbacks, redheads, and ruddy ducks.
Cattails, bulrush, whitetop (Fluminia festucacea),
willows (Salix sp), sedges, and reeds (Phragmites
communis) were listed by Stoudt (1982) in Mani-
toba. Bellrose ( 1978) reported that for nesting
cover, redheads appeared to have a preference for
bulrushes, cattails, and sedges, in that order. He also
stated that ruddy ducks seemed to select the vegeta-
tion characteristic of the nesting pond, availability
playing a key role. I have observed this on stock
dams in eastern Montana, when in good moisture
years, ruddy ducks will readily nest in a relatively
short emergent, longspike spikerush, the only emer-
gent cover present.

Most food habits studies of divers point out the
heavy intake of animal matter by young ducklings
and an increase in plant food as the birds get older
(Bartonek and Hickey 1969; Sugden 1973)- Although
specific foods will vary considerably with location
and availability within this species group, ruddy
ducks appear to feed most heavily on plant materials.
Both species of scaup are most inclined to an animal
diet.

Plants frequently used are pondweeds, musk-
grass (Chara sp. ), bulrush, wild celery (Valisneria
americana), duckweeds (Lemna sp. ), water milfoil,
and widgeon grass. Depending on the species of
plant and season of year, leafy parts, seeds, and tub-
ers of the various plants may be fed upon.

Eiders, Scoters, Mergansers, and Allies
(Mergini)

This relatively diverse tribe is represented in
North America by seven genera, all different in ap-
pearance as adults, but grouped together at least
partially as a result of similarity in voice and display
(Delacour and Mayr 1945). Members of this entire
group are referred to as "sea ducks," primarily be-
cause most of them winter along the sea coasts.
However, several members breed inland on fresh
water, and a few winter in the interior.

Included in the genus Somateria are three ei-
ders, the common eider (S. mollissima), king eider
(S. spectabilis), and spectacled eider (S. fischeri), all

376

Waterfowl

of which breed along the arctic coasts. The common
and king eiders winter along the northern coasts;
the spectacled eider's wintering areas remain some-
what obscure. Steller's eider (Polysticta stelleri), the
sole member of this genus, nests along the northern
coasts of Alaska and Siberia and winters along the
Alaska peninsula and the Aleutians (Bellrose 1978).

The harlequin duck (Histrionicus histrionicus)
has a distinct eastern and western range. It breeds
from the northern coasts of Canada and Alaska south-
ward into the Cascades and Sierra Nevadas and the
Rocky Mountains (Bellrose 1978), and winters along
both coasts as far south as California and Long Island.

Harlequin duck.

Oldsquaws (Clangula hyemalis) are found
throughout arctic Canada and Alaska during the
breeding season; they winter along both coasts, com-
monly south to northern California and northern
North Carolina and interior on the Great Lakes.

Three scoters are included under the genus
Melanitta: the black scoter (M. nigra americana),
surf scoter (M. perspicillata), and white-winged
scoter (M. fusca deglandi). The black scoter breeds
in Alaska and probably in Canada (Bellrose 1978).
Breeding surf and white-winged scoters are both
well distributed across Canada in the open and
closed boreal forest. All three species winter along
most of the east and west coasts of the U.S.

The genus Bucephala includes the bufflehead
(B. albeola) and two goldeneyes, the Barrow's (B.
islandica) and common (B. clangula americana).
The bufflehead has a rather extensive breeding range
throughout Canada; the largest numbers are in north-
western North America (Bellrose 1978). A few iso-
lated breeding populations are found in mountainous
areas of the U.S. and in some timbered areas of the
northern prairies. The primary breeding populations
of Barrow's goldeneyes are in northwestern North
America extending from the mountains of Wyoming

and northern California, north into Alaska. The com-
mon goldeneye is more widespread, breeding in
forest areas across the breadth of Canada and into
Alaska. All three species of this genus concentrate
along both the Atlantic and Pacific coasts during the
winter, but can also be found in small concentrations
along rivers and lakes in the interior.

Bufflehead.

Male and female common goldeneye.

Three species of mergansers (Mergus) are com-
mon to North America: the hooded (M. cucullatus),
red-breasted (M. serrator), and common (M. mer-
ganser americanus). The hooded merganser nests
throughout most of the eastern half of the U.S., ex-
tending into adjoining provinces in Canada. It also
nests in the northwest U.S. from northern California
to southeast Alaska (Bellrose 1978).

The red-breasted merganser breeds throughout
eastern Canada, across the southern tundra and
northern boreal forest into Alaska. The common
merganser has a somewhat similar distribution more

Waterfowl

377

to the south although with a broad overlap zone
where both may be found. The common merganser
breeds in the northeast and lake States, and extends
southward along mountain ranges in the West.

The hooded merganser winters along the east,
west, and gulf coasts and several southeastern States;
the red-breasted winters largely on the Great Lakes
and along both coasts. The common merganser, by
contrast, winters largely ( 58% ) in the interior
(Bellrose 1978).

Physical Features. Of the three Somateria eiders,
the common is the most marine, occurring primarily
along rocky shores and islands and only rarely on
open fresh water. The king and spectacled eiders,
although they both winter along seacoasts, breed
near fresh-water ponds, lakes, deltas, and tidal inlets.
The Stellar's eider nests along the coast or inland
near grassy ponds or lakes but winters along coasts
in shallow marine habitat.

The harlequin duck usually nests along fast-
flowing coastal or mountain streams and winters
along rocky sea coasts.

Oldsquaws, unlike the eiders, nest along the
coast and inland on lakes and ponds throughout the
tundra. Like most of the other sea ducks, it winters
along coastal areas but also on large inland lakes
(e.g., Great Lakes).

All three scoters nest in fresh-water situations,
the black and the surf showing a preference for
more brushy, taller cover near the water, whereas
the white -winged scoter is more prone to nest in
open prairie or tundra and frequently at considerable
distances from the water. Dense ground cover for
nesting is a prominent characteristic (Bellrose
1978).

Goldeneyes and buffleheads are cavity nesters,
so breeding habitat usually consists of trees sur-
rounding fresh-water lakes, ponds, and streams. All
three species winter along the coasts in bays or estu-
aries as well as lakes and rivers inland.

Hooded and common mergansers are also cavity
nesters. The hooded merganser seems more re-
stricted to tree cavities or wood duck houses than
the common merganser, which frequently uses cavi-
ties in rock cliffs along major western rivers. The
red-breasted merganser nests on the ground, often
under shrubs (Weller et al. 1969).

Wintering habitat for the three mergansers var-
ies: the common winters largely on open lakes and
rivers and only rarely in coastal habitat; the hooded
winters mostly in fresh water, but it is also com-
monly found in estuaries and bays; the red-breasted

Hooded merganser.

merganser winters primarily in coastal areas and
rarely in fresh-water.

Vegetation. Less can be concluded regarding the
vegetational aspect of sea duck habitat than with
most other tribes of waterfowl. Several reasons may
be put forth as a partial explanation.

Nesting by many of the species is carried out in
the tundra or in open or closed boreal forest. Resid-
ual vegetation from previous years' growth is critical
because the short season requires that nesting be
initiated before the appearance of much of the cur-
rent season's growth.

The importance of structural cover for nests
rather than species composition is reflected in the
selection of shrubs, rocks, burrows, and overhanging
banks by many species. Several species (eiders, olds-
quaw, some scoters) nest on islands in rather large
numbers (colonial) and as such rely less on cover
and more on the mammal-free island habitat.

Also, several species (goldeneyes, bufflehead,
common and hooded mergansers) are cavity nesters,
and as such the dominating feature is the proper-
sized cavity and not the plant species providing the
cavity. However, certain trees appear more prone to
natural cavities and species composition may thus
influence selection of a particular habitat. Erskine
(1972) in British Columbia noted that of 1 1 trees
that provided cavities for nesting buffleheads, 52%
were aspen (Populus tremuloides).

The relative minor importance of vegetation
structure and composition to sea ducks is also indi-
cated by the little se made of plants in their food
habits. Results from several studies showed that ani-
mal matter in the diet ranged from about 70 to 95%
for eiders, harlequins, oldsquaws, scoters, buffle-
heads, and goldeneyes, and 100% (largely fishes) for
mergansers (Bellrose 1978). Many of the insect lar-

378

Waterfowl

vae taken as food by young ducks of several tribes or
adult dabbling ducks are in close association with
aquatic vegetation. However, many of the animal
foods eaten by sea ducks are crustaceans and mol-
lusks taken during bottom feeding and have rela-
tively less relationship to aquatic macrophytes.

CHARACTERISTICS OF WINTER HABITAT

Winter habitat for waterfowl has only recently
received the attention it deserves. For 6 to 8 months
of the year, this habitat must provide security and a
food source capable of sustaining the bird. For many
species it must also provide an energy base for pair-
ing and initiation of migration. With the wide range
of species requirements present within the water-
fowl for breeding habitat, differences would also be
expected in winter habitat requirements.

Probably the least variability in winter habitat is
found among the tribes that are most aquatic.
Groups within the tribe Mergini, which feed largely
on aquatic invertebrates, are confined to coastal
waters, or if wintering inland, are usually found on
larger lakes or rivers. Heitmeyer and Vohs (1984), in
looking at wintering waterfowl habitat in Oklahoma,
found common mergansers and common goldeneyes
more abundant on the large reservoirs.

The diving ducks (Aythyini), another highly
aquatic group, tend to concentrate along certain
coastal areas. In contrast with many of the Mergini
which winter in marine habitat, the diving ducks
seek out estuarine habitat. In 1955, slightly less than
one-half of the canvasbacks recorded in North Amer-
ica were found in Chesapeake Bay (Perry et al.
1981 ), whereas about 78% of the redheads in exis-
tence normally winter on Laguna Madre along the
Texas coast (Weller 1964).

Both species of scaup make heavy use of coastal
areas, the greater, more abundant in the Atlantic
Flyway and the lesser, more prominent in coastal
marshes of the Mississippi Flyway. Both scaup also
winter inland as far north as the Great Lakes
(Bellrose 1978).

Dabbling ducks show considerable variability in
selecting winter habitat, both between and within
species. Mallards and pintails show different feeding
preferences depending on which portion of their
respective ranges they may be occupying (Chabreck
1979). Fredrickson and Drobney (1979) pointed
out the importance of a food resource on wintering
areas and referred to specific feeding habits that
permit several species of ducks to fully and at times
simultaneously exploit an aquatic food resource.

Thus, with high densities of ducks so often
found in wintering areas, a highly productive aquatic

food source is a key component of winter habitat.
Gadwalls and widgeon in Oklahoma wintered pri-
marily on wetlands where submergent and emergent
vegetation were abundant (Heitmeyer and Vohs
1984). All species of dabbling ducks preferred natu-
ral wetlands over farm ponds, probably a reflection
of the absence of biologically productive littoral
zones on the steep-sided farm ponds. In determining
values for 13 physical, limnological, and vegetative
features on a series of flood-prevention lakes in
Texas, Hobaugh and Teer ( 1981 ) concluded that the
most important characteristics influencing winter
waterfowl use were amounts of aquatic vegetation
and lake surface area.

Some species of waterfowl have been able to
capitalize on land-use changes by modifying their
food habits to include certain upland agricultural
crops. Historically, upland feeding in cultivated fields
in Manitoba was first recorded for geese, which
could be expected in light of the terrestrial habits of
this group. Use of cultivated fields by ducks seem-
ingly developed later and initially appeared with
mallards and pintails (Bossenmaier and Marshall
1958). Field feeding, although initially considered a
fall activity, has to be a major contributor in the
northward extension of wintering ducks (primarily
mallards) into Montana and the two Dakotas. Many
of these birds winter on large bodies of water that
are totally devoid of available aquatic vegetation,
thus forcing a total subsistence on cultivated grains.
On wintering areas farther south, American widgeons
(Anas americana), and blue-winged and green-
winged teal (A crecca) join the mallard and north-
ern pintail as regular field feeders (Baldassarre and
Bolen 1984).

POPULATION MEASUREMENT TECHNIQUES

Virtually all techniques employed for population
measurements of waterfowl are based on direct ob-
servations of the birds. This contrasts with tech-
niques, such as auditory counts, employed in
monitoring upland bird populations. The obvious
reason for this difference is that waterfowl are
closely associated with water for much of the time
and although they are no more equally visible be-
tween different bodies of water and between sea-
sons, they are easier to observe and count than many
other species of birds. Because the methods are
based on direct observation, they all detect presence
with varying degrees of precision and accuracy.

On a continental scale, breeding and winter
populations are surveyed annually by the U.S. Fish
and Wildlife Service, Canadian Wildlife Service, and
State and Provincial conservation agencies. Most
of such surveys are conducted from the air and are
used for determining long-term trends (Bellrose
1978:17-19).

Waterfowl

379

Breeding Population

Description. Annual trend data on North American
populations of breeding waterfowl are gathered in
May and June along aerial east-west transects
(Bellrose 1978:18). The transects, which are
censused cooperatively by State, Federal, and
Canadian biologists, are distributed throughout
prime breeding areas from the prairies to the boreal
forest and into the tundra. Because visibility varies
between species and years, segments of transects are
covered intensively by a ground crew within 24
hours of the air count and a visibility index applied
to the aerial count. This annual survey, which also
provides data on habitat (water) conditions, is
conducted over about 80,500 km ( 50,000 linear mi. )
representing about 2% of the breeding habitat
(Bellrose 1978).

This basic technique, involving ground counts
only, has been modified and refined making it more
useable on a smaller scale. The count, called a breed-
ing pair count, may be conducted on specific bodies
of water, along roadside transects, or on all water
areas within a predescribed area. The counts are
based on the knowledge that breeding pairs of dab-
blers, once established on a breeding home range,
will localize their activities to one or two ponds.
Thus, from the time of pair bond formation (often
on arrival because many species form pair bonds on
the wintering ground or enroute north) until the
female is well into incubation, the locations of pairs
or lone drakes are plotted during four or five cen-
suses. Pairs or lone drakes of any dabbler species
occurring on a given unit of habitat three or four
times are recorded as part of the assigned breeding
population (Dzubin 1969). The counts should be
made when the highest percentage of pairs are in
the pre-nesting, laying, or incubation stage. This may
include a period of 20 to 30 days to include early-
nesting pintails and mallards, as well as late-nesting,
blue-winged teal and gadwalls. Although this tech-
nique is often described to include all ducks, Dzubin
(1969) believed the large home ranges of canvas-
backs and redheads and the tendency for pairs of
these two species and lesser scaup to congregate on
deep ponds termed "primary waiting areas," pre-
cluded enumerating divers in smaller census units.

For additional precision, Dzubin (1969) pro-
posed an "Indicated Breeding Population" where, un-
til a certain date, grouped males and males in aerial
flights temporarily on censused ponds were also
considered pairs. After that date, grouped males of
five or more were not counted as pairs (six or more
were considered post-breeding males). Different
cutoff dates were established for early nesters (pin-
tails and mallards), intermediates (widgeon and
shovelers), and late nesters (blue-winged teal and
gadwalls); dates may vary between areas and years

but are based on nesting chronology of the respec-
tive groups.

Lastly, a sex-ratio corrected population figure
was suggested. This consisted of taking the mean-in-
dicated breeding population from four or five census
efforts and applying a sex-ratio correction factor
based upon sex-ratio data obtained from each spe-
cies before egg laying.

The following are some of the recommendations
by Dzubin (1969) for breeding-pair census on a
grassland type. This type, which has little visual ob-
struction from emergent vegetation, is appropriate
for much of the waterfowl habitat on public lands in
the West.

( 1 ) Census during that portion of the breeding
season when site attachment by pairs and
drakes is greatest (pre-nesting, laying, and early
incubation).

(2) Census between 0800 and 1200 hours; this is
the period of least mobility, and most pairs and
lone drakes will be on waiting stations.

(3) Census only on bright days with temperatures
above 4° C (40° F) and winds less than 24 km/
h (15 mph). Avoid inclement weather (rain,
heavy overcast, and low temperatures) which
affects mobility and visibility of ducks (Diem
and Lu I960).

(4) Conduct at least two censuses when sampling
a multiple species population; four to six
counts would be preferable, permitting a calcu-
lated average number of pairs for each species.

( 5 ) Take counts from a vehicle, positioned at a
vantage point at an adequate distance to pre-
vent flushing birds.

(6) Tally all lone pairs and lone drakes greater
than 5 m ( 1 5 ft ) apart as pairs; this type enu-
meration, if conducted four to six times on the
census area with proper timing, will provide
relative abundance data between areas or be-
tween years on a given area. Greater precision
can be obtained, if the objectives warrant, by
a more refined timetable as recommended
by Dzubin (1969:220).

Breeding populations of Canada geese are com-
monly censused along rivers either from the ground
(boat) or air at which time pairs and lone geese
(ganders) are recorded as pairs and groups as non-
breeders (Hanson and Eberhardt 1971; Allen et al.
1978). Unlike most ducks, geese remain paired
throughout the nesting season so lone ganders will
be available for tally (as pairs) throughout the incu-
bation period.

380

Waterfowl

Prime breeding areas on the prairies are censused in May
and June along aerial transects.

Accuracy and Precision. Sauder et al. (1971), in
evaluating the roadside census for breeding
waterfowl in South Dakota, found that to be within
20% of the mean at the 90% confidence level on an
86.4 km (54 mi.) route, at least four counts of lone
drakes and pairs were needed for blue-winged teal
and three for gadwalls and mallards. They also stated
that additional counts were necessary to maintain
the same degree of accuracy when routes were
shorter or when water areas were less numerous.
Although Dzubin (1969) did not recommend lone
drake and pair counts for divers, Sugden and Butler
(1980), after an evaluation, concluded that lone
male and pair counts for canvasbacks and total
female counts for redheads provided the best index
to breeding populations of these species.

In spite of the obvious shortcomings with the
breeding pair counts, indications are good that it will
provide useable data for certain management prob-
lems. Pair and lone drake counts were conducted on
33 stock ponds during a 5-year study to measure
waterfowl responses to a changing land use (Gjers-
ing 1975; Mundinger 1976). Although the changes in
waterfowl numbers were fairly pronounced over
the period, production data obtained from intensive
brood counts for the same period followed a similar
trend. For many evaluations of waterfowl habitat
on western ranges, carefully timed and conducted
breeding pair counts (lone pairs and drakes) should
fulfill the objectives.

Discussion. Without knowing what percentage of
the actual breeding population is being counted,
counts of breeding pairs at best provide yearly
trends. Timing of the count is critical. The
chronology of the nesting effort of a single species
may extend over a 70-day period (Humburg et al.
1978) and longer when several species are
considered. Thus for intensive work, several counts
are necessary, well-dispersed over the entire
breeding season.

Nest Counts

Nest counts are often incorporated in (or substi-
tuted for) breeding pair counts. Dzubin (1969:2217)
suggested this as a better method on block-type
studies for divers and ruddy ducks, although he
pointed out potential problems in uneven nest distri-
bution, failure to locate all nests, and difficulty in
separating first and second nest attempts.

Sugden and Butler ( 1980) considered nest
searching as an impractical technique for recording
canvasback and redhead breeding densities, primarily
because the variability in nest distribution made it
necessary to have a large sample. Nest counts are
effective for enumerating breeding populations of
Canada geese, particularly when nesting is largely ac-
complished on islands in rivers (Hansen and Eber-
hardt 1971) or ponds (Childress and Eng 1979).

A measure of various upland habitats may be
made by comparing nest densities or nest success.
Nests are located by systematically covering the area
in question at specified time intervals to locate early
and late nesting birds. This effort is frequently facili-
tated by the cable-chain drag method (Higgins et
al. 1969).

With early, intermediate, and late nesting spe-
cies plus renesting by most species, a continued
initiation of nests will be occurring throughout the
nesting season. It is not practical (nor advisable)
to search an area daily; thus many nests may be initi-
ated and destroyed between search periods and con-
sequently not found. Also, nests located early
(during laying or early incubation) suffer a greater
chance of destruction before hatching than a nest
first located shortly before hatching. If all nests are
considered equal regardless of their stage when first
found, the nest density would be biased downward
(e.g., nests destroyed early are not found) and the
hatch rate (nest success) biased upward. The recom-
mended method for standardizing waterfowl nest
density and success data is the Mayfield method
(Mayfield 1975; Miller and Johnson 1978) in which a
hatch rate is determined and applied to each nest
depending on its stage of development toward
hatching.

Brood Counts

Production surveys are conducted by U.S. Fish
and Wildlife Service biologists to assess annual wa-
terfowl breeding success. Brood counts are con-
ducted from the air along many of the same
transects as the breeding bird counts and are subject
to many of the same problems. However, Pospahala
et al. (1974) supported the general contention that
the number of Class II and III broods observed along
aerial transects in July, when adjusted for visibility,

Waterfowl

381

Nesting coot.

is related to the density of breeding birds observed
in the May survey.

Although broods are less mobile than flying
adults, they are usually more secretive. By the time
most broods are on the water, the emergent vegeta-
tion is more lush — reducing visibility. A serious need
for aerial-ground indexes as adjustments for the va-
garies of aerial counts alone was stressed by Diem
and Lu (I960).

Brood counts conducted more intensively from
the ground appear more acceptable. Gollop and Mar-
shall (1954) presented a method for brood counts
that basically calls for several counts to be made and
broods recorded by species and age, which tends
to eliminate duplication of counts. This method has
been employed extensively on studies where more
precise production data are important (Berg 1956;
Mundinger 1976; Kirby 1980).

Stock ponds in Montana generally have far less
emergent vegetation than prairie potholes. Mundin-
ger (1976), in making brood counts, would usually
approach the pond unobtrusively in an attempt to
count undisturbed broods. Even so, brood move-
ment ("transient broods") made it necessary to re-
cord broods by age class to prevent duplication.

Kirby ( 1980) analyzed intensive observations of
breeding pairs and broods of ducks in forested habi-
tats of northern Minnesota. Aerial censuses in such
areas appeared inadequate because of the difficulty
in observing birds at low densities. Kirby (1980)
found that indexes to breeding birds and brood
numbers were highly correlated, and that censusing
was a distinct possibility as a source for population
trend data in this type habitat.

Winter Surveys (Counts)

The annual January inventory was initiated in
the U.S. in 1935 and has increased in scope until it
now encompasses most wintering areas in Alaska,
Canada, the conterminous U.S., and Mexico. The
survey is coordinated by the U.S. Fish and Wildlife
Service, and although the purpose is to determine
the size and distribution of the midwinter waterfowl
population, the potential for error (at least in deter-
mining total population) is obvious and great. Black
ducks and wood ducks winter in small groups more
easily counted, but winter in habitat where they
would be more easily overlooked. Canvasbacks and
redheads often concentrate in open areas where they
can be easily seen, but in such large flocks that esti-
mates of numbers are often subject to error. Popula-
tion status of wintering canvasbacks on Chesapeake
Bay and coastal North Carolina were efficiently and
accurately obtained with low-level aerial photogra-
phy (Haramis et al. 1985).

A detailed description of the winter survey was
presented by Stewart et al. (1958:366-367). Bellrose
(1978:20) acknowledged that winter counts, for
several reasons, were undoubtedly underestimated.
However, he defended their use because they pro-
vided the distribution of ducks at that time of the
year and provided the major annual population data
for geese. Most geese nest in such widely distributed
parts of the arctic and subarctic that annual counts
are impractical.

DISCUSSION

Tiner (1984) provided an excellent summary of
the current status and recent trends of wetlands in
the U.S. Although some recent modest gains were
recorded as a result of man-made lakes and reser-
voirs, a net loss of 3-6 million ha (9 million a.) oc-
curred in a 20-year period starting in the mid-1950s.
Frayer et al. (1983) attributed 87% of these losses
to agricultural activities, 8% to urban development,
and 5% to other development. The geographical
distribution is no more evenly distributed, but
clearly related as to the causes. In the fertile farm-
lands of Iowa, 99% of the natural marshes have been
drained compared with 32% of Wisconsin's wetlands
(Tiner 1984, p. 34).

The major waterfowl habitats within the U.S. are
important primarily as breeding habitat, wintering
habitat, or both (Figure 1 ). The prairie pothole re-
gion, for example, encompasses the most valuable
waterfowl production marshes in North America.
Representing no more than 10% of the continent's
waterfowl breeding area, in a wet year it produces
over 50% of the ducks. Included are many of the
popular species including the mallard, pintail, teal,
widgeon, canvasback, and redhead. Of more recent

382

Waterfowl

AREA NAME

1 — Prairie potholes and parklands

2 — Central Valley of California

3 — Yukon Kiiskokivin Delta

4 — Middle Upper Atlantic Coast

5 — Lower Mississippi River Delta and Red River Basin

6 — Izembek Lagoon

7 — Upper Mississippi River and Northern Lakes

8 — Northern Great Plains

9 — Yukon Flats
10 — Intermountain West (Great Basin)
11 — Teshelpuk Lake
12— Middle Upper Pacific Coast
13 — Klamath Basin
14 — Upper Alaska Peninsula
15 — Copper River Delta
1(>—West Central Gulf Coast
17 — Upper Cook Inlet
18 — San Francisco Bay
19— NE United States
20 — Sandhills and Rainwater Basin
21 — Playa Lakes

Figure 1. Waterfowl habitats of major concern (after Tiner 1984).

Waterfowl

383

and justified concern has been the importance of
wintering habitat (Fredrickson and Drobney 1979)
and its rate of loss reported along coastal marshes
(Tiner 1984:35). Although wetland losses are not
confined to the areas illustrated in Figure 1, they
represent areas of high use, and consequently any
effort toward their preservation or enhancement
would pay great dividends toward waterfowl welfare.

One of the annual measures of potential water-
fowl productivity has been an index to the number
of water areas along the aerial transects used for
breeding birds. It has been assumed that good water
conditions in a given year would speak well for the
productivity of waterfowl in the area. However, with
ever-intensifying agricultural efforts in the surround-
ing uplands, at least the dabbling ducks, which rely
on secure upland nesting cover, may not be able
to respond favorably with good water conditions
only. Higgins (1977), in measuring waterfowl pro-
duction in an area of intense farming, believed that
upland nesting ducks could not reproduce them-
selves during most years in areas that were 85% or
more tilled. He recorded a 21% nest success on
his study area, with 76% of the failures attributed to
predators and 16% to farming operations.

In the northern Great Plains, grazing is a more
prevalent land use, and where grazing is not exces-

sive, waterfowl nest success is often high (77%,
Smith 1953; 67%, Rundquist 1973). Of the two ma-
jor ingredients to successful dabbling duck produc-
tion, quality water areas and secure upland nesting
habitat, the former is currently at a premium in the
northern Great Plains — the latter in the prairie pot-
hole area. Thus the two areas, although adjacent
to one another, offer different options in terms of
type of investment and return for effort expended
toward enhancing habitat.

Intensive agriculture, although detrimental on
the breeding habitat, may be providing at least an
acceptable substitute for some food sources on win-
tering areas. Fredrickson and Drobney (1979) sug-
gested that abundant food supplies (domestic grains)
may be holding some populations farther north than
previously, thereby reducing competition on tradi-
tional (but negatively altered) winter habitat. This is
particularly true for the mallard, which in the Co-
lumbia Basin in Washington, increased from a win-
tering population of 86,000 in the entire State in
1947 to 730,000 in I960, largely as a result of addi-
tional food sources from irrigated grain farms (Bell-
rose 1978). Baldassarre and Bolen (1984) suggested
a possible increasing importance of field-feeding by
wintering waterfowl, because of wetland distribution,
and presented management recommendations to-
ward enhancement of this activity.

384

Waterfowl

LITERATURE CITED

ADDY, C.E. 1964. Atlantic Flyway. Pages 167-184 in Lin
duska, J. P. ed. Waterfowl Tomorrow. U.S. Govt. Print.
Off., Washington, DC.

ALLEN, G.T., S.E. FAST, B.J. LANGSTAFF, D.W. TOMRDLE,
and B.L. TROUTMAN. 1 978. Census of Canada geese
on the Palouse River, Washington, during the spring
of 1977. The Murrelet 59:96-100.

AMERICAN ORNITHOLOGISTS UNION. 1983 Checklist
of North American birds, Sixth Edition. Allen Press,
Inc., Lawrence, KS.

BALDASSARRE, GA. and E.G. BOLEN. 1984. Field-feeding
ecology of waterfowl wintering on the southern high
plains of Texas. J. Wildl. Manage. 48:63-71.

BARTONEK, J.C. and J.J. HICKEY. 1969. Food habits of
canvasbacks, redheads and lesser scaup in Manitoba.
Condor 71:280-290.

BELLROSE, F.C. 1978. Ducks, geese and swans of North
America. The Stackpole Co., Harrisburg, PA, and
Wildl. Manage. Inst., Washington, DC. 540pp.

, KL. JOHNSON, and T.U. MEYERS. 1964. Relative

value of natural cavities and nesting houses for wood
ducks. J. Wildl. Manage. 28:661-676.

BERG, P.F. 1956. A study of waterfowl broods in eastern
Montana with special reference to movements and
the relationship of reservoir fencing to production. J.
Wildl. Manage. 20:253-262.

BISHOP, R.A. and R. BARRATT. 1970. Use of artificial nest
baskets by mallards. J. Wildl. Manage. 34:734-738.

BOSSENMAIER, E.F. and W.H. MARSHALL. 1958. Field-
feeding by waterfowl in southeast Manitoba. Wildl.
Monogr. 1. 32pp.

BREWSTER, W.G., J.M. GATES, and L.D. FLAKE. 1976.

Breeding waterfowl populations and their distribution
in South Dakota. J. Wildl. Manage. 40:50-59.

BUE, I.G., H.G. UHLIG, and D.J. SMITH 1964. Stock ponds
and dugouts. Pages 391-398 in Linduska, J.P. ed.,
Waterfowl Tomorrow, U.S. Govt. Print. Off., Washing-
ton, DC. 770pp.

CARLSEN, T.L. 1984. Waterfowl nesting in two ponds of
the Canyon Ferry Wildlife Management Area, Mon-
tana. M.S. Thesis. Montana State Univ., Bozeman.
91pp.

CHABRECK, R.H. 1979. Winter habitat of dabbling

ducks — physical, chemical, and biological aspects.
Pages 133-142 in Bookhout, T.A. ed., Waterfowl and
Wetlands — An Integrated Review. North-central Sect.,
The Wildl. Soc, Madison, WI

CHBLDRESS, DA. and R.L. ENG. 1979. Dust abatement
project with wildlife enhancement on Canyon Ferry
Reservoir, Montana. Pages 282-288 in The Mitigation
Symposium: a National Workshop On Mitigating
Losses of Fish and Wildlife Habitats. Colorado State
Univ., Ft. Collins.

COWARDIN, L.M., G.E. CUMMINGS, and P.B. REED, Jr.
1967. Stump and tree nesting by mallards and black
ducks. J. Wildl. Manage. 31:229-235.

DELACOUR, J. and E. MAYR. 1945. The family Anatidae.
Wilson Bull. 57:3-55.

DIEM, KL. and K.H. LU. I960. Factors influencing water-
fowl censuses in the parklands. Alberta, Canada. J.
Wildl. Manage. 24:113-133-

DUEBBERT, H.F. 1982. Nesting waterfowl on islands in

Lake Audubon, North Dakota. Wildl. Soc. Bull. 10:232-
237.

— and J.T. LOKEMOEN. 1976. Duck nesting in fields
of undisturbed grass-legume cover. J. Wildl. Manage.
40:39-49.
J.T. LOKEMOEN, and D.E. SHARP. 1983- Concen-

trated nesting of mallards and gadwalls on Miller Lake
Island, North Dakota. J. Wildl. Manage. 47:729-740.

DZUBIN, A. 1969. Assessing breeding populations of ducks
by ground counts. Pages 178-230 in Saskatoon Wet-
lands Seminar. Can. Wildl. Serv. Rep. Series 6.

ENG, R.L., J.D. JONES, and F.M. GJERSING 1979. Con

struction and management of stock ponds for water-
fowl. U.S. Dep. Inter., Bur. Land Manage. Tech. Note
327. Denver, CO. 39pp.

ERSKINE, A.J. 1972. Buffleheads. Can. Wildl. Serv. Monogr.
Series 4. 240pp.

EVANS, CD. and KE. BLACK 1956. Duck production
studies on the prairie potholes of South Dakota. U.S.
Dep. Inter., Fish and Wildl. Serv. Spec. Sci. Rep. Wildl.
32. 59pp.

EWASCHUK, E. and DA. BOAG. 1972. Factors affecting
hatching success of densely nesting Canada geese. J.
Wildl. Manage. 36:1097-1106.

FRAYER, W.E., T.J. MONAHAN, DC. BOWDEN, and FA.
GRAYBILL. 1983- Status and trends of wetlands and
deep water habitats in the conterminous United
States, 1950s to 1970s. Dep. of Forest and Wood Sci-
ences. Colorado State Univ., Ft. Collins. 32pp.

FREDRICKSON, L.H. and R.D. DROBNEY. 1979. Habitat
utilization by post-breeding waterfowl. Pages 119-131
in Bookhout, T.A. ed., Waterfowl and Wetlands — An
Integrated Review. North-central Sect., The Wildl.
Soc, Madison, WI.

GIROUX, J.F. 1981. Use of artificial islands by nesting
waterfowl in southeastern Alberta. J. Wildl. Manage.
45:669-679.

, D.E. JELINSKI, and R.W. BOYCHUK 1983- Use of

rock lands and round, straw bales by nesting Canada
geese. Wildl. Soc. Bull. 11:172-177.

GJERSING, F.M. 1975. Waterfowl production in relation to
rest-rotation grazing. J. Range Manage. 28:37-42.

GOLLOP, J.G. and W.H. MARSHALL. 1954. A guide for
aging duck broods in the field. Mississippi Flyway
Council Tech. Sect. 14pp. (mimeo.).

HAMMOND, M.C. and G.E. MANN. 1956. Waterfowl nest-
ing islands. J. Wildl. Manage. 20:345-352.

HANSEN, H.A., P.E.K SHEPHERD, J.G KING, and W.A.

TROYER. 1971. The trumpeter swan in Alaska. Wildl.
Monogr. 26. 83pp

HANSON, W.C. and L.L. EBERHARDT. 1971. A Columbia
River Canada goose population, 1950-1970. Wildl.
Monogr. 28. 6 lpp.

HARAMIS, G.M., JR. GOLDSBERRY, D.G. MCAULEY, and
E.L. DERLETH. 1985. An aerial photographic census of
Chesapeake Bay and North Carolina canvasbacks. J.
Wildl. Manage. 491:449-454.

HEITMEYER, ME. and PA. VOHS, Jr. 1984. Distribution
and habitat use of waterfowl wintering in Oklahoma.
J. Wildl. Manage. 48:51-62.

HIGGINS, KF., L.M. KIRSCH, and I.J. BALL. 1969. A cable-
chain device for locating duck nests. J. Wildl. Manage.
33(4):1009-1011.

. 1977. Duck nesting in intensively farmed areas of

North Dakota. J. Wildl. Manage. 41:232-242.

HILDEBRAND, B.D. 1979. Habitat requirements of molt-
ing Canada geese at Lima Reservoir, Montana. M.S.
Thesis. Montana State Univ., Bozeman. 79pp.

Waterfowl

385

HINES,J.E. and G.J. MITCHELL. 1983. Gadwall nest-site
selection and nesting success. J. Wildl. Manage.
47:1063-1071.

HOBAUGH, W.C and J.G. TEER. 1981. Waterfowl use

characteristics of flood-prevention lakes in north-cen-
tral Texas. J. Wildl. Manage. 45:16-26.

HUDSON, M.S. 1983. Waterfowl production in three age
classes of stock ponds in Montana. J. Wildl. Manage.
47.112-117.

HUMBURG, D.D., N.H. PRINCE, and R.A. BISHOP. 1978.
The social organization of a mallard population in
northern Iowa. J. Wildl. Manage. 42:72-80.

KAMINSKI, R.M. and H.H. PRINCE. 1977. Nesting habitat
of Canada geese in southeastern Michigan. Wilson
Bull. 89:523-531.

KEITH, L.B. 1 96 1 . A study of waterfowl ecology on small
impoundments in southeastern Alberta. Wildl. Mon-
ogr. 6. 88pp.

KIRBY, RE. 1980. Waterfowl production estimates on
forested wetlands from pair and brood counts. Wildl.
Soc. Bull. 8:273-278.

KROHN, W.B. and E.G. BIZEAU. 1979. Molt migration of
the Rocky Mountain population of the western Can-
ada goose. Pages 1 30- 1 40 in Management and Biology
of Pacific Flyway Geese Symp. N.W. Sect. Wildl. Soc.
Portland, OR.

LEMIEUX, L. 1959. The breeding biology of the greater
snow goose on Bylot Island, Northwest Territories.
Can. Field-Nat. 73:117-128.

LOKEMOEN, J.T., H.F. DUEBBERT, and D.E. SHARP. 1984.
Nest spacing, habitat selection, and behavior of water-
fowl on Miller Lake Island, North Dakota. J. Wildl.
Manage. 48:309-321.

LOW, J.B. 1945. Ecology and management of the redhead
(Nyroca americana) in Iowa. Ecol. Monogr. 15:35-69.

MACINNES, CD. 1962. Nesting of small Canada geese
near Eskimo Point, Northwest Territories. J. Wildl.
Manage. 26:247-256.

MAJ, ME. 1983- Analysis of trumpeter swan habitat on the
Targhee National Forest of Idaho and Wyoming. M.S.
Thesis. Montana State Univ., Bozeman. 102pp.

MAYFIELD, H. 1975. Suggestions for calculating nest suc-
cess. Wilson Bull. 87:456-466.

MENDALL, H.L. 1958. The ring-necked duck in the North-
east. Univ. of Maine Bull. 60(16). 317pp.

MILLER, H.W. and D.H. JOHNSON. 1978. Interpreting the
results of nesting studies. J. Wildl. Manage. 42:471-
476.

MUNDINGER, J.G 1976. Waterfowl response to rest-
rotation grazing. J. Wildl. Manage. 40:60-68.

NAGEL, J. 1965. Field-feeding of whistling swans in north-
ern Utah. Condor 67:446-447.

PATTERSON, J.N. 1976. The role of environmental hetero-
geneity in the regulation of duck populations. J. Wildl.
Manage. 40:22-32.

PERRY, M.C., RE. MUNRO, and G.M. HARAMIS. 1981.
Twenty-five-year trends in Chesapeake Bay diving
duck populations. Trans. North Am. Wildl. Nat. Re-
sour. Conf. 46:299-310.

POSPAHALA, R.S., DR. ANDERSON, and C.J. HENNY.

1974. Population ecology of the mallard. II. Breeding
habitat conditions, size of the breeding populations
and production indices. U.S. Dep. Inter., Fish and
Wildl. Serv. Resour. Publ. 115. 73pp.

RUNDQUIST, V.M. 1973. Avian ecology on stock ponds in
two vegetational types in north-central Montana.
PH.D. Thesis, Montana State Univ., Bozeman. 116pp.

RYDER, J. P. 1967. The breeding biology of Ross' goose in
the Perry River region, Northwest Territories. Can.
Wildl. Serv. Rep. Series 3- 56pp.

SAUDER, D.W., T.L LINDER, R.B. DAHLGREN, and W.L.
TUCKER. 1 97 1 . An evaluation of the roadside tech-
nique for censusing breeding waterfowl. J. Wildl.
Manage. 35:538-543.

SHERWOOD, G.A. I960. The whistling swan in the West,
with particular reference to Great Salt Lake Valley,
Utah. Condor 62:370-377.

SMITH, R.H. 1953- A study of waterfowl production in
artificial reservoirs in eastern Montana. J. Wildl. Man-
age. 17:276-291.

STEWART, RE., A.E. GEIS, and CD. EVANS. 1958. Distri-
bution of populations and hunting kill of the canvas-
back J. Wildl. Manage. 22:333-370.

STOUDT, J.H. 1982. Habitat use and productivity of can-
vasbacks in southwestern Manitoba, 1961-72. U.S.
Dep. Inter., Fish and Wildl. Serv. Spec. Sci. Rep. Wildl.
248. 31pp.

SUGDEN, L.G. 1973. Feeding ecology of pintail, gadwall,
American widgeon, and lesser scaup ducklings. Can.
Wildl. Serv. Rep. Series 24. 45pp.

and G BUTLER. 1980. Estimating densities of

breeding canvasbacks and redheads. J. Wildl. Manage.
44:814-821.

TINER, R.W., Jr. 1984. Wetlands of the United States: Cur-
rent status and recent trends. U.S. Dep. Inter., Fish
and Wildl. Serv. Habitat Resources. Newton Corner,
MA. 59pp.

VERMEER, K 1970. Some aspects of the nesting of ducks
on islands in Lake Newell, Alberta. J. Wildl. Manage.
34:126-129-

, D.R.M. HATCH, and J. A. WINDSOR. 1972. Greater

scaup as common breeder on northern Lake Winni-
peg. Can. Field-Nat. 86:168.

WELLER, M.W. 1964. Distribution and migration of the
redhead. J. Wildl. Manage. 28:64-103-

, D.L. TRAUGER, and G.L KRAPU. 1969. Breeding

birds of the West Mirage Islands, Great Slave Lake,
Northwest Territories, Can. Field Nat. 83:344-360.

386

Waterfowl

19

COLONIAL
WATERBIRDS

Steven M. Speich

Cascadia Research Collective
Waterstreet Building, Suite 201
218 1/2 W. 4th Avenue
Olympia, WA 98501

Editor's Note: Like marsh and shorebirds, colonial
waterbirds are a taxonomically diverse group, yet
they require similar management and survey tech-
niques. Because of their colonial nesting behavior,
they can be very sensitive to impacts and distur-
bances to nesting areas, including those caused by
biologists conducting inventories and monitoring
studies. This chapter provides guidelines on survey
techniques and methods to minimize impacts to
the birds from such efforts.

INTRODUCTION

Any bird that predominantly feeds in aquatic
systems, marine, fresh water, or both, and tends to
nest in groups of closely associated nests, is consid-
ered a colonial waterbird (Table 1 ). Some species
nest both in groups and singly, and all colonial nest-
ing birds nest as solitary pairs at times. Discussions
of the biology of colonial waterbirds are found in
Emlen and DeMong (1975), Evans (1982), Krebs
(1974, 1978), Kushlan (1978), and Pratt (1980).
Waterfowl (Anseriformes) are generally not thought
of as colonial nesting waterbirds and therefore, are
not discussed in this chapter.

Species of colonial waterbirds range from small
storm petrels to large pelicans. Some species, such as
storm petrels, lead a truly pelagic existence, except
when they come to offshore rocks and islands to
nest, often in dense colonies. Other colonial nesting
species (alcids and cormorants, brown pelican [Pele-
canus occidentalism and some gulls and terns) are
less pelagic, with coastal and nearshore affinities
throughout the year, or at least portions of the year.
A few species bridge the gap between marine and
fresh-water environments, with at least segments of
the species' population feeding in marine or fresh
water systems. Indeed, some individuals (some gulls
and terns, great blue heron [Ardea herodias], and
double-crested cormorant [Phalacrocorax auritus])
feed in both systems while nesting. And still other
species are truly specific to fresh water while nest-
ing, i.e., grebes; white pelican (P. erythrorhynchos);
and some egrets, gulls, and terns.

There are several reasons why it is important to
survey and census colonial nesting waterbirds. Most
species are near or at the top of food chains and
are thus sensitive to the health of fresh-water and
marine ecosystems. Changes or reductions in food
chain biomass and species composition often cause
stress in colonial waterbirds, expressed as failure
to breed, abandonment of eggs and young, late nest-
ing, depressed growth rates, or reduced fledging
success. Over long periods, populations may decline;
this can be monitored through a censusing program.

Colonial Waterbirds

387

Table 1. Habitat and nesting characteristics of colonial waterbirds found in western North America.

Family and Species

Habitat

Nest Type and Substrate

Nest Location

Grebes (Podicipedidae)

Red-necked (Podiceps

coastal waters, lakes

floating, reeds

surface

grisegena)

Eared (P. nigricollis)

lakes, bays, ocean

floating, reeds

surface

Western (Aechmophorus

lakes, sloughs, bays

floating, reeds

surface

occidentalis)

Fulmars (Procellariidae)

Northern (Fulmarus glacialis)

ocean

herbage, dried grass

seal cliffs

Storm Petrels (Hydrobatidae)

Fork-tailed (Oceanodroma

ocean

burrow

sea islands

furcata)

Leach's (0. leucorhoa)

ocean

burrow

sea islands

Ashy (0. homochroa)

ocean

under rocks

sea islands

Black (0. melania)

ocean

burrow

sea islands

Pelicans (Pelecanidae)

American White

lakes, marshes,

on ground or

surfaces, bare

(Pelecanus erythrorhynchos)

salt bays, beaches

in bulrushes

ground

Brown (P. occidentalis)

ocean, salt bays

on ground

islands

Cormorants (Phalacrocoracidae)

Double-crested (Phalacrocorax

coast, bays,

sticks, brush,

trees, marshes,

auritus)

lakes, rivers

seaweed

islands, or cliffs

Brandt's (P. penicillatus)

ocean, coast, littoral

seaweed, sticks

rocky island, cliffs

Pelagic (P. pelagicus)

coast, bays, sounds

seaweed, grass

ledges, cliffs

Red-faced (P. urile)

ocean, coast

seaweed, grass

rocky islands

Herons and Egrets (Ardeidae)

Great Blue Heron

marshes, swamps,

sticks

trees, rocky islands

(Ardea herodias)

streams, shores,
tideflats, ditches

Great Egret (Casmerodius

marshes, ponds,

sticks, brush

trees, brush over

albus)

shores

water

Snowy Egret (Egretta thula)

marshes, ponds,

sticks, tules,

willow thickets in

tideflats

bushes in marsh

marshes

Cattle Egret (Bubuleus ibis)

cow pastures, fields
and marshes

sticks

in bush or trees

Black-crowned Night Heron

marshes, lake margins,

sticks, stalks

trees, willow

(Nycticorax nycticorax)

shores

thickets in
marshes

Ibises (Threskiornithidae)

Glossy Ibis (Plegadis falcinellus)

marshes, lake margins

sticks, stalks

bushes in marshy
areas

Gulls and Terns (Laridae)

Laughing Gull (Larus atricilla)

salt water marshes,
bays

grass platforms

marsh islands

Franklin's Gull (L. pipixcan)

prairies, marshes, lakes

among reeds

prairie marshes

Bonaparte's Gull

ocean, bays, rivers,

in conifers

wooded muskeg

(L Philadelphia)

lakes

Mew Gull (L. canus)

coasts, rivers, lakes

rocks or trees

marsh or beach

Ring-billed Gull

coast, estuaries, rivers,

dried grass,

islands in lakes

(L delawarensis)

lakes, fields, refuse
dumps

weeds, sticks

California Gull (L. californicus)

coast, beach, lake,
river, farmland, cities

on ground

islands in lakes

Herring Gull (L. argentatus)

ocean coast, bays,

grass or seaweeds

ground or cliff,

beaches, lakes,

islands

farmlands, dumps

388

Colonial Waterbirds

Table 1. Habitat and nesting characteristics of colonial waterbirds found in western North America
(concluded).

Family and Species

Habitat

Nest Type and Substrate

Nest Location

Gulls and Terns (Laridae) (continued)

Thayer's Gull (L thayeri)

ocean coast, bays,
beaches

vegetation on rocks

on ground

Western Gull (L. Occidentalis)

coastal waters,

grass

offshore islets,

estuaries, beaches

seacliffs

Glaucous-winged Gull

ocean, coast, bays,

seaweed or grass

coastal islands,

(L. glaucescens)

beaches, piers,
dumps, waterfronts

headlands

Black-legged Kittiwake

ocean

seaweed

ledges of seacliffs

(Rissa tridactyla)

Red-legged Kittiwake

ocean

seaweed or grass

ledges of seacliffs

(R. brevirostris)

Sabine's Gull (Xema sabini)

ocean, tundra

grass

depressions in low

(summer)

tundra

Gull-billed Tern (Sterna nilotica)

lakes, marshes, fields

surface depression
lined with shells
and straw

sandy islands

Caspian Tern (S. caspia)

lakes, bays, coast

large, deep hollows

sandy islands or

lined with sticks

dikes of lakes

and shells

Elegant Tern (S. elegans)

coast

scrape on sand

islands

Common Tern (S. hirundo)

ocean, bays, beaches,
lakes (summer)

surface depression

lake islands

Arctic Tern (S. paradisaea)

ocean, coast,

surface depression

island, beach,

tundra lakes (sum-

tundra

mer)

Aleutian Tern (S. aleutica)

ocean, coast

depression in moss

islands

Forster's Tern (S. forsteri)

ocean, lakes, bays,

marsh vegetation

marshes or on

marshes, beaches

muskrat houses

Least Tern (S. antillarum)

ocean, bays, beaches,

scrape

sandy beaches or

estuaries

gravel bars

Black Tern (Chlidonias niger)

coasts, lakes, fresh-

marsh vegetation

floating or

water marshes

on muskrat
houses

Murres, Guillemots, Auks, etc. (Alci-

dae)

Common Murre (Uria aalge)

ocean, large bays

bare rock or

bare ledges, cliffs,

loose soil

or rocky islands

Thick-billed Murre (U. lomvia)

ocean

bare ledges

bare rock, sea cliffs

Black Guillemot

ocean, rocky coast

crevices, burrows

sea cliffs

(Cepphus grylle)

Pigeon Guillemot

ocean, rocky coast

crevices, burrows

sea cliffs

(C. columba)

Xantu's Murrelet

ocean

rock crevices

sea islands

(Synthliboramphus hypolecus)

Ancient Murrelet (S. antiquus)

ocean, sounds

burrows or bare rock

sea islands

Cassin's Auklet

ocean

crevices or burrows

sea islands

(Ptychoramphus aleuticus)

Parakeet Auklet

ocean

crevices

sea cliffs

(Cyclorrhynchus psittacula)

Least Auklet (Aethia pusilla)

ocean

under rocks

sea islands

Wiskered Auklet (A. pygmaea)

ocean

crevices

sea cliffs

Crested Auklet (A. cristatella)

ocean, tide-rip

crevices

sea cliffs

Rhinoceros Auklet

ocean, tide-rip

burrows

sea islands

(Cerorhinca monocerata)

Tufted Puffin

ocean

burrows

sea islands or

(Fratercula cirrhata)

headlands

Horned Puffin (F. corniculata)

ocean

burrows or crevices

sea cliffs

Colonial Waterbirds

389

Colonial nesting waterbirds have served, coinci-
dently, as indicators of man-made contaminants in
the environment. The effects have included repro-
ductive failure, which can be detected through care-
ful censusing and surveying programs. Knowledge
of the numbers and locations of colonies provides an
immediate source of biological materials for the geo-
graphical monitoring of the levels and types of pollu-
tants in the environment and their subtle effects on
body and egg tissues.

Information that can be obtained from censusing
and surveying programs, includes quantification of
the numbers and geographic locations of colonies,
numbers and species present in each colony, and
habitats occupied by species and colonies. Collecting
quantified information about colonies for long pe-
riods, decades and preferably longer, allows a mean-
ingful monitoring of populations. There is often a
large variation in reproductive success from year to
year. Long-term population trends cannot be reliably
derived from short-term studies (those lasting 5
years or less).

Determining the numbers of waterbirds is more
efficient when they are concentrated at breeding
sites. During the nonbreeding portion of the year,
the individuals of most species are dispersed over
large geographic areas, such as all or part of North
America, Central and South Americas, the continental
margins, or the Pacific Ocean. Even when numbers
are determined away from breeding colonies in re-
mote areas, the origin of observed birds is almost
always unknown. The efficiency and accuracy obtain-
able from censusing breeding colonies is usually
superior to those from censusing wintering colonies.

The material presented here is intended to aid
investigators in the design and execution of census-
ing, surveying, and monitoring programs. Primarily,
the biology of colonial nesting waterbirds relevant to
such programs, measurement techniques, examples
and comparisons of techniques actually applied, and
limitations on observers of colonial nesting water-
birds relevant to disturbance and its effects are dis-
cussed. These materials are meant to guide potential
users in formulating specific programs, according
to species, nesting sites, and program goals.

COLONIAL WATERBIRD BIOLOGY

Several aspects of the biology of colonial nesting
waterbirds are both unique and important. Since
most of these aspects vary greatly, it is helpful to be
aware of them. Knowledge of these aspects of biol-
ogy will improve the design and conduct of any
inventory or monitoring study. The following aspects
of the biology of colonial nesting waterbirds are
relevant to study design and execution.

Colony Size

In most cases, it is not possible to make a defini-
tive statement of the exact size of a given colony.
The failure of nesting birds, and the occurrence of
renesters and late nesters, makes any determination
imprecise. Thus, any statement of colony size must
be accompanied by qualifying statements, such as
survey date, stage of nesting, and timing of the sur-
vey in relation to the colony nesting cycle. If precise
information on any of these nesting aspects is
known, it should accompany the colony census data.

For each species, there is a range in the number
of active nests found in its colonies. Solitary nesting
pairs occur in most species. The upper colony size
limit varies with each species. If a histogram of col-
ony size were constructed for each species, all
would be different. The frequency distribution of
colony size is, more or less, species specific (Black-
lock et al. 1978; Buckley and Buckley 1972; Des-
Granges and Laporte 1979, 1981; DesGranges et al.
1979; Sowls et al. 1978, 1980; Speich and Wahl
1985; Thompson 1982; Werschkul et al. 1976,
1977).

Colony size varies geographically in some spe-
cies. In marine species, a good example is the com-
mon murre {Uria aalge). In the Pacific Ocean basin,
it breeds along the continental margin from central
California to Alaska. Colonies in California, Oregon,
and Washington usually number hundreds, thou-
sands, or low tens-of-thousands. Colonies of just a
few pairs or tens-of-pairs are present. However, in
Alaska, colonies are known to contain up to
hundreds-of-thousands of birds. A similar pattern is
present in Leach's storm petrel (Oceanodroma leu-
corhoa), and to a lesser extent, tufted puffin (Frater-
cula cirrhata). However, other species do not show
geographic variation in colony size along the same
coast line: double-crested cormorant, pelagic cormo-
rant (Phalacrocorax pelagicus), glaucous-winged
gull (Larus glaucescens), and pigeon guillemot (Cep-
phus columba) (Sowls et al. 1978, 1980; Speich
and Wahl 1985).

Inland fresh-water species also show a range in
colony size, but the upper limit is dramatically lower
than the large colonies of marine birds (English
1978).

Colony Size Stability

The number of nesting individuals of a species
at a site varies over time. To document this, it is
necessary to determine the number of nests or nest-
ing adults at a site with the same methods at the
same time in the nesting cycle over several years,
preferably at least a decade (Erwin et al. 1981 ). In
general, survey data are lacking for most species and
most sites over long periods of time, seldom allow-

390

Colonial Waterbirds

ing analysis of colony site trends. Data are usually
fragmentary and imprecise. There are examples of
species colonizing new sites, ceasing to nest at sites,
and switching sites (Ainley and Lewis 1974; Buckley
and Buckley 1980; Conover and Conover 1981;
DesGranges 1980; Drent and Guiguet 1961; Drent et
al. 1964; Nettleship 1975; Ogden 1978; Penland
1982; Pratt 1972, 1974; Speich and Wahl 1985;
Stowe 1982; Thompson et al. 1979; Tuck 1961).

Nesting

Of critical importance to the censusing and
monitoring of colonial nesting species is the knowl-
edge of the time of nesting and the breeding syn-
chrony of the subject species, as a whole, and its
colonies. The time of breeding and synchrony are
distorted by renesting and late nesting pairs. Al-
though the time of nesting at any site tends to take
place at the same time each year, there is, nonethe-
less, variation in the time of nesting (Bayer and
McMahon 1981, Drent 1965; Ogden 1978; Pratt
1970, 1972, 1974; Thompson and Tabor 1981; Tuck
1961 ). At times this can vary as much as a month
(Drent et al. 1964; Speich and Wahl 1985). A few
studies have documented geographical variation in
the time of nesting. Even colonies close together can
be out of phase (Bailey and Terman 1983; Bayer
and McMahon 1981; Speich and Wahl 1985; Thomp-
son and Tabor 1981; Tuck 1961; Werschkul et al.
1977; pers. obs.).

Renesting. Generally egg laying is clumped and can
be described statistically by calculating mean,
variance, range, etc. However, some nesters fail
through loss of nest, eggs, young, or even mates. The
number of failed nests can range from a small
percentage to, at times, the entire colony.

Often the individuals that failed in their initial
attempt will renest later in the same season. After
nest failure, there is a recycle time before they at-
tempt to renest. The re-entering of these individuals
into the colony prolongs the active nesting phase
of the colony and negates the usual pronounced
breeding synchrony observed at the onset of nesting
in the colony.

Non-nesters. In most populations there is a
segment that does not breed. It may be quite small,
or sizable and measurable, depending on the species.
These non-nesting individuals may not be present
in the nesting colony at all, or may be present for
only a short time. Generally, non-nesters are thought
to be young inexperienced birds, without previous
nesting experience (Rodgers 1978). However,
occasionally yearling birds do attempt to nest (Lowe
1954; Pratt 1973; Rodgers 1978; Speich and
Manuwal 1974).

Late Nesters. Although nesting colonies often
initially appear to be synchronous in nesting, this
frequently is dispelled by the arrival of late nesting
individuals. The numbers can be just a few pairs
or so numerous that they double the colony size, or
more (Rodgers 1978). The appearance of late
nesting birds is often gradual, or in mass (Massey
and Atwood 1981).

Mixed Species Colonies. In mixed species
colonies, nesting species are often out of phase with
one another. This is usually preceded by different
arrival times for each species (Burger 1978; Drent et
al. 1964; McCrimmon 1978a; Thompson 1981;
Thompson and Tabor 1981; pers. obs.).

A colony of adult nesting pelicans.

A mixed-species colony consisting of great blue herons,
double-crested cormorants, and black-crowned night her-
ons.

Colonial Waterbirds

391

Attendance Patterns

When censusing colonial nesting species and
counting adults, it is critical to determine daily col-
ony attendance patterns of the nesting adults be-
cause the proportion of nesting adults in the colony
varies with the time of day, the stage of nesting,
and other environmental patterns (Birkhead 1978;
Brun 1972; Burger 1976; Conover and Miller 1980;
Custer and Osborn 1978; Drent 1965, 1967; Galusha
and Amlaner 1978; Gibson 1950; Lloyd 1972; Manu-
wal 1974b; Patterson 1965; Slater 1980; Tuck 1961).

Distribution and Habitat

When working with colonial nesting waterbirds,
two aspects of distribution should be considered:
( 1 ) the distribution of colony sites over often broad
geographic regions and (2) the distribution of the
nests in a colony.

The geographic distribution of a species' colo-
nies may cover a considerable area, such as large
portions of western North America (Godfrey 1966;
Palmer 1962; Ryder 1978; Sowls et al. 1978, 1980;
Speich and Wahl 1985; general bird guides). The
distribution of colony sites within overall ranges may
evidence one or more distribution patterns. Colonies
may be distributed non-randomly, with clusters near
lakes, streams, or coastlines (Drent and Guiguet
1961; English 1978; Sowls et al. 1978, 1980; Speich
and Wahl 1985). Within an area, such as a corridor
along a river or around a lake, the nests there may
be evenly or randomly spaced. Knowledge of the
study species' distribution pattern is important to
any plan to locate colony sites for all programs.

The geographic distribution of a species is par-
tially determined by the interactions of its ability

to reach new areas, the extent it is a specialist or
generalist in feeding and nesting requirements, and
the extent of its preferred habitat(s) (Udvardy
1969).

Colony Site Selection

The factors involved in colony site selection are
only partially understood and undoubtedly vary with
each species. They may include ( 1 ) suitable nest
site substrate or support for nest placement, (2)
shelter from adverse environmental conditions, (3)
apparent freedom from predators, and (4) proximity
to food sources (English 1978; Buckley and Buckley
1972). Individuals of colonial nesting species usually
nest in colonies with conspeciflcs and often with
other species as well, while suitable nearby sites are
often left unused with all individuals in the area nest-
ing at one or a few sites (Weise 1978). Behavioral
factors may be as important as those of selecting
nest sites (Buckley and Buckley 1980).

Nest Site Selection

Nest density varies with each species. Inter-nest
(how far apart nests are constructed) distance is
the result of nest site selection and appears to be
affected by nest site structure as it relates to habitat
suitability and social interactions (Buckley and Buck-
ley 1977; Burger 1978; McCrimmon 1978a).

Habitat Suitability. The placement of a nest
ultimately depends on whether the substrate is
appropriate for a particular species' nests. The nests
of some species are found in several habitat types;
the habitat structure for nest placement is apparently
important (Burger 1978). Great blue heron nests
are found on the ground, cliff ledges, in bushes and
shrubs, deciduous trees, coniferous trees, dead trees,
and man-made structures (DesGranges 1979; Henny
and Kurtz 1978; Werschkul et al. 1976). In this

SM

Typical white pelican colony.

Great blue heron nests in cottonwood grove.

392

Colonial Waterbirds

species, if the substrate will support the nest, it can
be used for nesting. Finding suitable places that
support a nest in a colony limit in part nest
placement, nest density, and colony size. For most
species that nest in vegetation, the species of plants
used are not necessarily important, as long as the
structures of the plants meet the support and shelter
requirements of the species' nest (Bailey and Terman
1983; Bayer and McMahon 1981; Werschkul et al.
1976).

Burrow nesting species require a substrate that
allows the digging of a burrow, yet will not collapse
once the burrow is complete. In rocky soil, density
may be limited by a scarcity of "clean" soil that al-
lows burrowing. Talus slopes contain only a limited
number of openings that lead deep enough into the
slope to afford nesting sites for alcids and storm-
petrels (pers. obs.).

Social Interactions. The presence of a species
nesting in a colony may affect the choice of nest
sites of another species (Burger 1978). If all suitable
nest sites are taken, other individuals (conspecincs
or otherwise ) have four choices: ( 1 ) displace the
occupants of a nest site, ( 2 ) wait until a site is open,
(3) nest in a less favorable site, or (4) go elsewhere
(Thompson 1981).

Inter-nest distance is determined in part by so-
cial interaction (territory defense) with conspecincs
(Buckley and Buckley 1977; Meyerriecks I960) or
other species (Burger 1978). The space between
nests is open if a new nest, territory, can "fit" be-
tween existing nests.

The density of burrowing birds, such as Cassin's
auklet, is limited by aggressive interactions on the
colony surface (Manuwal 1974a, b; Thoresen 1964).
Burrow occupants vigorously defend the burrow
opening and small area around the burrow entrance.
Glaucous-winged gulls and western gulls (L. occiden-
talis) nest in higher densities where the substrate
surface is irregular with many large objects on the
ground. These are visual barriers and allow closer
packing of the nests. Herons and egrets nest in lower
densities, allowing for large body size and the reach
of long necks of birds at their nests (Burger 1978).
Pelecaniformes defend nest sites and a small area
around nests (Van Tets 1965), probably resulting
from the observed nest densities in
colonies.

Nest Types, Placement, and Habitat

Colonial nesting waterbirds in western North
America construct a variety of nests and place them
in characteristic locations and habitats depending on
the needs of the particular species.

Floating Platform. Floating nests are generally
constructed in marshes or areas of slow moving
water. The nests are attached to the edge of patches
of anchored, emergent vegetation; areas of variable-
plant density; and often in thick stands of emergent
vegetation. The visibility of nests varies, with edge
nests easily observable, while nests in areas of dense
vegetation are often visible only when the observer
is next to the nest. Nests in dense stands are often
obscured from sight from the air; the nests are built
with vegetation forming a natural overhead shelter.
(For species nest descriptions see Bent 1922;
Godfrey 1966; Palmer 1962.)

Subsurface Chamber. Subsurface nests are used or
constructed by a variety of marine birds. (For
species' nest descriptions see Bent 1919; Palmer
1962; Sowls et al. 1978.) In some cases, nesting
chambers are naturally occurring. The sites range
from talus slopes, rubble piles, boulder piles, sea
caves, driftwood, various holes and cracks in cliffs,
and burrows dug by other species (Manuwal 1974b).
Existing holes are often extended and enlarged, but
at least usually cleaned out. Birds nesting in natural
cavities are often particularly difficult to census and
the numbers present in a colony cannot always be
determined from surface features (Byrd et al. 1983;
Sowls et al. 1978).

In other cases, nesting chambers are con-
structed. For instance, many species of marine birds
excavate burrows for nesting. Chambers are often
excavated new each season or used for several nest-
ing seasons. Burrow sites vary with each species
and show considerable range in some species. Sites
must allow burrowing and be stable enough to per-
mit the structure to survive the nesting season. Bur-
rows vary in length, shape, and depth, depending
on the species and the nesting substrate. Surface
cover is not critical providing the vegetation can be
penetrated.

Glaucous gull.

Colonial Waterbirds

393

Surface Nests. Nests placed on the surface of the
ground range from simple, virtually no nest at all, to
large complex structures. (For species' nest
descriptions see Bent 1919, 1921, 1922, 1927;
Godfrey 1966; Palmer 1962; others.)

The simplest nests are in reality no nest at all,
with the egg placed directly on the substrate. This is
seen in the ledge nesting common murre and thick-
billed murre (U. lotnvia). The murres bring a few
small stones and other objects to the nest site (Tuck
1961). Birds nesting on ledges are usually highly
visible at the proper angle of observation.

The simplest "nest" of ground nesting birds is
the scrape, a small hollow made in the soil, often
sand. This is common in terns, such as the Caspian
tern {Sterna caspia; Penland 1981). These nests are
precarious, and are often filled with drifting sand.
Nests of this type are difficult to see from a distance,
especially from aircraft. The birds themselves are
counted, unless the colony is entered.

The next order of nest complexity is seen in the
small rings of material that are constructed of gath-
ered vegetation and other debris. These nests have a
bottom and a rim that hold the eggs in place and
off the substrate. These are commonly seen in gulls
and cormorants. Nests of sticks, such as those built
by cormorants, often survive the winter and are used
again, with new material added. Other rim nests,
made of more substantial non-woody plant material
often survive to be used again. Many species, such as
the kittiwakes, compact the nest with mud giving
permanency to the nest. Nests of these types are
often clearly visible and are often easily counted
when the observer is remote from the colony.

Finally, in some species the nest can become a
sizeable structure. Such nests usually are the result
of use over several years, with each nesting pair

Nesting double-crested cormorants.

adding more material to the top of the nest (e.g.,
double-crested cormorant, Brandt's cormorant [P.
auritus], and brown pelican). These nests are easily
observed from a distance, are easily counted, and are
usually discernible in photographs (Sidle and Fergu-
son 1982).

Arboreal Nests. Arboreal nests are built at various
heights from just a few centimeters to tens of
meters. The nest support can vary from bulrushes,
willows, and sagebrush to large alders, cottonwoods,
or conifers. Nest construction material varies
considerably with the large nests of some herons and
cormorants, often containing sticks and small
branches. Nests are often in full view from the
ground or air; however, arboreal nests of some
species may be placed in concealed locations, and
locating them can be difficult. Nests placed in trees
may be fully exposed at the onset of nesting, but are
soon concealed by emerging leaves.

POPULATION MEASUREMENT CRITERIA

This section deals with the mechanisms and
necessary requirements to design and conduct a sur-
vey, monitor, or census program (McCrimmon
1976).

Biologists should remember, two important
points when working with colonial nesting water
birds: ( 1 ) that actions can severely affect the study
species, and (2) that any program undertaken must
be done well. This includes taking steps to ensure
the program results are available to present day re-
searchers and to those in the distant future (see
McCrimmon 1978b; Herman 1980). Following are
the requirements needed before a survey or moni-
toring and census program can be designed or
conducted.

What to Measure: Populations or Habitats

There are two different types of nesting distribu-
tions that must be considered. One is the distribu-
tion of the study species' colonies over often broad
geographic areas of one or more habitat types. The
other is the distribution of nests within colonies. The
approaches for sampling colonies versus nests are
usually different; the sampling program used depends
on the program goals. If the goal is general distribu-
tion and total numbers of colonies, with imprecise
knowledge of species and their numbers, then sur-
veys over large areas are employed. However, if ac-
curate numbers of species at each colony are
needed, then censuses of colonies are also required.

Waterbirds (those that commonly nest in colo-
nies) usually concentrate their nests in a few colo-
nies, leaving apparently suitable nesting habitat
unused. The colony sizes and species present usually

394

Colonial Waterbirds

vary. Sampling a portion of apparently available and
suitable habitat to derive values for the number of
colonies and individuals per unit area and then ex-
trapolating to the total area are done only with cau-
tion. The reliability of derived values is often
improved with larger samples, when colony size
variation is small, and species composition is con-
stant. To determine the total number of colonies
present for any species it is desirable to locate each
colony. However, this is not always practical and
extrapolations from samples must be considered.

When colonies are censused, the goal is usually
to determine the total number of nesting individuals
of each species present. The possibilities of accom-
plishing this depend on the species present, their
numbers, the visibility of nests, the resources avail-
able, and the limitations imposed by the conse-
quences of disturbance. Preferably, a direct count of
each nest in the colony by personnel visiting each
nest is most desirable. However, this is often not
possible or desirable. In such cases, samples of the
colony are obtained and values extrapolated to the
total area occupied by the colony. A variety of tech-
niques are available to obtain sample counts (Buck-
ley and Buckley 1978).

Where to Look

The process of locating colonies can take con-
siderable time and resources; therefore, learning
as much as possible about the study species' colonies
in the study area is needed before field work begins.
The time spent reading literature, field notes, and
reports, and talking with local observers is usually a
good investment. When surveys do begin, the known
sites can serve as starting points. Later as more expe-
rience is gained through study, other habitats can
be searched.

If the habitat occurs as widely scattered patches,
as continuous belts, or large expanses, the use of an
airplane for observations may be necessary. Depend-
ing on the species, colonies may be visible from
the air and easily plotted on maps. The use of a boat
may be required to locate colonies along rivers,
waterways, marshes, and coastlines. For secretive or
hard to observe species it may be necessary to
closely inspect likely colony sites, on foot or from a
boat.

The colonies of some species are not specific to
a particular habitat or a few habitats; other colonies
are concealed. In both cases, these colonies would
be difficult to locate by direct searches. The loca-
tions of colonies can be narrowed down to smaller
areas by observing where foraging birds are located.
Observers can return to these areas and attempt to
follow birds to their colonies. It is also possible to
determine the locations of colonies by observing
where birds emerge from vegetation when leaving

their colonies to forage. Nocturnal species may re-
quire systematically checking colony sites at night, a
potentially long process. Burrowing species may
require the excavation of burrows to determine spe-
cies composition of colonies.

Once the survey of an area or habitats is com-
pleted, observers can return to the located sites and
make systematic censuses of the species, their num-
bers and status. The methods used to locate colonies
in the beginning can then be used for monitoring
the colonies. Unfortunately there is no one way to
survey for the presence of colonies. The methods
must be adjusted for the study species, the habitats
present, the area covered, the resources available,
and the program goals. The researcher must decide
the precision needed to cover the target area. Should
all the colonies be located, or just a few for later
projects? In recent years, there were several pro-
grams to locate and count all breeding colonies and
species of the North American continental margin
(for examples see Blacklock et al. 1978; Clapp et al.
1982a, b; Erwin 1979; Erwin and Korschgen 1979;
Korschgen 1979; Nesbitt et al. 1982; Portnoy 1977;
Portnoy et al. 1981; Sowls et al. 1978, 1980; and
Speich and Wahl 1985).

How to Observe

Counting the number of active nests is an un-
likely prospect for many species at many sites, al-
though it can be done under ideal situations. Often,
nests are hard to find, and may be concealed in vege-
tation, or both. Even when nests can be counted, it
is unlikely that the status of every nest can be deter-
mined. Impacts from disturbance on the study spe-
cies also need to be considered. In this case,
observers must often obtain sample counts and sta-
tus determinations from portions of the colony and
apply correction factors to samples.

High density nesting species, such as white pelicans, suc-
cessfully reproduce with no social conflict.

Colonial Waterbirds

395

If nests cannot be counted, less precise methods
must be employed. These include determining the
net movements of adults to and from feeding areas,
counting adults at feeding areas and, for nocturnal
burrowing species, trying to determine numbers
from the sounds of adults. Although imprecise, these
methods are often the best that can be used.

When to Observe

Equally important to knowing how and what to
observe is to know when to observe. Counts of nests
or burrows, and adults on nests, are usually made
during daylight hours. The time of day observations
are made can be important. If adults are counted,
it is ideal to count when the adult-to-nest ratio ap-
proaches 2:1. When adults might be frightened from
their nests, especially if the nests contain eggs and
nestlings that would be exposed to overheating,
chilling, or predation, a time of day for observations
should be chosen when stress is minimal. Ideally,
numbers are best determined after the peak of egg
laying when most nests at a colony have incubating
adults.

MEASUREMENT TECHNIQUES
Counts of All Nests or Adults

to colony conditions and program goals. Whether
the samples are obtained by personnel in the colony,
at remote vantage points or in aircraft, or through
stereo photography, is decided on a colony-by-col-
ony basis. The sample plots are adjusted as necessary
to match the habitat differences within the colony.

If sample counts of adults are required, the pro-
cedures are similar to those used to obtain sample
counts of nests.

Adult Counts Away from the Colony

When estimates of the numbers of birds nesting
in a colony cannot be obtained, counting adults
away from the colony is an alternative. One method,
flight line counts, counts adult birds as they fly to
and from foraging areas. This is accomplished by
observing birds from vantage points that offer a clear
view of the area surrounding the colony. Numbers
derived from net movement counts are affected in
several ways. In part these are taken into considera-
tion in calculations to derive total numbers. The
total number derived can be affected by ( 1 ) the por-
tion of the colony's birds that come by the point of
observation, (2) the portion of the adults observed
that are actually nesting, and ( 3 ) the portion of the
birds observed that are from the subject colony and
not from other colonies.

Counting every nest is usually practical only in
small colonies; larger colonies can be counted when
there are sufficient personnel. When making direct
counts of burrows, ground nests, or arboreal nests, it
is desirable to mark each nest visited and then re-
count all nests, marked and unmarked, to derive
correction factors. It is useful to partition the colony
to ensure a thorough census. Counts from aircraft
are necessarily less precise, but often necessary.
However, counting nests from aerial photographs,
especially through stereo photography, is often a
practical way to determine numbers of nesting birds.
The counting of all nests, especially by observers in
the colony, is limited by the effects of disturbance.

As stated, counts of adults are desirable, but
often impractical, especially when nests are difficult
to recognize. Counts of adults are usually made from
remote vantage points, from aircraft, or through
stereo photography. Helicopters are often superior
to fixed wing aircraft, as they allow prolonged view-
ing of the colony from chosen vantage positions
above the colony.

Samples of Nests or Adults

Estimates of the total numbers of nesting birds
are derived from counts of sample sections in the
colony. Sample plots are established through random
strip transects, evenly spaced belts, point-centered-
quarter, random plots, or other methods suitable

Another method is to obtain counts of birds
when they are on foraging areas away from the col-
ony. The considerations are the same as for flight
line counts. Additionally, for diving birds, at any
moment of observation, a portion of the birds may
be below the water surface, out of view. Counts
of the birds are made directly from vantage points,
aircraft, or through stereo photography. The time
taken to understand the foraging schedule of the
birds is well spent.

Correction Factors

When definitive numbers of birds of any species
nesting at any given colony site cannot be obtained,
numbers can be approximated through a census
program. The number derived is qualified by the
date of the census in relation to the nesting stage of
the colony. Usually nests, burrows, or adults are
counted, but only a proportion of the adults nesting
at a colony may be present at the time of the census.
Likewise, a count of nests or burrows will contain
sites that are not active. And, of the active nests
found on any given census, only a portion will be
successful, producing fledging young.

Correction factors are applied to counts to bring
them more in line with the actual number of adults
probably present and nesting at the colony. The
usual approach is to correct for the nests or burrows
that are not active, and for those missed during a

396

Colonial Waterbirds

census. The only difference is that generally the cor-
rection factors for burrowing species are much
harder to obtain, especially for nocturnal species.

If adults are counted, the most important ratio
to determine is between adults present and the num-
ber of active nests. Thus, counts from the air or from
remote observation points are verified by site checks
of sample sections in the colony where the ratio is
determined (Nisbet 1973; Kadlec and Drury 1968).

Checking Accuracy

Determining the numbers of a species present
and nesting in a colony is often elusive because of
the known, uncontrollable and usually unquantifi-
able, variables involved. However, the numbers gen-
erated are often good. The accuracy or inaccuracy of
numbers obtained may preclude any detailed analysis
(Ralph 1981 ). This and large values may allow state-
ments only when numbers are different by factors
of 2, 4, or even orders of magnitude. The application
of statistical tests may indicate significant differences
or trends in the data, but significant statistical differ-
ences may not translate to significant biological dif-
ferences or be real.

Nevertheless, the accuracy of a survey or census
should be checked. There are two main approaches.
First, the survey or census can be repeated by the
same observers or different observers. Hopefully, all
methods are constant and the environmental condi-
tions remain the same. Obtaining the same colony
attendance pattern during the repeat of the survey
or census is desirable. Repeating a survey or census
with different methods could be employed when
there is reason to suspect the results of the original
survey or census.

Sample Methods and Comparisons

The following examples of methods used to
observe and obtain numbers of colonial nesting wa-
terbirds are only a sample; many other methods
could be included (Buckley and Buckley 1978). The
examples show the results of actual studies from a
variety of methods, on several species in different
habitats. These examples should help the researcher
select the methods needed to satisfy program goals.
Studies where different methods are compared are
presented in more detail to demonstrate the variabil-
ity of results and efficiency of methods in specific
situations. These examples also show that considera-
ble effort must be extended to obtain data on spe-
cies and their numbers, nesting in waterbird
colonies, and that the inherent difficulties should not
be underestimated!

Aerial Photography — Ground Nest Count. On

June 2 and 3, 1980, ground counts were conducted
and aerial photographs taken of nesting white

pelicans on Small Island, Chase Lake National
Wildlife Refuge, North Dakota (Sidle and Ferguson
1982). The aerial photography revealed 1,368
pelican nests, and the ground census 1,355 pelican
nests, a difference of only 1 % .

Ground Nest Counts — Duplications. On May 24,

1980, herring gulls (L. argentatus) and lesser black-
backed gulls (L. fuscus) nests were counted by 10
observers on 10 separate plots in Flat Holm, United
Kingdom (Ferns and Mudge 1981 ). The goal of the
observers was to systematically search their
respective plots for all nests and place a spray paint
mark by each nest found. The observers then
switched plots, and counted all marked and
unmarked plots, establishing a ratio of the two nest
counts. For nine plots the average percentage of
nests missed during the first count was 16.9 ± 33%
(SE). The range for nine plots was 5.0% to 27.3%.
In total 2,758 nests were marked and counted, and
the corrected total number of nests was 3,651.

Similar results were obtained from duplicating
the count and marking the nests in gull colonies
in Massachusetts. The error ranged from 4% to 22%
on eight sets of counts. Plot totals varied from 27
to 442 nests, on plots of 0.06 to 0.73 ha (0.15 to
1.8 a.) with varied vegetation density (Erwin
1980b). Even in sparsely vegetated small plots, there
is considerable error in finding ground nests.

Ground Censuses of Ground Nests. A few of the

methods used to determine the numbers of ground
nesting species by counting on the ground are
presented here. Numbers of ring-billed gulls (L.
delaivarensis) nesting in a colony were determined
by dividing the colony into sections and then
counting all the nests in each section (Southern and
Southern 1981 ). Total counts of nesting gulls and
terns in colonies were obtained by counting and
marking all the nests in distinct areas of the colony.

Ring-billed gull.

Colonial Waterbirds

397

The areas were then counted again and the numbers
of marked and unmarked nests were recorded. A
Lincoln Index was applied to get the "true" total
(Erwin 1980b).

Strip transects were randomly placed through
large colonies, and all nests were counted in the
transects. These transect counts were continued un-
til 10% to 20% of the colony area was surveyed.
The densities obtained were used to calculate the
total numbers nesting in the whole colony ( Erwin
1980b).

In large wading bird colonies, belt transect
counts are a practical way to obtain an idea of the
total numbers nesting. Transects are made through
the colony and all nests are counted in the belt.
These are repeated until 10% of the total colony
area is censused. This technique at least obtains or-
der of magnitude counts for a colony (Portnoy
1980). Although this technique may be practical for
arboreal or surface ground nesting species, the pros-
pect of obtaining a 10% area sample in a large bur-
rowing species colony is discouraging.

In a burrowing species colony, the nests, in
many cases, must be individually examined to iden-
tify species and to determine whether the nests are
active before correction factors can be calculated
(pers. obs.). The point-centered-quarter technique of
obtaining samples has been used. Between 8 and 1 5
samples are taken along an arbitrarily drawn line
through a colony. However, the strip transect
method is recommended (Erwin 1980b). Total
counts of nests are practical only in small colonies
(Portnoy 1980) and the accuracy can be quite good
(pers. obs.).

Censusing — Helicopter, Airplane, and Ground
Nest Counts. The census results obtained from
a helicopter, a fixed wing aircraft, and on the
ground, of a mixed species colony of mangrove
(Rhizosphora mangle) were compared (Kushlan
1979). For large white and dark birds, nesting in the
tree tops, the helicopter results had an error 10%
to 16% of the ground nest count. The error for fixed
wing aircraft compared to ground censuses ranged
from 32% to 100% (also see Erwin 1980a).

142%. Five samples were below 11%, and one each
at 26% and 54% . Point-centered-quarter error rates
ranged from 3% to 400% . Most were larger than the
corresponding strip transect sample errors. The strip
transect method is apparently equally reliable
regardless of the spatial distribution of the sample
population. Laboratory studies with artificial
populations — random, uniform, and patchy — yielded
similar results (Erwin 1980b).

Flight Line Count — Ground Count. Erwin (1981)
compared flight line counts and ground nest count
results at 1 3 mixed species wading bird colonies. For
three hours in the morning, all birds entering and
leaving colonies were counted and identified. Then
all nests were counted or belt transects were used to
generate the total nests present. Species and colony
differences in flight rate were observed. Individual
colony size error reached a factor of 2, and the
actual populations were predicted with 10% error.
Species correction factors to flight line data may
be necessary. Hour to hour variation in flight rates
can be substantial, and the calculation of a mean rate
is suggested. For the use of flight line counts,
comparisons with other methods of obtaining the
number of nesting birds present, and for diurnal
burrowing species, see the discussion in Byrd et al.
(1983).

Aerial Visual Estimates, Aerial Photograph
Counts, and Transect Sampling. Portnoy (1980)
compared aerial visual estimates, aerial photograph
counts, and transect counts in 19 colonies of nesting
wading birds. Aerial visual estimates of great egrets
(Casmerodius albus) at these colonies were greater
4 times and less 15 times than the actual numbers
present. Of 33 colonies, mean colony size was 511
nests; the average error of the visual estimates was
— 4.2% (SD 12.2). Of six snowy egret colonies,
mean size was 1 ,992 nests; the visual error was
-8.7% (SD 92.0). For Louisiana heron {Hydranassa
tricolor) colonies, n = 8; mean size, 3,192 nests; and
the visual error was 79.1% (SD 25.1). And in black-
crowned night heron colonies, n = 7; mean size, 573
nests; and the visual error was —84.4% (SD 14.7).
Clearly, aerial- visual samples of colonial nesting
wading birds greatly underestimate numbers of
smaller cryptic and concealed nesting species in
trees or marsh vegetation (see also Erwin 1980a).

Census Accuracy — Total Count, Strip Count, and
Point-Centered-Quarter. Three census methods
were compared in six gull and three heron colonies
(Erwin 1980b). Strip transect and point-centered-
quarter methods were compared to known total
counts in the colonies. Sample areas were within
10% to 20% of plots for the strip transect method.
Between 8 and 1 5 points were sampled for the fixed
point-centered-quarter method. Only the strip
transect method showed reasonable accuracy in the
field, and the observer error ranged from 3% to

Gull Colony Ground Nest Counts, Aerial
Photograph Counts, and Aerial Visual Estimates.

Three methods of determining nesting numbers of
gulls at colonies were compared (Kadlec and Drury
1968). Photographs of colonies and visual estimates
were made of colonies from helicopters. Double
ground counts of gull nests were made in the
colonies. On the first ground count, 95% of all nests
were found. At some colonies, two additional ground

39S

Colonial Waterbirds

counts were made. The ratios of adults, counted on
photographs and from aerial estimates, to ground
counts were established for all sites. Ratios tended to
be similar within each census, although the ratios
varied between censuses. Confidence intervals are
given for predicting the number of nests from
photograph counts of adults and visual estimates of
adults. For regional determinations of gull numbers,
visual estimations are a valid approach. More
detailed studies within a region require photography.
In predicting numbers from visual estimates and
photograph counts, errors of 19% to 31% and 26%
to 35% , respectively, can be expected. Annual
changes of 25% or less may not be detectable with
either method.

Belt Transect Sampling and Confidence Limits.

In 1976, gulf coast heronries were sampled with
randomly placed ground belt transects, in which all
nests were sampled (Portnoy 1980). Transects were
run until 10% of the colony area was sampled. The
95% confidence limits generated for nest densities
per transect were large: for estimated numbers of
nests, n= 16,880; the confidence interval was 2,672
nests; n= 14,938, 3,090; n= 14,279, 1,809;
n= 12,600, 1,337; n = 3,240, 810. When dealing with
largely inaccessible colonies, i.e., concealed nests,
colonies over water, and cryptic species, this is
about the best one can expect. When surveying large
areas, and a large number of difficult large colonies,
economic and personnel limits may not allow better
work. In large colonies, order-of-magnitude
measurements may be all that can be expected.

Estimating Numbers from Aerial Photographs —
An Experiment. Nine observers were asked to
estimate the numbers of birds on 10 different
photographs each of 5 consecutive days. The
observers were in three groups: inexperienced, past
experience, and recent experience. The effects of
observer differences, prior experience, training, and
numerical magnitude were examined. With no
training, experienced observers were more accurate
than inexperienced observers. Trends in
underestimation were observed. With training,
estimates by eight observers were within 10% of the
actual number. This study by Erwin (1982) is
interesting and suggests that training prior to
censuses or surveys is important.

Aerial Photography of Large Colonies of Cliff
and Island Top Nesting Species. Aerial
photography was used to determine colony size of
nesting common murres (Nettleship 1980). This
species can nest in numbers in the 100,000s. It nests
on cliff ledges, the flat tops of islands, in boulder
fields, and in caves. There is no easy way to census
these large colonies. Correction factors for counts
from photographs are generated from field counting
and counting from photographs, groups of about 200

birds. If photographs cannot be obtained for all
areas, then numbers may be estimated for these
areas from densities in the known areas. The total
"number" may be a composite from photograph
counts, area extrapolations, and direct counts or
estimates from land or aircraft. Correction factors for
flat, accessible island top areas are generated by
counting birds on photographs, field counting adults,
or both, and then flushing the adults and counting
their eggs in distinct areas. For large colonies, this
procedure is very time-consuming, taking one
observer up to 6 weeks to complete a census. To
obtain population samples and detect population
trends, numbers are determined in sample plots and
extrapolated to known areas of similar habitat. In a
large colony, this can take an observer up to 10 days.
Several photographs appear in this publication
showing birds, habitat, and sampling areas (see also
Nettleship 1976).

Aerial Photography of Cliff and Island Top
Nesting Species. Aerial photographs of nesting
gannets on ledges, on cliff faces, and on flat ground
at the tops of cliffs were used to determine nesting
numbers (Nettleship 1975). Overlapping aerial
photographs of incubating birds were taken from a
fixed-wing aircraft with a 35mm camera, using a
50mm lens and Kodak Plus-X, black-and-white film.
The photographs were taken about 600 m ( 1 ,980 ft )
from the colony. Glossy prints, 7 x 10 in. or 9 x
1 3 in. were made. It was possible to count the
number of birds associated with each nest from the
photographs by using hand lenses and 1 x 1 cm
(0.4 in.) plastic overlays. This gave a count of nest
site holders, not the actual number of birds that
build nests and actually lay eggs. The error was
thought to be less than 20% . The error largely
resulted from the imprecise boundaries on

Common murre and nest in typical cliff ledge site.

Colonial Waterbirds

399

photographs and difficulty in counting the nests at
the back of broad, flat areas at cliff tops.

Aerial Photograph Counts and Estimates of
Ground Nests. Estimates of the total numbers of
ground nests were made from airplanes, and these
were corrected with ground counts of nests (Erwin
1980a). Counts of ground-nesting terns and black
skimmers {Runchops niger) were determined from
aerial photographs and then corrected with ground
counts of nests (Portnoy 1980).

Aerial Estimates of Arboreal Nests. Total
estimates of all visible arboreal nests were made
from an airplane, and the counts were then
corrected with "good" data (nest counts, etc.; Erwin
1980a). Large, conspicuous wading birds can be
counted in this manner, but the technique is not
suitable for determining the numbers of cryptic, and/
or concealed, nesting species (Portnoy 1980).

DISTURBANCE

Disturbance includes any activity by an observer
that results in any perturbation of the "normal" con-
dition on the colony under observation. The distur-
bance created ranges from that undetectable by
observers to major disruptions of the colony (Table
2). Unless observations are conducted from a posi-
tion remote and distant to the colony, they will re-
sult in disturbance to the colony.

Sources of Disturbance

The nature and magnitude of disturbance often
depend on the experience and skill of the observers.
The level of disturbance is usually a function of the
platform of observation, and the techniques used
to make observations. Table 2 lists types and effects
of disturbance on several colonial-nesting species,
with references to pertinent studies.

Table 2. Types and Effects of Disturbance.

Types of Disturbance

Authors

Species Groups

AIRCRAFT:

HELICOPTERS:
OBSERVERS IN OR NEAR
COLONIES:

Dunnet 1977
Burger 1980
Kushlan 1979
Burger 1981
Robert and Ralph 1975

seabird colonies
herring gulls
wading bird colonies
gulls
gulls

Effects of Disturbance

AGGRESSIVE BEHAVIOR:

Burger 1980, 1981
Robert and Ralph 1975

gulls
gulls

THERMAL STRESS:
(when adults leave eggs or
nestlings unattended)

Bartholomew et al. 1953

Bartholomew and Dawson 1954

Vermeer 1963

Drent et al. 1964

Harris 1964

Drent 1967

Hunt 1972

Anderson and Keith 1980

Dawson and Bennett 1981

Bennett et al. 1981

white pelicans
pelicans, gulls, herons
glaucous-winged gulls
seabird colonies
gulls

herring gulls
herring gulls
brown pelicans
Western gulls
Western gulls

PREDATION:

Drent et al. 1964
Verbeek 1982
Manuwal 1974b
Penland 1981

by crows at colonies
by crows at colonies
by gulls at colonies
of Caspian tern eggs

PREDATION FROM OBSERVER-
INDUCED DISTURBANCES:

Anderson and Keith 1980
Anderson and Keith 1980
Ellison and Cleary 1978
DesGranges and Reed 1981
Buckley and Buckley 1972
Kury and Gochfeld 1975
Johnson 1938

pelicans by gulls and ravens
tern eggs/young by gulls
cormorant eggs by gulls
eggs/young by gulls
tern eggs by gulls
cormorant eggs by gulls
murre eggs/young by gulls

CHICK MOBILITY:
(flushing of young from ground
nests)

Robert and Ralph 1975
Penland 1981
Burger 1981

gulls

Caspian terns
gulls

400

Colonial Waterbirds

Table 2. Types and Effects of Disturbance (concluded).

Effects of Disturbance

Authors

Species Groups

NEST AND COLONY
ABANDONMENT:
(disturbance during nesting)

Werschkul et al. 1976
Manuwal 1978
Anderson and Keith 1980
Jehl 1973

Tremblay and Ellison 1979
Ellison and Cleary 1978
Burger 1981
Penland 1981
Manuwal 1978
Manuwal 1978

great blue herons

storm petrels

brown pelicans

brown pelicans

black-crowned night herons

double-crested cormorants

gulls

Caspian terns

tufted puffins

rhinoceros auklets

DESTRUCTION OF NESTS:

Giles and Marshall 1954

marsh-nesters

CANNIBALISM:
(killing of eggs and young by
conspecific adults)

Anderson and Keith 1980

Hand 1980

Harris 1964

Hunt and Hunt 1975, 1976

Parson 1971

Patterson 1965

Penland 1981

Robert and Ralph 1975

Vermeer 1963

seabirds
Western gulls
gulls
gulls

herring gulls
black-headed gulls
Caspian terns
Western gulls
glaucous-winged gulls

REPRODUCTIVE SUCCESS:
(additional studies)

Anderson et al. 1976
Ashmole 1963
Brun 1972
Gillett et al. 1975
Kadlec and Drury 1968b
Schreiber 1979
Schreiber 1979

seabirds

tropical oceanic birds

gannet

glaucous-winged gulls

herring gulls

brown pelicans

brown pelicans

DISCUSSION AND CONCLUSIONS

A large variety of species of birds are considered
colonial nesting waterbirds. They are found in the
interior of the continent in association with aquatic
systems and in marine environments along the conti-
nental margin. The habitat preferences for nest sites
vary between species and often within species. Nests
range from burrows dug in the soil, ground nests of
various sizes and types, floating nests, and arboreal
nests. Nests can be completely concealed or easy to
see. The feeding methods, positions on food chains,
foraging habitats, and feeding schedules are varied,
and the sizes of the species range from tens of grams
to the low kilograms. These factors are usually im-
portant in the design and execution of any program.

The large variance in the breeding biology of
colonial nesting waterbirds requires that many differ-
ent measurement techniques be employed in survey,
census, and monitoring programs. The specific meas-
urement techniques depend on the program goals,
the resources available to perform the program, the
specific characteristics of the placement and visibility
of nests and nesting adults, the size of the nesting
colonies, the geographic distribution of the colonies,
and the effects of disturbance from performing the

program. Unfortunately, no one method or few
methods can be considered for obtaining quantified
information. Each program must be tailor-made and
adapted to each species and program goals. It is ad-
visable to be well-prepared before a program begins,
but flexibility must be allowed for inevitable adjust-
ments. However, despite all the inherent problems in
working with colonial nesting waterbirds, reasonably
accurate results can be obtained from measurement
programs.

A variety of methods can be employed to meas-
ure the numbers of individual species in colonial
nesting waterbird colonies. Observations are made in
the colony, from the edge of the colony or other
remote positions on land, from boats or aircraft, and
by stereo photography. Aircraft are efficient when
large areas need to be covered quickly. Observations
and photographs of colonies are made from aircraft.
Helicopters provide stable platforms and can be ac-
curately positioned above colonies for prolonged
viewing and photography. Population estimates can
be derived from counts of nests or adults. Counts of
adults or nests by stereo photography are often effi-
cient. In some situations, every nest or adult in a
colony can be counted, but most often, sample
counts from plots in the colony are required.

Colonial Waterbirds

401

A variety of methods can be useful for partition-
ing a colony for sample counts. Correction factors
should be applied to census numbers to adjust for
inactive nests, adults not nesting, adults not present
in the colony, when counts of adults are made, etc.
Duplicate counting of nests or adults in colony plots
generates accuracy correction factors for the per-
centage of nests or adults missed on the first count-
ing. Additional counts or estimates can be obtained
by using a completely different method to check
accuracy.

Before colonies can be censused, they first have
to be located. This is accomplished first by searching
literature, reports, and field notes, and interviewing
local observers. This first step can save time and give
good indications of the locations of colonies, the
species present, and the habitats they occupy. If the
goal of the program is to find all colonies, then wide
area searches are necessary, with the first effort con-
centrated in likely habitats or locations. Surveys are
conducted on foot and from cars, boats, and aircraft.
Once the species, habitats, and geography become
familiar, the survey can be refined. If the program
goal is the numbers of each species in one or more
colonies, then a census is conducted at each colony.

The main factor limiting the methods and ulti-
mately the accuracy of the data obtained from pro-
grams, especially censuses of colonies, is the effects
of observer-induced disturbance to the nesting birds.
Disturbance to colonies in the past has resulted in
significant mortality to eggs and nestlings. Every
precaution must be taken to ensure that disturbance
by observers is kept to a minimum. When working
in and near colonies, constant monitoring of the
colony is required. The effects of disturbance are not
always obvious and immediately evident.

Colonial waterbirds are censused from the ground, at
water level, and from the air.

402

Colonial Waterbirds

LITERATURE CITED

AINLEY, D.G. and T.J. LEWIS. 1974. The history of the
Farallon Island marine bird populations, 1854-1972.
Condor 76:432-446.

ANDERSON, D.W. and JO. KEITH. 1980. The human
influence on seabird nesting success: Conservation
implications. Biol. Conserv. 18:65-80.

,J.E. MENDOZA, and JO. KEITH. 1976. Seabirds in

the Gulf of California: A valuable, international re-
source. Nat. Resour. J. 16:483-505.

ASHMOLE, N.P. 1963. The regulation of numbers of tropi-
cal oceanic birds. Ibis 103b:458-473

BAILEY, V. and MR. TERMAN. 1983. A comparative study
of a Great Blue Heron colony in Chase County, Kan-
sas. Trans. Kansas Acad. Sci. 86:81-88.

BARTHOLOMEW, G.A. and W.R. DAWSON. 1954. Tern
perature regulation in young pelicans, herons, and
gulls. Ecol. 35:466-472.

, , and E.J. O'NEILL. 1953. A field study of

temperature regulation in young white pelicans, Pele-
canus erythrotyncbos. Ecol. 34:554-560.

BAYER, R.D. and E. MCMAHON. 1981. Colony size and
hatching synchrony of Great Blue Herons in coastal
Oregon. Murrelet 62:73-79.

BENNETT, A.F., W.R. DAWSON, and R.W. PUTMAN. 1981.
Thermal environment and tolerance of embryonic
western gulls. Physiological Zoology 54:146-154.

BENT, AC. 1919. Life histories of North American diving
birds. U.S. Natl. Mus. Bull. 107.

. 1921. Life histories of North American gulls and

terns. U.S. Natl. Mus. Bull. 1 1 3.

. 1922. Life histories of North American petrels and

tion 24:177-184.
— . 1981. Effects of human disturbance on colonial

pelicans and their allies. U.S. Natl. Mus. Bull. 176.
— . 1927. Life histories of North American marsh

birds. U.S. Natl. Mus. Bull. 135.
BIRKHEAD, T.R. 1978. Attendance patterns of Guillemots

(Uria aalge) at breeding colonies on Skomer Island.

Ibis 120:219-229-
BLACKLOCK, G.W., DR. BLANKINSHIP, S.D. KENNEDY,

KA. KING, R.T. PAUL, R.D. SLACK, J.C. SMITH, and

R.C. TELFAIR II, (compilers). 1978. Texas Colonial

Waterbird Census, 1973-1976. Texas Park Wildl.

Dept., FA Rep. 1 5.
BRUN, E. 1972. Establishment and population increase of

the Gannet (Sulci bassana) in Norway. Ornis Scand.

3:27-38.
BUCKLEY, F.G. and PA. BUCKLEY. 1972. The breeding

ecology of royal terns, Sterna ( Thalasseus ) maxima

maxima. Ibis 114:344-359.
and . 1980. Habitat selection and marine

birds. Pages 69-112 in Burger, J., B.L. Olla, and HE.

Winn, eds. Behavior of Marine Animals, Vol. 4.

Plenum Publishing Co.
BUCKLEY, PA. 1978. Guidelines for the protection and

management of colonially nesting waterbirds. North

Atlantic Regional Office, Natl. Park Serv., Boston, MA.
and F.G. BUCKLEY. 1977. Hexagonal packing of

royal tern nests. Auk 94:36-43-
BURGER, J. 1976. Daily and seasonal activity patterns in

breeding laughing gulls. Auk 93:308-32.3.
. 1978. The pattern and mechanism of nesting in

mixed-species heronries. Pages 45-58 in Sprunt, A.S.,

IV, J.C. Ogden, and S. Winckler, eds. Wading birds.

Natl. Audubon Soc, Res. Rep. 7.
. 1980. Behavioral responses of herring gulls (Larus

species, particularly gulls. Colonial Waterbirds
4:28-36.

BYRD, G.V., R.H. DAY, and E.P. KNUDTSON. 1983. Pat-
terns of colony attendance and censusing of auklets at
Buldir Island, Alaska. Condor 85:274-280.

CLAPP, R.B., R.C. BANKS, D. MORGAN-JACOBS, and W.A.
HOFFMAN. 1982a. Marine birds of the southeastern
United States and Gulf of Mexico. Part I. Gaviiformes
through Pelecaniformes. U.S. Dept. Inter., Fish Wildl.
Serv., FWS/OBS-82/01.

, , , and . 1982b. Marine birds

of the southeastern United States and Gulf of Mexico.
Part II. Anseriformes. U.S. Dept. Inter., Fish Wildl.
Serv., Biol. Serv. Prog, FWS/OBS-82/02.

CONOVER, MR. and DO. CONOVER. 1981. A docu-
mented history of ring-billed gull and California gull
colonies in the western United States. Colonial Water-
birds 4:37-43.

and D.E. MILLER. 1980. Daily activity patterns of

breeding Ring-billed and California Gulls. J. Field
Ornith. 51:329-339.

CUSTER, T.W. and R.G. OSBORN. 1978. Feeding habitat
use by colonially breeding herons, egrets, and ibises
in North Carolina. Auk 95:733-743.

DAWSON, W.R. and A.F. BENNETT. 1981. Field and labo-
ratory studies of the thermal relations of hatching
western gulls. Physiological Zoology 54:155-164.

DESGRANGES, J.L 1979. Abandoned windmill used as a
nesting site by great blue herons. Can. Field-Naturalist
93:439-440.

. 1980. Competition entre le cormoran a aigrettes et

Le Grand Heron au moment de la nidification. Natu-
raliste Can. 107:199-200.

and P. LA PORTE. 1979. Second tour of inspection

of Quebec heronries, 1978. Can. Wildl. Serv., Prog.
Notes 105.

and . 1981. Third tour of inspection of Que-

bec heronries, 1979. Can. Wildl. Serv., Prog. Notes

123.
— . and A. REED. 1981. Disturbance and control of

selected colonies of double-crested cormorants in

Quebec, Colonial Waterbirds 4:12-19.
— , , and G. CHAPELAINE. 1979. First tour of

inspection of Quebec heronries, 1977. Can. Wildl.
Serv., Prog. Notes 93.

DRENT, R.H. 1965. Breeding biology of the pigeon guille-
mot (Cepphus columba). Ardea 53:99-160.

. 1967. Functional aspects of incubation in the her-
ring gull (Larus argentatiis Pont. ). Behaviour Supple-
ment 17:1-32.

and C.J. GUIGUET. 1961. A catalog of British Co-
lumbia seabird colonies. Occ. Papers British Columbia
Provincial Mus. 12.

, G.F. VAN TETS, F. TOMPA, and K VERMEER.

argentatiis) to aircraft noises. Environmental Pollu-

1964. The breeding birds of Mandarte Island, British
Columbia. Can. Field-Naturalist 78:208-263.

DUNNET, G.M. 1977. Observations on the effects of low
flying aircraft at seabird colonies on the coast of
Aberdeenshire, Scotland, Biol. Conserv. 12:55-63-

ELLISON, L.N. and L. CLEARY. 1978. Effects of human dis-
turbance on breeding double-crested cormorants.
Auk 95:510-517.

EMLEN, S.T. and N.J. DEMONG 1975. Adaptive signifi-
cance of synchronized breeding in a colonial bird: A
new hypothesis. Science 188:1029-1031.

ENGLISH, S.M. 1978. Distribution and ecology- of Great

Colonial Waterbirds

403

Blue Heron colonies on the Willamette River, Ore-
gon. Pages 235-244 in Sprunt, AS., IV, J.C. Ogden,
and S. Winckler, eds. Wading birds. Natl. Audubon
Soc, Res. Rep. 7.

ERWIN, R.M. 1979. Coastal waterbird colonies: Cape Eliza-
beth, Maine, to Virginia. U.S. Dept. Inter. Fish Wildl.
Serv., Biol. Serv. Prog., FWS/OBS-79/10.

. 1980a. Censusing colonial waterbirds: Problems

and progress. Atlantic Naturalist 33:19-22.

. 1980b. Censusing waterbird colonies: Some sam-
pling experiments. Trans. Linnaean Soc. New York
9:77-86.

. 1981. Censusing wading bird colonies: An update

on "flight-line" count method. Colonial Waterbirds
4:91-95.

. 1982. Observer variability in estimating numbers:

An experiment. J. Field Ornithol. 53:159-167.1

, J. GALLI, and J. BURGER. 1981. Colony site dy-
namics and habitat use in Atlantic coast seabirds. Auk
98:550-561.

and C.E. KORSCHGEN. 1979. Coastal waterbird

colonies: Maine to Virginia, 1977. U.S. Dept. Inter.
Fish Wildl. Serv., Biol. Serv. Prog., FWS/OBS-79/08.

EVANS, R.M. 1982. Flock size and formation in black-billed
gulls. Can. J. Zoology 60:1806-1811.

FERNS, P.N. and G.P. MUDGE. 1981. Accuracy of nest

counts at a mixed colony of herring and lesser black-
backed gulls. Bird Study 28:244-246.

GALUSHA, J.G., Jr. and CJ. AMLANER, Jr. 1978. The ef-
fects of diurnal and tidal periodicities in the numbers
and activities of herring gulls (Larus argentatus) in
a colony. Ibis 120:322-328.

GIBSON, J.A. 1950. Methods of determining breeding cliff
populations of guillemots and razorbills. British Birds
43:329-331.

GILES, L.W. and D. B. MARSHALL. 1954. A large heron and
egret colony on the Stillwater Wildlife Management
Area, Nevada, Auk 71:322-325.

GILLET, W.H., J.L. HAYWARD, Jr., AND J.F. STOUT. 1975.
Effects of human activity on egg and chick mortality
in glaucous-winged gull colony. Condor 77:492-495.

GODFREY, WE. 1966. The birds of Canada. Nat. Mus.
Can., Bull. 203, Biol. Serv. 73.

HAND, J.L. 1980. Human disturbance in western gull (La-
nds occidentalis livens) colonies and possible amplifi-
cation by intraspecific predation. Biol. Conserv.
18:59-63.

HARRIS, MP. 1964. Aspects of the breeding biology of the
gulls (Larus argentatus, L. fiiscns, and L. marinits).
Ibis 106:432-456.

HENNY, C.J. and J.E. KURTZ. 1978. Great blue herons
respond to nesting habitat loss. Wildl. Soc. Bull.
6:35-37.

HERMAN, S.G. 1980. The naturalist's field journal: A man-
ual of instruction based on a system established by
Joseph Grinnell. Published privately, S.G. Herman,
Biol. Dep., The Evergreen State College, Olympia, WA.
200pp.

HUNT, G.L., Jr. 1972. Influence of food distribution and
human disturbance on the reproductive success of
herring gulls Ecol. 53:1051-1061.

and M.W. HUNT. 1975. Reproductive ecology of

the western gull: The importance of nest spacing. Auk
92:270-279.

and . 1976. Gull chick survival. The signifi-
cance of growth rates, timing of breeding and terri-
tory size. Ecol. 57:62-75.

JEHL, JR., Jr. 1973 Studies of a declining population of

brown pelicans in northwest Baja, California. Condor

75:69-79.
JOHNSON, R.A. 1938. Predation of gulls in murre colonies.

Wilson Bull. 50:161-170.
KADLEC, J.A and W.H. DRURY. 1968a. Structure of the

New England herring gull population. Ecol.

49:644-676.
and . 1968b. Aerial estimation of the size of

gull breeding colonies. J. Wildl. Manage. 32:287-293.
KORSCHGEN, C.E. 1979. Coastal waterbird colonies:

Maine. U.S. Dept. Inter., Fish Wildl. Serv., Biol. Serv.

Prog., FWS/OBS-70/07.
KREBS, JR. 1974. Colonial nesting and social feeding as

strategies for exploiting food resources in the great

blue heron (Ardea herodias). Behaviour 51:99-134.
. 1978. Colonial nesting in birds, with special refer-
ence to the Ciconiformes. Pages 299-311 in Sprunt,

AS., IV, J.C. Ogden, and S. Winckler, eds. Wading

Birds. Natl. Audubon Soc, Res. Rep. 7.
KURY, C.R. and M. GOCHFELD. 1975. Human interference

and gull predation in cormorant colonies. Biol. Con-
serv. 8:23-24.
KUSHLAN, J.A. 1978. Feeding ecology of wading birds.

Pages 249-297 in Sprunt, A.S., IV, J.C. Ogden, and S.

Winckler, eds. Wading Birds. Natl. Audubon Soc, Res.

Rep. 7.
-. 1979. Effects of helicopter censuses on wading

bird colonies. J. Wildl. Manage. 43:756-760.
LLOYD, C.S. 1972. Attendance at auk colonies during the

breeding season. Skokholm Bird Observation Rep.

1972:15-23.
LOWE, FA. 1954. The heron. Collins, St. James Place,

London.
MANUWAL, DA. 1974a. Effects of territoriality on breed-
ing in a population of Cassin's Auklet. Condor

55:1399-1406.
. 1974b. The natural history of Cassin's auklet (Pty-

choramphns aleuticus). Condor 76:421-431-
. 1978. Effect of man on marine birds: a review.

Proc J.S. Wright Forestry Conf. 4:140-160.

MASSEY, B.W. and J.L ATWOOD. 1981. Second-wave

nesting of the California least tern: Age composition
and reproductive success. Auk 98:596-605.

MCCRIMMON, DA. 1976. A review of some methods and
considerations for the assessment of breeding popula-
tions of colonial waterbirds. Proc. 4th Ann. Meeting,
Texas Fish-Eating Bird Conf. 1976:36-49.

MCCRIMMON, DA, Jr. 1978a. Nest-site characteristics
among five species of herons on the North Carolina
coast. Auk 95:267-280.

. 1978b. The collection, management, and exchange

of information on colonially nesting birds. Pages 187-
196 in Sprunt, A.S., IV. J.C. Ogden, and S. Winckler,
eds. Wading Birds, Natl. Audubon Soc, Res. Rep. 7.

MEYERRIECKS, A.J. I960. Comparative breeding behavior
of four species of North American herons. Publ. Nut-
tal Ornith. Club 2.

NESBITT, S.A., J.C. OGDEN, H.W. KALE, II, B.W. PATTY,
and LA. ROWSE. 1982. Florida atlas of breeding sites
for herons and their allies: 1976-78. U.S. Dept. Inter.,
Fish Wildl. Serv., Biol. Serv. Prog., FWS/OBS-81/49.

NETTLESHIP, D.N. 1975. A recent decline of gannets,
Morns bassanus, on Bonaventure Island, Quebec.
Can. Field-Naturalist 89:125-133-

. 1976. Census techniques for seabirds of arctic and

eastern Canada. Can. Wildl. Serv., Occ Pap. 25.

. 1980. Census methods for murres, Uria species: A

unified approach. Can. Wild. Serv., Occ. Pap. 43-

404

Colonial Waterbirds

NISBET, I.C.T. 1973- Terns in Massachusetts: Present num-
bers and historical changes. Bird-Banding 44:27-55.
OGDEN, J.C. 1978. Recent population trends of colonial

wading birds on the Atlantic and Gulf coastal plains.

Pages 137-153 in Sprunt, AS., IV, J.C. Ogden, and

S. Winckler, eds. Wading Birds. Natl. Audubon Soc,

Res. Rep. 7.
PALMER, R.S. (ed.). 1962. Handbook of North American

birds. Vol. 1. Loons through flamingos. Yale Univ.

Press., New Haven, CT. 567pp.
PARSON, J. 1971. Cannibalism in herring gulls. British

Birds 64:528-537.
PATTERSON, I.J. 1965. Timing and spacing of broods in

the black-headed gull, Lanis ridibundus. Ibis

107:433-459.
PENLAND, S. 1981. Natural history of the Caspian tern in

Grays Harbor, Washington. Murrelet 62:66-72.
. 1982. Distribution and status of the Caspian tern

in Washington State. Murrelet 63:73-79.
PORTNOY, J.W. 1977. Nesting colonies of seabirds and

wading birds — coastal Louisiana, Mississippi, and

Alabama. U.S. Dept. Inter., Fish Wild!. Serv., Biol. Serv.

Prog., FWS/OBS-77/07.
. 1980. Censusing methods for gulf coast waterbirds.

Trans. Linnean Soc. New York 9:127-134.
-, R.M. ERWIN, and T.W. CUSTER. 1981. Atlas of gull

and tern colonies. North Carolina to Key West, Flor-
ida (including pelicans, cormorants, and skimmers).

U.S. Dept. Inter., Fish Wildl. Serv., Biol. Serv. Prog.,

FWS/OBS-80/05.
PRATT, H.M. 1970. Breeding biology of great blue herons

and common egrets in central California. Condor

72:407-416.
. 1972. Nesting success of common egrets and great

blue herons in the San Francisco Bay region. Condor

74:447-453.
. 1973. Breeding attempts by juvenile great blue

herons. Auk 90:897-899.
. 1974. Breeding of great blue herons and great

egrets at Audubon Canyon Ranch, California. Western

Birds 5:127-136.

1980. Directions and timing of great blue heron

foraging flights from a California colony: Implications
for social facilitation of food finding. Wilson Bull.
92:489-496.

RALPH, C.J. 1981. Terminology used in estimating num-
bers of birds. Studies Avian Biol. 6:577-578.

ROBERT, H.C. and C.J. RALPH. 1975. Effects of human

disturbance on the breeding success of gulls. Condor
77:495-499.

RODGERS, J.A., Jr. 1978. Display characteristics and fre-
quency of breeding by subadult Little Blue Herons.
Pages 35-39 in Sprunt, A.S., IV, J.C. Ogden, and S.
Winckler, eds. Wading Birds. Natl. Audubon Soc, Res.
Rep. 7.

RYDER, R.A. 1978. Breeding distribution, movements, and
mortality of snowy egrets in North America. Pages
197-205 in Sprunt, AS., IV, J.C. Ogden, and S. Winck-
ler, eds. Wading Birds. Natl. Audubon Soc, Res.
Rep. 7.

SCHREIBER, R.W. 1979. Reproductive performance of the
eastern brown pelican. Contr. Sci., Natl. Hist. Mus.
Los Angeles Co. CA.317T-43.

and E.A. SCHREIBER. 1972. Studies of the brown

pelican. Wilson Bull. 84:119-135.

SEALY, S.G. and J. BEDARD. 1973. Breeding biology of the
parakeet auklet (Cyclorrhynchus psittacula) on St.
Lawrence Island, Alaska. Astarte 6:59-68.

SIDLE, J.G. and EL. FERGUSON. 1982. White pelicans
populations at Chase Lake, North Dakota, evaluated
by aerial photography. Prairie Naturalist 14:13-26.

SLATER, J. B. 1980. Factors affecting the numbers of guille-
mots ilria aalge present on cliffs. Ornis Scand.
11:155-163.

SOUTHERN, W.F. and L.K SOUTHERN. 1981. Colony

census results as indicators of pre-hatching perturba-
tions. Colonial Waterbirds 4:143-149-

SOWLS, A.L., A.R. DE GANGE, J.W. NELSON, and GS. LES-
TER. 1980. Catalog of California seabird colonies. U.S.
Dept. Inter., Fish Wildl. Serv., Biol. Serv. Prog., FWS/
OBS-80/37.

, S.M. HATCH, and C.J. LENSINK 1978. Catalog of

Alaskan seabird colonies. U.S. Dept. Inter., Fish Wildl.
Serv., Biol. Serv. Prog., FWS/OBS-78/78.

SPEICH, S.M. and DA. MANUWAL. 1974. Gular pouch
development and population structure of Cassin's
auklet. Auk 91:291-306.

and T.R. WAHL. 1985. Catalog of Washington sea-
bird colonies. U.S. Dept. Inter., Fish Wildl. Serv., Biol.
Serv. Prog., FWS/OBS. In press.

STOWE, J.J. 1982. Recent population trends in cliff-breed-
ing seabirds in Britain and Ireland. Ibis 124:502-510.

THOMPSON, B.C. and J.E. TABOR. 1981. Nesting popula-
tions and breeding chronologies of gulls, terns, and
herons on the upper Columbia River, Oregon and
Washington. Northwest Sci. 55:209-218.

THOMPSON, L.S. 1981. Nest-tree sharing by herons and
cormorants in Montana. Can. Field-Naturalist
95:257-260.

. 1982. A 1979 census of Great Blue Heron colonies

in Montana. Proc Montana Acad. Sci. 41:23-27.

THOMPSON, S.P., CD. LITTLEFIELD, and R.A. RYDER.

1979. Historical review and status of colonial nesting
birds on Malheur National Wildlife Refuge, Oregon.
Proc. Colonial Waterbird Group 3:156-164.

THORESEN, AC. 1964. The breeding behavior of the
Cassin's auklet. Condor 66:456-476.

TREMBLAY, J. and L.N. ELLISON. 1979. Effects of human
disturbance on breeding of black-crowned night her-
ons. Auk 96:364-369.

TUCK, L.K 1961. The murres. Can. Wildl. Serv.
Monogr. 1.

UDVARDY, M.D.F. 1969. Dynamic zoogeography: with
special reference to land animals. Van Nostrand
Rheinhold Co., New York. 445pp

VAN TETS, G.F. 1965. A comparative study of some social
communication patterns in the Pelecaniformes. Am.
Ornith. Union, Ornith. Monogr. 2.

VERBEEK, N.A.M. 1982. Egg predation by northwest
crows: Its association with human and bald eagle
activity. Auk 99:347-352.

VERMEER, K 1963. The breeding ecology of the glau-
cous-winged gull (Larus occidentalism on Mandarte
Island, B.C. British Columbia Province Mus., Occ Pap.
13.

WEISE, J.H. 1978. Heron nest-site selection and its ecolog-
ical effects. Pages 27-34 in Sprunt, AS., IV, J.C. Og-
den, and S. Winckler, eds. Wading Birds. Natl.
Audubon Soc, Res. Rep. 7.

WERSCHKUL, D.F., E. MC MAHON, and M. LEITSCHUH.
1976. Some effects of human activities on the great
blue heron in Oregon. Wilson Bull. 88:660-662.

, , , S. ENGLISH, C. SKIBINSKI, and G.

WILLIAMSON. 1977. Observations on the reproduc-
tive ecology of the great blue heron {Ardea herodias)
in western Oregon. Murrelet 58:7- 12.

Colonial Waterbirds

405

20

Upland Game
Birds

Robert L. Eng

Biology Department
Montana State University
Bozeman, MT 59717-0001

Editor's Note: Upland game birds, like other eco-
nomically important species, have received much
attention. And even though most of these species
are avidly hunted, in most cases habitat quantity
and quality is the factor controlling abundance
and distribution. Although much research has been
conducted on upland game bird habitat, much of
this has focused on species associated with agricul-
ture such as pheasants, rather than on wildland
species. For example, the effects of logging grazing
and recreation on species of grouse are poorly
understood Therefore enlightened management of
habitat for these species will require better identifi-
cation of limiting habitat factors, together with
careful inventory and monitoring of habitats and
populations.

INTRODUCTION

Usually a reference to upland game birds is an
exclusive reference to members of the family Phas-
ianidae: non-migratory, chicken-like birds including
the partridge, grouse, turkey, and quail. Recently,
increasing mention may be found to migratory up-
land game birds in reference to members of the fam-
ily Columbidae, the pigeons and doves.

Emphasis here will be placed on members of
the family Phasianidae, primarily because their non-
migratory status requires a more labor-intensive in-
ventory and assessment of year-long habitat needs.
Also, within this group, emphasis will be placed on
those members in the western U.S. that inhabit non-
cultivated lands, at least for a major portion of the
year.

Although life histories or the biology of each
species will not be dealt with here, some basic
knowledge of a species' life history is necessary be-
fore sound management decisions can be made.

This does not imply the need for an in-depth
study of each species. Rather, many species have
certain habits similar enough to permit the use of
common denominators in monitoring the birds or
their habitat. For example, most quail and partridge
feed heavily on seeds from a variety of annuals that
may show wide year-to-year fluctuations and are
often products of an early serai stage. These birds
have evolved with a very high reproductive poten-
tial, including a high rate of first-year breeding and
large clutch sizes, and thus can quickly respond to a
bountiful food supply. Conversely, blue grouse (Den-
dragapus obscurus) and sage grouse (Centrocercus
urophasianus) feed during the winter on the vegeta-
tive parts of a few species of dominant plants, a food
supply that is produced in a climax or near climax
stage, and shows comparatively little annual variation
in abundance. These birds have a relatively low re-
productive potential with respect to age of consist-

Upland Game Birds

407

ent first breeding and clutch size; this appears to
be in keeping with their more stable food supply.
The contrasts in the two facets of life history of
these two groups would result in equally contrasting
habitats and habitat management concerns.

That there is a close relationship between the
quality of a bird's habitat and its welfare (and result-
ing population status) requires little explanation. The
difficulty arises in quantifying the quality of a habitat
which for a given species may be found in a contin-
uum from marginal to excellent. Available in the
literature are qualitative descriptions of game bird
habitats listing food and cover components which
may be preferred or required by the species in ques-
tion. When measurements are taken of the amount
and distribution of these components within a unit
area, the description assumes a quantitative air. Fi-
nally, with the assistance of a series of assumptions,
numbers are assigned to the descriptions and meas-
urements, and a numerical rating is assigned placing
that particular piece of habitat at a fixed point along
the continuum. However, the quality of the overall
habitat, which may rate high as a composite, may in
reality rate rather low as a result of poor quality of
one facet. For example, two areas, one a homogene-
ous stand of smooth brome and the other 85%
smooth brome and 15% forbs, may both be re-
corded as ungrazed, unmowed pastures of smooth
brome. However, the field with a sprinkling of forbs
would provide better cover and a greater diversity of
insects, making it a superior habitat for brood
rearing.

The above is not intended to entirely discredit
efforts at measuring the quality of game bird habitats.
Certainly, the available descriptions provide land
managers with the capability to recognize the gen-
eral habitat type in which they could expect to find
the various species. They could also, with some ex-
perience, place an assessment of quality within a
broad scale. However, a more realistic measure of
the quality can be attained through a measure of the
bird's response to its habitat, the home range size,
or the population density.

The average size of a home range, whether sea-
sonal or annual, when compared between habitats,
will reflect the quality of a given habitat. Birds in
high quality, secure habitat tend to have smaller
home ranges than those in habitats of lesser quality.
Unfortunately, home range determination is time and
budget consuming, usually involving trapping, mark-
ing, and relocating individual birds.

Census techniques, whereby relative densities
can be determined, present a far more workable
solution. Most techniques permit only an index to
the population; but if conducted by standardized

methods, data between areas or time periods provide
workable comparisons. For the land managers inter-
ested in habitat quality, population density compari-
sons between areas provide a measure of the various
areas to support birds of that species. Similarly, com-
parable data gathered from a single area over a pe-
riod of several years may provide a measure of the
population trend which in turn could reflect changes
in habitat quality.

HABITAT FEATURES CORRELATED WITH
SPECIES GROUPS

Major habitat characteristics of the various spe-
cies are presented in summary form with selected
references in Table 1. Supplemental statements to as-
sist in habitat identification are presented in the text.

Gray Partridge (Perdix perdix)

The European gray or Hungarian partridge is
one of three exotic game birds successfully intro-
duced into the U.S. The others are the ring-necked
pheasant (Phasianus colcNcus) and the chukar
(Alectoris chukar). The "nun" is native to Europe
and Asia although North American stock appears to
have originated largely in central Europe.

Although present in a rather wide variety of
topographic conditions, the more dense populations
of gray partridges are generally associated with flat
or moderately rolling land. Stiehl (1984) referred to
three geographic populations in North America:
western, central, and the Great Lakes.

Gray partridge are frequently associated with a
combination of cultivated and non-cultivated lands.
Weigand (1980) reported land uses from his study
area, in decreasing order of abundance, as small
grain and fallow land, rangeland, hayfields, and agri-
culturally idle areas. Small grain fields and fallow
lands were used heavily for feeding (both seeds and
green vegetation) during a large part of the year
whereas the other types provided an herbaceous
cover primarily for loafing, roosting, and nesting.
Their affinity for edges probably accounts for their
great ability to successfully cope with modern agri-
culture (Allen 1984).

Weigand (1980) found wheat and barley seeds
and leaves a major food component throughout the
year; seeds and leaves of such forbs as dandelion
(Taraxacum officinale), knotweeds (Polygonum sp.),
mustard ( Brassica kaber), and white clover ( Trifol-
ium repens) were prominent at various seasons. He
also reported that of 10 nests located, 5 were con-
cealed by grass, the others by forbs.

408

Upland Game Birds

Table 1. Synopsis of habitat features for species or species groups.

Species or
Species Group

Habitat Characteristics

(Physical features, vegetation, and species composition)

References

Gray partridge

(Perdix perdix)

Agricultural lands, primarily small grains, interspersed with idle
areas Croplands provide major food source and idle areas pro-
vide loafing, roosting, and nesting cover. Nesting cover often
in grassland/shrub areas — successful nesting efforts dependent
on residual cover from the previous year. Seeds from cultivated
grains and a variety of forbs and other green succulent vegeta-
tion provide primary food

Johnsgard 1973, Wei-
gand 1980; Smith et al.
1982; Mendel and Peter-
son 1983; Potts 1984;
Stiehl 1984

Chukar

(Alectoris chukar)

Rugged, rocky terrain in arid to semi-arid climates, about equally
divided between rocky outcrops and low shrub/grassland. Low
winter temperatures are less devastating than persistent snow
cover. Water availability may significantly affect summer distribu-
tion. Rocky outcrops may provide loafing and roosting cover.
Nesting cover similar to shrub/grassland types used for foraging,
often on a south facing slope up from creek bottoms Sagebrush
(Artemisia sp), cheatgrass (Bromus tectorum), and wheatgrass
(Agropyron sp ) are common species with cheatgrass seeds and
leaves providing a major food source.

Galbreath and Moreland
1953; Bohl 1957; Harper
et al. 1958; Christensen
1970; Molini 1976.

Ring-necked pheasant
(Phasianus colchicus)

Agricultural land highly interspersed with non-cultivated land (i.e.,
dense clumps of grasses, forbs, and shrubs). Non-cultivated
areas (preferably idle areas such as roadsides, railroad rights-of-
way, and temporary wetlands) are necessary for nesting, roosting,
and loafing cover, particularly during winter and early spring
when crops are removed. Farm crops, primarily waste cereal
grains, may account for over 75% of the food with weed seeds
associated with cultivation making up another 10-15%.

Hiatt 1947; Yeager et al.
1951; Snyder 1974;
Weigand and Janson
1976.

Spruce grouse

(Dendragapus
canadensis)

Coniferous forests of Canada and northern U.S. A trend toward
increased use of upland over lowland areas from eastern to
western distribution. Preference shown for coniferous stands that
are neither mature nor extremely dense, thus permitting adequate
understory for food and cover. Spruce (Picea sp.) and jackpine
(Pinus banksiana) in Michigan or lodgepole pine (Pinus contorta)
in the mountain west are key to the species. A variety of berries
(Vaccinium sp., Symphoncarpus sp., Rubus sp.) are important
summer foods whereas coniferous needles (Picea, Pinus, Larix)
dominate the winter diet.

Stoneberg 1967; Pender-
gast and Boag 1970;
Robinson 1980; Ratti et
al. 1984.

Blue grouse

(Dendragapus obscurus)

Coniferous forests of western and coastal mountain ranges Win-
ter at high elevations in stands of conifers; in spring, found at
lower elevations at forest-grassland ecotone, while females with
broods may follow brushy draws 2 to 3 miles (3 to 5 km) below
conifers. Prefer open, dry sites but near clumps of conifers for
escape cover. Structural height of vegetation at least 8 inches (20
cm) preferred in brood rearing areas. Douglas fir (Pseudotsuga
menziesn) a key cover and food plant (winter). Berries (Vaccinium
sp., Symphoncarpos sp.) and leafy forage (Trlfolium sp., Fragana
sp.) are heavily used at other times of the year.

Mussehl 1963; Bendell
and Elliott 1966; Martinka
1972.

Ptarmigan
Willow

(Lagopus lagopus)

Rock

(Lagopus mutus)

White-tailed

(Lagopus leucurus)

Willow ptarmigan inhabit open tundra and show a preference for
areas with level to moderate terrain, heavily vegetated with
grasses, forbs, and shrubs. Rock ptarmigan frequent the open
tundra but on more hilly terrain and with less luxuriant vegetation
than utilized by the willow ptarmigan. White-tailed ptarmigan
occupy alpine tundra, ridges, and steep slopes where vegetation
is sparse and dwarfed; rocks provide the structural cover. Willow
(Salix sp.) is a regular shrub for all three species, the buds and
twigs being dominant food for a large part of the year. Also used
are a variety of leaves, flowers, and berries as they become
available during a relatively short growing season.

Choate 1963; Weeden
1963, 1965, 1967; Johns-
gard 1973.

Upland Game Birds

409

Table 1. Synopsis of habitat features for species or species groups (continued).

Species or
Species Group

Habitat Characteristics

(Physical features, vegetation, and species composition)

References

Ruffed grouse

(Bonasa umbellus)

Deciduous forests or mixed deciduous/coniferous forest edges. In
the Mountain West, this species is more prevalent along stream
courses or other moist areas. Drumming males prefer areas of
high vegetative density (overhead cover important) whereas
broods prefer more open stands but with high degrees of herba-
ceous cover. Aspen (Populus tremuloides or P. grandidentata)
key species in preferred stands A preference shown for early
serai stages of aspen although an interspersion of different age
classes desirable for yearlong food source (i.e., aspen buds from
mature stands, berries, fruits, and leafy forage from younger
stands).

Svoboda and Gullion
1972; Gullion 1977, 1984;
Stauffer and Peterson
1985, 1985b; Kubisiak
1985.

Sage grouse

(Cenfrocercus
urophasianus)

Semi:arid sagebrush or sagebrush/grassland types of the western
plains and intermountain basins at elevations from 750 m (2,500
ft) to 2,100 m (7,000 ft) Preference for gentle topography over
steep-sided canyons or slopes. Sagebrush is essential to survival
with stands of greatest canopy coverage (20-35%) preferred in
winter and least (5-10%) in summer. Food is reflected in this
seasonal cover preference: sagebrush in winter, a variety of suc-
culent forbs in spring and summer.

Patterson 1952; Eng and
Schladweiler 1972; Wal-
lestad 1975; Wallestad et
al. 1975.

Greater prairie chicken

(Tympanuchus cupido)

Arid native grasslands of the Southwest intermixed with shrub or
half-shrub component. A portion (5-40%) of an area in cultivation,
if well interspersed, may enhance the habitat through increased
food source. Shrubs, less than 1 m (3 ft) tall in savannah aspect,
are desirable although shrub species may determine preferred
shrub density; high densities of sand sagebrush (Artemisia fili-
folia) appear more desirable than high densities of shinnery oak
(Quercus havardii).

Jones 1963; Crawford
and Bolen 1976; Davis et
al. 1980; Cannon and
Knopf 1981a; Doerr and
Guthery 1983.

Sharp-tailed grouse

(Tympanuchus
phasianellus)

Grasslands that have a prominent deciduous shrub or woodland
component. Although both western races (T. p. columbianus
and jamesi) inhabit a grassland/shrub/tree complex, the Colum-
bian sharptail is usually found in habitat with a higher overall
shrub/tree cover. Columbian sharptails are associated with a vari-
ety of shrubs and trees (sagebrush, hawthorne [Crataegus sp.],
serviceberry [Amelanchier sp], Gambel oak [Quercus gambellii],
and aspen) which may be used for cover and produce food in
the form of berries or winter buds. A similar group of plants exists
for the prairie race (juniper [Juniperus sp.], buffaloberry [Shep-
herdia sp.]), cottonwood, and aspen.

Miller and Graul 1980;
Nielson and Yde 1981.

Wild turkey

{Meleagns gallopavo)

Open, mature, or nearly mature forests, deciduous but often
deciduous-coniferous, particularly in mountainous areas. Mature
forests provide a more dependable mast crop and permit greater
visibility. Merriam's turkey shows a preference during spring,
summer, and fall for ponderosa (Pinus ponderosa) forests with an
open grassland understory; in winter it often frequents a lower
elevation in a mast producing habitat or in the north, grain fields.
The Rio Grande race, found in more arid areas to the south, often
occupies river bottom habitat, but also oak-grassland intersper-
sions. Scattered openings in forest habitat (10-40% of the total)
are highly desirable. Mast producers such as oak (Quercus sp.),
pinyon pine (Pinus edulis), and juniper and seeds and leaves
from a variety of grasses are major food sources.

Ligon 1946; Spicer 1959;
Hoffman 1962; Jonas
1966; Beasom 1970.

Mearn's quail
(Cyrtonyx montezuma
mearnsi)

Oak-grasslands and pine-oak woodlands ranging from 1,100 to
3,600 m (4,000 to 12,000 ft) in the highlands of the Southwest.
Preference is shown for open woodlands with a grass-forb un-
derstory. Grass is used commonly for cover; the bulbs and seeds
of woodsorrel (Oxalis) and flat sedges (Cyperus) are prominent
foods. During years of scarcity of forbs and sedges, acorns may
be a common food.

Leopold and McCabe
1957; Ligon 1961; Bishop
and Hungerford 1965;
Brown 1979; Brown 1982

410

Upland Game Birds

Table 1. Synopsis of habitat features for species or species groups (concluded).

Species or
Species Group

Habitat Characteristics

(Physical features, vegetation, and species composition)

References

Masked bobwhite

(Colinus virginianus
ndgwayi)

Open desert grasslands at elevations from 300 to 1 ,200 m (1 ,000
to 4,000 ft). Although mesquite (Prosopsis sp.) and catclaw (Aca-
cia sp.) are frequently found in the area, this bird seemingly
prefers the grassy areas adjacent to shrubs. Coarse grasses
(Sporobolus sp.) provide cover whereas species of grama (Bou-
teloua sp.) and three-awns (Anstida sp.) provide both food and
cover.

Tomlinson 1972a, b;
Johnsgard 1973.

Scaled quail

{Callipepla squamata)

Desert grassland, usually below 2,100 m (6,900 ft), composed of
low growing grasses, forbs, and shrubs with an overall ground
cover between 10 and 50%. Shrubs used for overhead cover with
preference shown for those less than 2 m (7 ft). As little as 10%
of an area left in clumps of mesquite will provide cover and food
(seeds). Major vegetative foods consist of conspicuous forb
and shrub- seeds such as croton (Croton sp.), cycloloma (Cyclo-
loma atnplicifolium), snakeweed (Gutierrezia sarothrae) and
mesquite.

Campbell et al. 1973,
Davis et al. 1975; Good-
win and Hungerford
1977, Ault and Stormer
1983

Gambel's quail

(Callipepla gambelii)

Desert shrub type usually in habitats below 1,800 m (6,000 ft)
Preference is shown for shrub cover of sufficient density to shade
50 to 75% of the ground and where 60 to 80% of the shrubs are
taller than 2 m. Mesquite, hackberry (Celtis reticulata), wolfberry
(Lycium sp), and catclaw provide cover for most of the bird's
activities Many preferred broods are derived from members of
the Leguminosae, i.e., mesquite, catclaw, mimosa (Mimosa sp.),
deervetch (Lofus sp.), paloverde (Cercidium sp.), and lupine
(Lupinus sp.)

Gullion 1960; Hungerford
1962; Goodwin and
Hungerford 1977

California quail

(Callipepla
calif or nica)

May occur in desert, rangeland, and dry and irrigated farmland at
elevations from sea level to 2,600 m (8,500 ft). Rangeland with
the following components are most widely used: clumps of woody
vegetation for roosting, shrubs or herbaceous growth for escape
cover, and herbaceous cover for nesting. Food habit studies
repeatedly show importance of legumes (esp. seeds). Some im-
portant legumes include the following: bur clover (Medicago sp),
lupines, deervetches, clover (Trifolium sp), and vetches (Vicia
sp.). Filaree (Erodium sp.) leaves and seeds occur frequently in
the diet.

Emlen and Glading 1945;
Edminster 1954; Shields
and Duncan 1966; Johns-
gard 1973; Gutierrez
1980.

Band-tailed pigeon
Coastal race

(Columba fasciata
monilis)
Interior race
(C. f. fasciata)

The coastal race, although primarily an inhabitant of mountain
habitat, can be found from sea level to 4,200 m (13,850 ft). The
race may occupy a variety of habitats, but the preferred type
consists of large conifers and deciduous trees interspersed with
berry and mast producing trees and shrubs. The interior race also
shows preference for mountainous areas supporting a mixed
conifer-deciduous type and achieves greatest densities in this
type at elevations between 1 ,675 m (5,500 ft) and 2,575 m (8,500
ft). Both races feed heavily on mast produced by species of
pine and oak as well as a variety of berries produced by such
shrubs as elderberries (Sambucus sp.), wild cherries (Prunus
sp.), huckleberries, and dogwood (Cornus sp). During certain
years, both races will move to lower elevations and feed on culti-
vated grain.

Braun et al. 1975; Jeffrey
1977; Tomlinson 1983.

Mourning dove

(Zenaida macroura)

The wide breeding distribution of this species almost precludes
describing habitat features with precision. It is primarily an inhab-
itant of woodland-grassland edge Thus, both clearing of large
areas of forest in the East and planting of shelterbelts in the
Plains States enhanced mourning dove habitat. Although tree
nesting is most common, in the absence of trees or shrubs, the
nests are readily placed on the ground Food supplies for doves
have been improved by species that produce more seeds than
native grasses.

Keeler 1977; Dunks et al
1982

Upland Game Birds

411

Chukar (Alectoris chukar)

The chukar has been widely introduced into the
U.S., but was successfully established only in arid
areas of the West. Seven species of this genus have
been recognized, all of which are native to parts
of Europe or Asia. It is believed that the only species
that have been successful in the U.S. were those
originating in India (Christensen 1970).

Chukars characteristically inhabit very rugged,
rocky terrain in arid to semiarid climates. They have
been found from sea level to 4,800 m (16,000 ft)
in their native Asian habitat and to 3,600 m ( 1 2,000
ft) in North America. The Great Basin habitat of the
West with its combination of mountain range-valley
(basin) floor provides excellent conditions for this
exotic species (Christensen 1970).

Chukars frequently use the rugged topography
for cover, although they are also found in a shrub-
grassland type. The understory may consist of an
abundant grass-forb combination of the desert. Sage-
brush (Artemisia tridentata) is the dominant shrub in
chukar habitat in Nevada (Christensen 1970) and
probably over most of its range in the U.S. Other
structurally similar shrubs such as greasewood (Sar-
cobatus sp.), rabbitbrush (Chrysothamnus sp.), and
bitterbrush (Purshia tridentata) may be found in-
stead of sagebrush in different parts of this bird's U.S.
distribution. A food source is largely derived from
an understory of grasses and forbs although grasses
frequently provide the major portion. Christensen
(1970) reported heavy use of seeds and leaves of
cheatgrass (Bromus tectorum) and believed that the
introduction of this exotic grass into the western
U.S. may have greatly influenced the later success of
the chukar introduction.

within and adjacent to agricultural lands, the preser-
vation of grass/forb/shrub cover is frequently the
key to maintaining a healthy pheasant population.

Although the pheasant is primarily associated
with agricultural lands, a few populations exist in the
western U.S. completely removed from cultivation.
In such areas, usually a drainage supporting some
woody cover, the birds substitute weed seeds and
fruits from shrubs such as buffaloberry (Shepherdia
sp.), rose (Rosa sp.), and snowberry (Symphoricar-
pos sp. ) for the cultivated grains that usually make
up a large part of the diet. In such areas, pheasant
densities are much lower than in comparable cover
adjacent to agriculture, but the existence of even the
low numbers is highly dependent upon the shrubby
(protective) cover.

Ring-necked pheasant.

Ring-necked Pheasant (Phasianus
colchicus)

Ring-necked pheasants consist of a group of
subspecies from Asia that were successfully intro-
duced into the U.S. in the 1880s. The habitats of this
group of birds in Asia were similar to those found
in the midwestern and western U.S., and the birds
became established in many of these areas as a very
popular upland game bird.

The pheasant is closely allied with agricultural
lands. In the West, areas placed under irrigation ini-
tially created islands of excellent pheasant habitat,
where as little as 5% of the area in permanent, well
interspersed cover provided adequate nesting, brood
rearing, and winter cover (Yeager et al. 1951). As
agriculture became more intensive, land use changes
(largely the removal of permanent cover) reduced
the potential for producing pheasants (Baxter and
Wolfe 1972; Snyder 1974; Taylor et al. 1978). Thus,

Spruce Grouse (Dendragapus canadansis)

The western representative of this small timber
grouse, for many years considered a separate species,
is presently regarded as a race — the Franklin's spruce
grouse. Among the many local names applied to this
bird, the most common is undoubtedly the fool hen,
a direct result of its lack of fear of man.

Spruce grouse are distributed across the whole
of Canada and portions of the northern U.S., closely
coinciding with the distribution of the coniferous
forest. Robinson (1980), in reviewing literature on
habitat for this bird, suggested a trend toward in-
creased use of upland over lowland areas in its distri-
bution from east to west. The most dense forests
are seemingly avoided because they do not provide
an adequate ground cover and openings large
enough for territorial display flights (Stoneberg 1967;
Robinson 1980).

412

Upland Game Birds

Although early descriptions of spruce grouse
habitat emphasize the mature coniferous forest with
species of spruce being the primary ingredient, more
recent studies have shown that habitat often consists
of mixed stands of spruces (Picea sp.) and jack pine
(Pinus banksiana) in Michigan (Robinson 1980) and
spruces and lodgepole pine (Pinus contorta) in west-
ern States (Stoneberg 1967; Ratti et al. 1984).

Spruce grouse.

Blue Grouse (Dendragapus obscurus)

The blue grouse is the largest of the North
American forest grouse, the males commonly weigh-
ing 1,200-1,250 g (2.6-2.7 lb) or about twice the
weight of spruce or ruffed grouse. The subspecies
(D. fuliquinosus) inhabiting the coastal mountain
ranges is often referred to as the sooty grouse
whereas the one in the Rocky Mountain States is
often called the dusky grouse.

Blue grouse often participate in a seasonal, alti-
tudinal migration. In winter, they are most fre-
quently found at high elevations in dense stands of
coniferous forest. In spring, a downward migration
occurs into a forest-grassland ecotone where males
establish territories. Females may brood in open
meadows at this elevation, or in some parts of their
range, will follow small brushy draws out into prairie
foothill types. In Montana and Colorado, sharp-tailed
grouse ( Tympanuchus phasianellus), sage grouse,
and blue grouse have been observed on common
brood-rearing ranges in late summer and early fall
(pers. obs.).

Although an inhabitant of forested, mountain
areas, blue grouse appear to prefer meadows and
grassland adjacent to woody cover (Bendell and
Elliott 1966). Breeding habitat is frequently an open,
dry site with scattered shrubs but near dense clumps
of trees for escape cover (Martinka 1972). Brood

rearing often occurs in meadows or an open stand of
timber with a good ground cover of grass and forbs.
Mussehl (1963) listed the following basic physical
requirements of blue grouse brood range: a relatively
high degree of canopy coverage, an effective height
of ground cover of about 20 cm (8 in.), which is
an important consideration in areas grazed by do-
mestic livestock, diversity of plant life forms, and
small amounts of bare ground. Winter habitat, which
usually includes a more dense canopy of conifers
for a food source, is nonetheless often in a parkland
type or along open ridges.

In overall distribution, blue grouse seem closely
associated with Douglas fir (Pseudotsuga menziesii).
Needles from this species are an important winter
food source although others (Larix sp., Pinus sp., and
Abies sp. ) have been reported as common in the
winter diet.

Blue grouse.

In breeding and brood rearing habitat, life form
is probably as critical as species composition. Mar-
tinka ( 1972) pointed out the importance of Douglas
fir thickets as escape cover for males displaying in
open logged-over stands of ponderosa pine (Pinus
pondcrosa). Mussehl (1963), in discussing brood
habitat, described the bunchgrass-balsamroot open-
ings at the edge of coniferous stands in one study
area and in another, a similar grass-forb association
in an open stand of mature ponderosa pine. In years
when balsamroot was not abundant, broods were
found in association with snowberry or large sedges
(Carex sp.). In late summer, before moving to higher
elevations and a conifer diet, blue grouse in the
Rocky Mountains feed heavily on berries such as
currants (Ribes sp.), juneberries (Amelanchier sp.),
bearberry (Arctostaphylos sp.), and huckleberry
(Vaccinium sp.) (Beer 1943).

Upland Game Birds

413

Willow Ptarmigan (Lagopus lagopus)
Rock Ptarmigan (I. mutus)
White-tailed Ptarmigan (I. leucurus)

Although considerable overlap exists in the
North American geographical distribution of these
three species, particularly so with willow and rock
ptarmigan, differences in seasonal habitat selection
assist in maintaining them as separate species. Be-
cause willow (Salix sp.) is a shrub group commonly
used by all three species, the major difference in
habitat characteristics is structural. From willow to
rock to white-tailed ptarmigan, the trend is toward
shorter and less dense vegetation and increased top-
ographic relief. Thus during the spring and summer,
the willow ptarmigan shows a preference for level
ground or gentle to moderate slopes, luxuriant plant
growth with shrub height usually 1-2.5 m (3-8 ft),
and an elevation at the upper edge of timberline
among widely scattered trees. The rock ptarmigan
shows a preference for moderate slopes in hilly boul-
der-strewn country, vegetative cover complete but
sparse with shrubs (in ravines) 0.3 to 1.4 m (1-4 ft),
and most are found at an elevation of 30-300 m
(100-1,000 ft) above timberline. The white-tailed
ptarmigan prefers steep slopes with rocky outcrops,
with plant cover rarely continuous, shrubs sparse
and in a dwarf form, and are usually found 1 50-600
m (500-2,000 ft) above timberline (Weeden 1965).

All three species rely heavily on willow for food.
Dominant winter foods eaten by the willow ptarmi-
gan are willow buds and twigs; dwarf birch (Betula
nana) buds and catkins are second in importance.
Rock ptarmigan eat the same species but in reverse
order of prominence. During spring, summer, and
fall, both species feed commonly on leaves and ber-
ries of cranberry and blueberry, horsetail (Equisetum
sp.) tips, leaves of mountain avens (Dryas octope-
tala), and crowberries (Empetrum nigrum) (Weeden
1965). White-tailed ptarmigan in the Rocky Moun-
tain States (primarily Colorado) rely heavily on al-
pine willows (Salix nivalis and S. arctica), and the
distribution and abundance of these plants dictate
the distribution of this ptarmigan (Braun 1970).

Ruffed Grouse (Bonasa utnbellus)

Ruffed grouse, the most widely distributed
grouse in North America, does not have the same
popularity as a game bird in the western U.S. that it
has in the Lake States and farther east. This distribu-
tion of popularity is also reflected by the distribution
of ruffed grouse research conducted to date.

Ruffed grouse are closely associated with decid-
uous forests or mixed deciduous-coniferous forest
edges. This permits a more continuous distribution
in the Lake States where large expanses of hardwood
forest are present. In the mountain West, ruffed
grouse are more prevalent along (but not entirely

confined to) stream courses or moist areas where
deciduous trees and shrubs are found. During the
winter in mountain habitat, the birds prefer a south-
ern exposure and elevations above the stream bot-
tom (Stauffer 1983), both related to higher
temperatures, snow melt, and resulting food
availability.

The plant most often associated with thriving
ruffed grouse populations is quaking aspen (Populus
tremuloides). Several authors (Svoboda and Gullion
1972; Doerr et al. 1974) have suggested that aspen
buds (especially flower buds from mature male as-
pen) are critical winter and spring foods. Gullion
(1977), in discussing forest management for ruffed
grouse, pointed out the value of each different age
aspen stand to a particular seasonal ruffed grouse
requirement. In the West, fire suppression in many
areas has probably prevented regeneration of com-
plete stands of aspen, and early serai stages are few
or non-existent (Stauffer and Peterson 1985b).

Ruffed grouse.

In many areas of the mountain West, aspen is
limited in distribution and seldom in the large con-
tinuous stands so common to the Lake States. Under
these conditions, ruffed grouse rely on a variety of
understory shrubs for food and cover such as willow,
serviceberry ( Amelanchier sp.), blueberry, haw-
thorne (Crataegus sp.), chokecherry (Prunus sp.),
and snowberry.

Sage Grouse (Centrocercus urophasianus)

The sage grouse is decidedly the largest grouse
in North America; an adult male may weigh 3,175 g
(7 lb ), about twice the size of the next largest, the
blue grouse. Unlike many other North American
grouse, sexual dimorphism is pronounced in the sage
grouse in both size and plumage.

414

Upland Game Birds

Sage grouse are characteristic of the semi-arid
sagebrush or sagebrush-grassland types found in the
western plains and intermountain basins. Generally
following the distribution of the plant genus Artemi-
sia, sage grouse can be found at elevations from
about 750 m (2,500 ft) in the western Dakotas to
over 2,100 m (7,000 ft) in some intermountain
basins.

Sagebrush is essential to sage grouse survival. A
proper combination of height and canopy of this
plant must be available to meet the birds' seasonal
needs. Dependency on the plant is greatest during
the late fall, winter, and early spring. A sage grouse
food habits study in Montana (Wallestad et al. 1975)
showed that during 8 months of the year. (October
through May ), the frequency and volume of big sage
leaves found in crops remained above 60% and 90% ,
respectively. For three of these months (December
through February) it was the only food present. Thus
for a large part of the year, sage grouse are found in
or near stands of sagebrush with a canopy-coverage
exceeding 20% (Eng and Schladweiler 1972; Walles-
tad and Schladweiler 1974). In addition to the sage-
brush canopy coverage requirement, winter ranges
are frequently on areas with little or no slope (Eng
and Schladweiler 1972; Beck 1977), although Beck
(1977) reported some use of slopes exceeding 10%
and also use of windswept ridges for feeding sites.

During the 4 months of summer and early fall,
sage grouse seek out succulent forbs, and frequently
leave the dense sage and move to scattered sage
distribution with a 1-10 and 10-25% canopy cover-
age (Wallestad 1975). Sage grouse at this time are
often found near seeps, streams, or irrigated fields
where succulent forbs are available. By mid Septem-
ber to late September, frequently coinciding with
regular heavy frosts, the birds move toward the win-
tering areas. This movement between summer and

winter ranges may be as little as 3-5 km (1.8-3 mi.)
(Eng and Schladweiler 1972) or up to 80 km
(48 mi.) (Dalke et al. 1963).

Although big sage is the species most frequently
associated with sage grouse, several others (Artemi-
sia cana, A. frigida, A. ludoviciana, A. nova) have
been listed as used for both food and cover (Patter-
son 1952; Wallestad et al. 1975). Species of forbs
that are prominent in the diet of sage grouse are
dandelion, prickly lettuce (Lactuca sp.), salsify (Tra-
gopogon sp.), gumweed (Grindelia sp.), yarrow
(Achillea sp.), and sweet clover (Melilotus sp.). Al-
falfa (Medicago sp.) is frequently heavily used in late
summer and early fall.

Prairie Chicken (Tytnpanuchus cupido)

This species, also called the pinnated grouse,
was once well represented throughout the central
U.S. from the southern prairie provinces of Canada to
the Gulf Coast, coinciding closely with the distribu-
tion of mid- and tallgrass prairie. The 1957 American
Ornithological Union (A.O.U.) checklist recognized
the lesser and greater prairie chicken as separate
species; the 1983 version lists the two as a superspe-
cies although many consider them to be a single
species.

The lesser prairie chicken is characteristic of the
more arid grasslands of the Southwest, whereas the
greater is associated with the more moist tallgrass
prairie of the eastern great plains (Johnsgard 1973).
Both races appeared to have benefited from the
breaking of the prairies for cultivation, providing
winter food in the form of grain, and permitted a
northward range extension (Johnsgard and Wood
1968). This extension was relatively short-lived for
both races and as the agricultural effort became
more intense, distribution receded to the remnant
status of today (Johnsgard 1973).

€§y^i^'i '^%'%f> ■^^vj%W"»*«**»<i

m

Sage grouse droppings.

Sage grouse at water hole.

Upland Game Birds

415

A very explicit comparison of habitat for the
lesser and greater prairie chickens in Oklahoma was
presented by Jones (1963). He stated "... the habi-
tat of the lesser prairie chicken consisted of small
units of shortgrass prairie intermixed with larger
units of shrub or half-shrub vegetation; that of the
greater prairie chicken consisted of small units of
shortgrasses or midgrasses intermixed with larger
units of tall grasses." This difference was also shown
in day-resting activities with the lesser using primar-
ily half-shrub vegetation whereas the greater favored
the edges of tallgrass and midgrass vegetation. Both
tended to rear broods in areas of abundant forb
growth, at least in part for the insect fauna in associ-
ation with such areas.

The emphasis of shrub component in the lesser
prairie chicken habitat is more pronounced than
with the greater prairie chicken. The half-shrub (Ar-
temisia filifolia) is common in Oklahoma and New
Mexico (Jones 1963). In New Mexico, the shinnery
oak ( Quercus havardii) is heavily used; the habitat
consists of a shinnery oak-tallgrass community which
included such grasses as sand bluestem (Andropogon
hallii), little bluestem (Andropogon scoparius),
three-awn (Aristida sp.), dropseed (Sporobolus sp. ),
and hairy grama (Bouteloua hirsuta) in addition to
a variety of forbs (Davis et al. 1980). However, in
Oklahoma, densities of displaying males were nega-
tively correlated with percentage coverage of brush
and positively correlated with percentage coverage
of grass in a shinnery oak rangeland; the reverse was
found on a sand sagebrush rangeland (Cannon and
Knopf 1981a).

Sharp-tailed grouse (Tympanuchus
phasian ellus)

Six races of sharp-tailed grouse are recognized
(Hamerstrom and Hamerstrom 1961), two of which
extend their ranges south of Canada in the western
U.S. The Columbian sharptail ( T. p. columbianus)
is found in western Montana and eastern Washington
south to Utah and western Colorado. The plains race
( T. p. jamesi) is the most abundant and widespread
of the races in the U.S., being found in eastern Mon-
tana, Wyoming, and Colorado, and in western Ne-
braska and the Dakotas.

Sharp-tailed grouse prefer habitat with structural
features intermediate to the early forest stages of
ruffed grouse and the tallgrass prairie habitat of the
greater prairie chicken. Columbian sharptails cur-
rently exist in small isolated populations, with the
most viable segments in western Colorado; they
associate largely with Kuchler's (1964) sagebrush
steppe type (Miller and Graul 1980). The plains race
is the most prairie dwelling of the three, and is
found in association with patches of shrubs on the
prairie.

Both races utilize a combination of trees-shrubs-
grasslands but in varying degrees of composition of
these structural components. The Columbian sharp-
tail favors rolling hills or benchland with a bunch-
grass, forb, and shrub combination. In valleys of
western Colorado, this bird overlaps the zones of
sagebrush and Gambel oak ( Quercus gambellii)
(Hoffman and Alexander 1980). Frequently agricul-
tural fields are interspersed on the ridgetops or val-
ley floors and may serve as feeding sites in the fall
before snow cover.

The plains sharp-tail habitat consists of short to
mid-grass prairie, interspersed with shrubby draws
and agricultural land. Grazing intensity greatly influ-
ences the distribution of grouse in this habitat. Niel-
son and Yde ( 1981 ) noted a uniform use of pastures
where grass height averaged 22.8 cm (91 in.) com-
pared with those with grass averaging 9-11 cm ( 3.6-
4 in. ) where use was more confined to sections
where shrubs provided cover.

Sharp-tailed grouse.

Columbia sharptails are most frequently associ-
ated with aspen, sagebrush, hawthorne, serviceberry,
and rose. In western Colorado, a rather close winter
association is frequently seen with Gambel oak
which may provide both mast and buds for food.
Plains sharptails are linked with such key species as
aspen, cottonwood ( Populus sp. ), buffaloberry, juni-
per (Juniperus sp. ), snowberry, and rose. Nielson
and Yde ( 1981 ) indicated a close association with
buffaloberry during winters of heavy snow. Swenson
( 1985) quantified a shift by plains sharptails from
feeding in grain stubble to shrubs as snow depth
increased beyond 140 mm (5.5 in.).

Wild Turkey (Meleagris gallopavo)

Two subspecies of wild turkey, the Merriam's
(M. g. merriami) and the Rio Grande (M. g. interme-

416

Upland Game Birds

dia), are common in portions of the western U.S. A
third, the Mexican or Gould's turkey (M. g. mexi-
cana), has limited distribution in the U.S., being con-
fined to a couple of small areas in New Mexico and
Arizona (Lee 1959). The Merriam's turkey was origi-
nally associated with the mountains of Colorado,
New Mexico, and Arizona. However, trapping and
transplanting has extended the range to most States
west of the Mississippi to the Canadian border. This
range expansion is not continuous; rather the distri-
bution consists of discrete populations, many iso-
lated in small islands of mountain habitat. The Rio
Grande turkey is found largely in Texas and the
northeast corner of New Mexico.

The Merriam's turkey in its original range is a
mountain-dwelling bird ranging from 1,800 to 3,000
m (6,000 to 10,000 ft) in elevation in the ponderosa
pine-oak forest (Ligon 1946). It frequently has a
distinct summer and winter range, moving to lower
elevations in the winter to zones producing mast
crops, which are more available due to lighter snow
conditions. Many of the more successful transplant
sites north of its original distribution are character-
ized by relatively low rainfall (and snowfall), ponde-
rosa pine and grassland areas, and foothill grainfields
adjacent to stringers of timber or shrubby draws.
Mast producing trees and shrubs are mostly absent
from the northern areas and domestic grain is fre-
quently a substitute.

Ponderosa pine forests with open grassland un-
derstory characterize Merriam's turkey spring, sum-
mer, and fall habitat. In a study in southeastern
Montana, Jonas (1966) observed 6,271 turkeys in
468 groups, recording the observations by plant
communities. Over a third of the observations, ex-
cluding those made during the winter, were in grass-
lands, but only 9% were more than 100 m (109
yd) from tree or shrub cover. Twenty-seven percent
of all observations were in snowberry, but mostly
in the brush-grass ecotone and not in pure stands.
Much use of snowberry was for cover by young
poults — older poults used the forest for cover. Over
20% of all observations were in pine stands and 78%
of these were in stands of pole size class. Stands of
pole size pine, in contrast to saplings, were open
enough to permit growth of a variety of low growing
food plants but still provided overhead protection
from avian predators. Only in winter did the use of
pine forests decrease to less than 20% . During this
time, 73% of the observations were in deciduous
tree-shrubby draws at the lower elevations where
weather was less severe. Beasom (1970) reported on
Rio Grande turkeys inhabiting mesquite (Prosopsis
juliflora) and live oak (Quercus virginiana) vegeta-
tive types in south Texas. Kothmann and Litton
(1975) reported that this subspecies extended its
range into 14 million ha (35 million a.) of shrub
mesquite prairie in west Texas.

Ponderosa pine is undoubtedly the single most
important plant species in Merriam's turkey habitat.
Sapling doghair stands of this species are used for
cover (Hoffman 1962), the older more open stands
are used for cover and feeding areas, and large ma-
ture trees for roost trees and mast producers for
winter feed. Other mast producers in the more
southern ranges are oak, primarily Gambel oak, pin-
yon pine (Pinus edulis), and alligator juniper (Juni-
perus deppeana) (Ligon 1946).

Because most mast producers do not produce
heavy crops annually, other food sources must often
be relied upon. This is frequently provided by the
grass family, both seed heads and vegetative parts,
from a variety of species (Ligon 1946; Spicer 1959;
Hoffman 1962; Jonas 1966). In northern extensions
of the range, mast crops are even less dependable,
and grasses play a greater role in winter foods.
Often, waste domestic grains in stubble fields pro-
vide supplemental winter foods, particularly when
grazing by domestic livestock is excessive and native
grass seed production is nil.

Wild turkeys.

The extension of the Rio Grande turkey beyond
its primary range into the relatively treeless area of
west Texas was assisted by man-made structures
(primarily power lines and poles) for roosting, and
available food supplies in the form of grasses, grass
seeds, and forbs (Kothmann and Litton 1975). The
importance of roost sites to Merriam's turkeys is
illustrated by a reported 64% reduction in popula-
tion following a pinyon-juniper reduction program
which isolated roost sites over 300 m (330 yd) from
cover (Scott and Boeker 1977).

Mearn's Quail (Cyrtonyx montezuma
mearnsi)

The largest part of the overall distribution of
this quail is in Mexico, although one race, the

Upland Game Birds

417

Mearn's ( C. m. mearnsi), is found in westcentral
Texas, southwest New Mexico, and southeast Ari-
zona. The male has a striking facial plumage of
bluish-black patches mixed with white, hence the
common name harlequin quail.

Throughout its range, the Mearn's quail dis-
places the Gambel's ( Callipepla gambellii) and scaled
quails (G squamata) of the desert grasslands at
higher elevations (1,200-3,600 m [4,000-12,000 ft]).
In Arizona, Mearn's quail inhabit the oak-grasslands
and oak-woodlands (Bishop and Hungerford 1965).
In New Mexico, Mearn's quail habitat is typified
more by pinyon pine and junipers as well as ponde-
rosa pine (Ligon 1961).

Although this species of quail was considered to
be rare in the U.S., Brown (1982) noted that major
populations are present in mountain ranges of south-
eastern Arizona in oak woodland and grassland type,
and is common enough to be an important game
bird in Arizona (Brown 1979). In the Mexican high-
lands, Leopold and McCabe (1957) believed that this
quail was so closely associated with the pine-oak
forest that it could well be considered an avian indi-
cator of the type.

The overstory of pine and oak is not the impor-
tant habitat component of this species. Rather it
signifies a climatic zone harboring a complex of per-
ennial forbs and sedges that provide the food and
water source in the form of bulbs (Leopold and
McCabe 1957). Although heavy grazing by livestock
was believed responsible for a severe reduction in
quail food, more recent studies have shown that
forage removal > 55% will eliminate a quail popula-
tion, not because of a short food supply but because
of cover removal and subsequent vulnerability of
the birds (Brown 1982).

Most food habit studies show that acorns are
fairly high in both frequency and volume, particu-
larly in years when some of the other foods may be
scarce (Bishop and Hungerford 1965). Thus the
various species of scrubby oaks so characteristic of
the U.S. range of this quail probably play a dual role
as food and cover. Woodsorrel ( Oxalis sp. ) and
sedges are both heavily used for food (bulbs and
seeds) and only during July through September did
these two genera make up less than 50% of the vol-
ume of food in a study in Arizona (Bishop and
Hungerford 1965).

Masked Bobwhite (Colinus virginianus
ridgwayf)

The masked bobwhite quail formerly occupied
mesquite-grassland areas of southern Arizona and
northern Sonora, Mexico. They disappeared from Ari-
zona by 1900, presumably because of cover removal

resulting from a series of droughts coupled with
excessive grazing (Tomlinson 1972b).

Descriptions of masked bobwhite habitat have
been obtained from studies conducted in Mexico
(Tomlinson 1972a, b) or on transplanted birds in Ar-
izona (Goodwin and Hungerford 1977). Masked
bobwhite show preference for flat, mesquite-grass-
land at elevations from 300 to 1,200 m (1,000 to
4,000 ft). This is in contrast to the other quails of
the area that exhibit a preference for a more
wooded and broken terrain (Tomlinson 1972b). Pen-
reared bobwhites released in Arizona also showed a
preference for open, grass-forb sites but adjacent
to mesquite lined drainages (Goodwin and
Hungerford 1977).

Masked bobwhites prefer heavy forb growth in
moderate to dense stands of grasses. Grasses com-
mon to their Mexican habitat are several species of
grama (Bouteloua sp.), three-awn, bristlegrass (Se-
taria sp.), and panic grasses (Panicum sp.; Tomlinson
1972a). Quail released on Arizona sites were later
found in areas with a dense understory of pigweed
(Amaranthus sp. ) and an overstory of mesquite, pa-
loverde (Cercidium sp.), and wolfberry (Lycium sp.).
Like the birds observed in Sonora, a preference was
shown for a dense grass-forb complex with adjacent
brush or trees (Goodwin and Hungerford 1977). The
woody cover was used for cover by the quail during
the winter (Tomlinson 1972a).

Scaled Quail {Callipepla squamata)

The scaled or blue quail, as it is often called, has
a geographic distribution which conforms closely to
the Chihuahuan desert and adjacent grasslands
(Johnsgard 1973). In New Mexico, it is found over
most nonforested sections up to 2,100 m (6,990 ft)
(Ligon 1961). Hoffman (1965) described the scaled

Scaled quail.

418

Upland Game Birds

quail range in Colorado as extremely variable from
the flat farmlands at 1,020 m (3,400 ft) to the rocky
slopes and canyons adjacent to the Rocky Mountains
at 2,100 m (7,000 ft).

In Arizona good scaled quail habitat consists of
low-growing grasses, forbs, and shrubs with an over-
all ground cover between 10 and 50% (Goodwin
and Hungerford 1977). In contrast to Gambel's quail
(which seemed to prefer taller shrubs — 60 to 80%
over 2 m [6.6 ft]), scaled quail were found to fre-
quent areas where less than 2% of the shrubs ex-
ceeded 2 m (6.6 ft) in height. In the absence of
shrubby cover in western Oklahoma, they readily
used man-made structures (Schemnitz 1961). In
Colorado, scaled quail are most commonly found in
sand sagebrush, dense cholla cactus (Opuntia sp.), or
yucca (soapweed [Yucca glauca])-grassland, and pin-
yon- juniper woodlands (Hoffman 1965).

Species of plants used for cover by scaled quail
seemed secondary to structure. Schemnitz (1961)
identified habitat use in Oklahoma based on over
2,000 observations in three vegetation types and
found 29% in man-made cover (such as buildings
and machinery) and 54% in a shrub life-form, pri-
marily skunkbush (Rhus aromatica), small soapweed -
sandsage, and small soapweed. In New Mexico,
prominent shrubs in scaled quail habitat were mes-
quite, catclaw (Acacia greggii), whitehorn (Acacia
constricta), and snakeweed (Guticrrezia sp.)
(Campbell et ai. 1973).

Based on the analysis of 221 crops collected in
the fall and winter in New Mexico, seven foods
made up over 75% of the total volume: seeds, leaves,
and stems of whitehorn, snakeweed, doveweed ( Cro-
ton sp.), whitemargin euphorbia (Euphorbia albo-
marginata), insects, mesquite, and amaranth (Ama-
ranthus sp.) (Campbell et al. 1973).

Gambel's Quail (Callipepla gatnbeUii)

Gullion (I960) reported the Gambel's quail as
the most promising upland game bird in parts of the
Southwest because in many desert areas it is the
only resident upland game species in abundance. Al-
though records exist of scattered populations of
Gambel's quail occurring at higher elevations, maxi-
mum elevations of normal occurrence are more in
the 1,220- to 1,830-m (4,000- to 6,000-ft) range.
Only marginal populations exist where annual snow-
fall is greater than 50 cm (20 in.) or where 2.5 cm
(1 in. ) or more snow is on the ground for more than
40 days each year (Gullion I960).

A preference for dense stands of desert shrub
was clearly evident from Goodwin and Hungerford's
(1977) Arizona study. They found this species of
quail favoring shrub cover of sufficient density to
shade from 50 to 75% of the ground and where 60

to 80% of the shrubs were taller than 2 m (6.5 ft).
They suggested that a dense understory was unim-
portant to Gambel's quail.

Considerable overlap exists in the distribution
of scaled and Gambel's quail and although they may
be found in the same habitats, their habitat require-
ments are different. Generally, Gambel's quail are
found in the river bottoms and are more associated
with dry washes or brushy draws; scaled quail will
more likely occur on the surrounding mesas and
plains in association with grassy open lands (R.L.
Tomlinson pers. commun.).

Mature mesquite, hackberry ( Celtis reticulata),
wolfberry, and catclaw provided cover for feeding,
loafing, roosting, nesting, and raising broods in Ari-
zona (Goodwin and Hungerford 1977). Gullion
(I960) listed many of the same plants plus other
shrubs providing a comparable structure. In addition,
he listed a variety of noxious weeds such as redroot
amaranth (Amaranthus rctroflexux), goosefoot
(Chenopodium sp.), sunflower (Helianthus annuus),
and white sweetclover (Melilotus alba) as low grow-
ing forms providing a good food source.

Hungerford (1962), in studying food habits of
this quail in Arizona, reported members of the Leg-
uminosae as very important food producers. Six of
nine genera listed as important food items were of
the legume family including mesquite, catclaw, mi-
mosa (Mimosa sp.), deervetch (Lotus sp.), paloverde,
and lupine.

Gambel's quail.

California Quail (Callipepla californica)

This close relative of the Gambel's quail is dis-
tributed along the West Coast. It has also been suc-
cessfully introduced into the interior in Washington,

Upland Game Birds

419

Oregon, Idaho, Nevada, and Utah. Although hybridi-
zation with Gambel's quail occurs where the two
species coexist, ecological differences prevent exten-
sive overlap in distribution (Johnsgard 1973).

With the extensive north-south distribution of
this quail, at least eight races are found over a rather
wide range of climatic and moisture conditions. In
California, this quail can be found from near sea
level to 2,550 m (8,500 ft). Although several races
occur in California and a wide variety of habitat con-
ditions are utilized, brushy vegetation (stiff- twigged,
dense-foliaged trees for night roosting) adjacent to
open grassy, weedy types and available water charac-
terize this species' requirements throughout the
state (Grinnel and Miller 1944). In comparing habi-
tats of the mountain quail ( Oreortyx pictus) and
California quails, Gutierrez (1980) listed the latter as
preferring the open woodland, grassland, and chapar-
ral habitats on less steep hillsides.

California quail habitat reported by Emlen and
Glading (1945) includes the following general types:
desert, rangeland, dry farming land, and irrigated
land. Rangeland is the most extensive and most
widely used of the four types. Each type provides for
the basic requirements of the quail in various de-
grees. Water, which is apparently a more critical
factor than with Gambel's quail, is usually at a pre-
mium on dry-land farms and portions of the desert
habitat. Cover (roosting, loafing, and nesting) is often
lacking in both irrigated and dry-land farm areas.
Deserts and rangelands (particularly the latter ) will
most often provide the variety of conditions needed
for food, water, and various cover components to
house the more dense populations (Emlen and
Glading 1945).

Like many of the quail that feed on a wide vari-
ety of foods (including seeds and leafy parts of
plants), species composition of cover is probably far
less important than structural or life -form character-
istics. This is particularly true of the California quail
in light of its very wide distribution through diverse
vegetational zones. A variety of food habit studies
repeatedly show the importance of legumes, espe-
cially the seeds (Edminster 1954; Shields and Dun-
can 1966; Gutierrez 1980). Important legumes
appearing in food include bur clover (Medicago
hispida) lupines, deervetches (Lotus sp.), clover (Tri-
folium sp. ), and vetches ( Vicia sp. ). Another herb
that appears frequently as food, both as leafage and
seeds, is filaree ( Erodium sp. ).

Band- tailed Pigeon (Columba fasciata)

The band-tailed pigeon is a large migratory dove
similar in size and color to the non-migratory blue-
phase domestic pigeon (Columba livia). Two races
of this bird, the interior race (C. f. fasciata) and the
coastal race (C. f. monilis), have been identified

north of Mexico. The interior race migrates into
Mexico for 6 months during the winter; the coastal
race winters primarily in southern California (Braun
et al. 1975; Tomlinson 1983).

Band- tailed pigeon.

Both races of band-tailed pigeons are closely
associated with forests that provide their major food
source in the form of mast, berries, and small fruits.
In some areas, when natural forest foods are scarce,
usually during migration, grain fields and orchard
crops are invaded to a point of human conflict. The
coastal race occupies a diverse array of mountain
forests from moist coast forests in Washington and
Oregon to drier Sierran montane forests further
south and east (Jeffrey 1977). The interior race oc-
cupies a wide range of habitats ranging from agricul-
tural types near forests to berry-producing forest
sites at elevations to 3,300 m ( 10,900 ft). Habitat is
almost always mountain related although occasion-
ally feeding birds will be observed in grassland or
desert shrub.

Habitat structure for the coastal race in its
northern range consists of forested land, well inter-
spersed with forest openings in early successional
stages favoring berry-producing shrubs. California
habitat includes mountain forests, woodlands, and
chaparral if accompanied by abundant oak. Ranges of
the interior race are primarily diverse mountain con-
iferous forests but most frequently accompanied by
a common denominator — a pine-oak combination
(Jeffrey 1977). In southern Arizona and New Mex-
ico, breeding habitat may be found in oak communi-
ties (Braun et al. 1975).

In the northern habitat of the coastal race, a
variety of overstory trees such as western hemlock
(Tsuga heterophylla), red cedar (Thuja plicata), and
Douglas fir are dominants, whereas in California the
redwood (Sequoia sempervirens) predominates in
the coastal forest. However, because food supplies

420

Upland Game Birds

appear to control the distribution and abundance of
bandtails more than any other single factor, species
of understory trees and shrubs are probably most
important. Various species of oak are key deciduous
trees. Several species of shrubs that are good food
producers, and most vigorous and productive in
serai stages following logging or fire, are elderberries
(Sambucus sp.), wild cherries, huckleberries, and
dogwood (Cornus sp. ) (Jeffrey 1977).

The primary winter range of the coastal race is
in California. During winter they primarily inhabit
the pine-oak woodland and coastal chaparral plant
associations, although wintering populations are mo-
bile and may shift areas of concentration from year
to year depending on available food sources
(Tomlinson 1983).

Although the interior race can be found at ele-
vations supporting Engelmann spruce (Picea engel-
manni) and alpine fir (Abies lasiocarpa), it is most
abundant at the lower ponderosa pine/Gambel oak
zone. The close association with various species
of oak and pine is related to the staple fall and win-
ter foods of acorns and pine nuts.

Mourning Dove (Zenaida macroura)

Two races of mourning doves breed in the U.S.,
Z. m. carolinensis in the eastern one-third and Z.
m. marginella in the western two-thirds. These doves
breed in all of the lower 48 States, extending into
the Canadian provinces to the north and Mexico to
the south. Most mourning doves migrate, spending
the winter in the southern U.S., Mexico, Central
America, or the West Indies (Keeler 1977).

The mourning dove has probably benefited from
or has been able to adapt to man's activities more
than most other native bird species. It is primarily a

Mourning dove.

bird of open woodland or the edge between forest
and prairie. Thus, the clearing of large areas of decid-
uous forest in the East and the planting of trees on
the prairie (i.e., shelterbelts for farms and fields, ur-
ban tree planting) has enhanced the dove population
over large areas. The dove is basically a tree nester
but will commonly nest on the ground in the ab-
sence of trees or shrubs. Early in the nesting season
before deciduous trees leaf out, doves prefer conifers
for nest sites. Thus the inclusion of conifers in shel-
terbelts and urban plantings has also favored this
bird. The conversion of large tracts of treeless prairie
to domestic grains and farmsteads (trees) has cre-
ated an excellent combination of food (waste grain)
and nesting cover for doves. Intensive grazing on
many ranges has encouraged invader plant species
that often produce more seeds than native grasses.
Also intensive grazing management includes stock-
dams that provide water in areas where it is at a
premium, and often trees and shrubs for nesting and
loafing (Dunks et al. 1982).

Population Measurement Techniques

Most techniques for inventorying upland game
birds provide direct or indirect data on the presence
or absence of the target species within the area
being sampled. Such techniques may consist of enu-
merations of actual sightings or audible responses by
the bird to a stimulus (direct) or the observation of
droppings, feathers, or tracks (indirect), either of
which can provide the manager with at least a basis
for initial habitat management decision.

Many techniques, primarily those based on a
seasonal activity of the bird (usually breeding activ-
ity), show sufficient precision to provide relative
abundance data. Thus comparisons can be made be-
tween areas or between years on a single area per-
mitting an evaluation of the bird's response to
different or changing habitat conditions.

Few if any techniques measure the actual num-
ber of birds present. Even on limited areas where
a large percentage of the male population may be
engaged in a common activity (breeding display), an
unknown percentage of the males are not displaying
(Dorney et al. 1958; Rippin and Boag 1974).

Although a few upland game census techniques
involve complete coverage of a unit, most are based
on variations of transect sampling. Likewise, a basic
walking transect census, such as the King Census for
ruffed grouse originally described by Leopold
(1933), has been modified for a variety of species
and conditions as pointed out by Hayne (1949) in
his examination of the technique. Thus, many meth-
ods for determining indexes to upland game bird
numbers are derived from a basic technique but
modified slightly to accommodate a particular spe-
cies, habitat type, or season. A very comprehensive

Upland Game Birds

421

and exhaustive coverage of transect sampling was
presented by Eberhardt ( 1978).

All of the techniques listed below will at least
indicate presence or absence. The degree of preci-
sion and accuracy beyond that will be indicated with
each method.

Dropping Counts.

Description. Although this technique (pellet-
group counts) has been used extensively with large
ungulates, it has been little used with birds. Pyrah
(1972) used this method successfully in determining
relative use by sage grouse on several study areas
subjected to various degrees of herbicide treatment.
In areas previously gridded, points were randomly
selected on which to establish 30-m (100-ft)
transects. At 7.5-m (25-ft) intervals along the
transect, 9-m (100-ft ) circular plots (radius [1.7
m]5.64 ft) were established and searched for
droppings. Distinction was made between single
droppings and clusters resulting from a night roost.

Accuracy and Precision. No attempt was
made to correlate the data with the actual number of
birds using the area. However, based on field obser-
vations during this rather intensive study, the data
reflected relative use by sage grouse among the dif-
ferent areas during the previous winter (D. Pyrah,
pers. commun.).

Discussion. This indirect census method has
some distinct advantages over many for game birds.
Sampling can be conducted throughout the day in
contrast to short periods of bird activity. By clearing
permanent plots after each count, seasonal use can
be measured in a relatively short period of time
compared with measuring directly through actual
bird observation. Gates (1983) used the same tran-
sects and in part the sample plots to determine rela-
tive use by sage grouse, pygmy rabbits (Sylvilagus
idahoensis), black-tailed jack rabbits (Lepus californi-
cus), and pronghorn (Antilocapra americana).

The method also has some disadvantages: its use
is largely restricted to habitats with moderate to
light ground cover and to the period after snow melt
and before excessive green-up; it is probably re-
stricted to species (several of the grouse) on a diet
including adequate roughage to prevent rapid deteri-
oration of the droppings; and until more information
on defecation and dropping deterioration rates is
available, the method will be restricted to a relative
use index.

King Strip Census.

Description. The King Strip census has been
modified and discussed many times since first

employed on ruffed grouse by King ( 1973). In its
simplest form, a series of parallel transects of known
length are walked, and birds observed are recorded
including the flushing distance. The average flushing
distance is calculated, which when multiplied by
the total length of the transects provides the area
covered. The total area divided by the number of
birds observed provides the density index. A more
detailed description of this method is presented by
Overton (1971, pp. 420-424).

Accuracy and Precision. Under most condi-
tions this method would denote presence and with
adequate samples may indicate distinct trends in the
population. One of the longest continuous uses of
this method on a single area (Marshall 1954) sug-
gested similar magnitudes in population when com-
pared to grouse numbers with a second index from
the same area.

Discussion. This method has some merit in
areas or seasons where the use of the other methods
are not feasible; it can indicate presence and to a
degree, relative abundance of the target species.
Because so many variables are potentially present in
using such a method, it is imperative that conditions
be standardized as much as possible, i.e., time of year
(phenologically), time of day, and comparable
weather conditions. One advantage of this method is
that it forces the biologist onto the ground to sys-
tematically walk through the habitat.

This method does not provide usable indexes
when used on low density populations — the amount
of sampling necessary would be prohibitive.

Roadside Counts.

Description. As the name implies, this method
consists of driving along a predetermined route
under a prescribed set of conditions and counting
the number of birds observed (Kosicky et al. 1952).
One should attempt to standardize driving speed,
number of observers, season, time of day, and
weather conditions (wind speed, temperature, and
cloud cover). Thus many of the variables
encountered in walking strip counts (King Strip
census) would be the same as in roadside counts;
the latter permits more extensive coverage in less
time.

Accuracy and Precision. Basically the same as
for the preceding method.

Discussion. Roadside counts have been used
extensively in the Midwest to obtain indexes on
pheasants, quail, doves, and cottontails. In the West
where game bird densities are frequently lower than
in the Midwest, the roadside count is superior to a
strip count conducted on foot. In addition to popula-
tion indexes, roadside counts have also been used

422

Upland Game Birds

to obtain cover or habitat preferences by recording
habitat use in relation to availability along the route
(Hoffman 1965).

Like the strip census, the greatest drawback to
this method is the number of variables that can influ-
ence the results. The availability of cottontails for
enumeration can vary considerably from day to day
as discussed by Newman (1959). The basic problems
would be similar for game birds.

Complete Census.

Description. The complete census attempts to
tally all birds in a unit of habitat of known size. The
procedure can vary from one individual carefully
cruising back and forth to one or two persons with
well trained dogs, to many individuals essentially in a
drive. In all instances care must be exercised to
prevent duplication in tally. The distance between
lines, whether covered singly or by many, is largely
dictated by the type of cover.

Accuracy and Precision. This technique,
when properly executed in cover that is not too
dense, has been considered to be more accurate and
reliable than many others.

Discussion. For most species of game birds,
the term complete count would be a misnomer.
However, on tracts without heavy cover, with spe-
cies more prone to flush than run, and with adequate
personnel to keep intervals between participants at
a minimum, a drive count can be effective in making
a large percentage of the birds available for tallying.
The results are influenced by many of the same vari-
ables mentioned for strip counts, and maximum
standardization of the variables is very critical to
comparing the results between areas or years.

This method is more manpower-expensive than
strip sampling or roadside counts and does not sam-
ple as large an area for the effort. One modification
that has been employed on a number of game birds,
particularly covey species, is to incorporate well-
trained pointing dogs into the effort. This can be
very effective but all too often simply adds another
variable. This method is obviously impractical for
use over wide areas (county or state); its application
is best confined to areas of intensive study.

Auditory Census.

Description. As used here, this method
includes all sounds produced by a bird. Most of the
sounds are vocalizations although some, such as
those produced by a drumming ruffed grouse, are
mechanically produced. The sounds used in
obtaining population indexes are associated with the
breeding season.

The most basic auditory method is to travel
over a predetermined route, stopping and listening
at established intervals (listening stops) for a given
length of time, during which period the observer
records the number of sounds produced or the num-
ber of birds that can be heard per stop. Applications
or evaluations of this method have been reported for
a variety of game bird species: crowing count for
pheasants (Kimball 1949), a drumming count for
ruffed grouse (Petraborg et al. 1953), mourning
dove call-count survey (Armbruster et al. 1978), and
woodcock (Scolopax Minor) singing ground count
(Tautin 1982).

A modification of this method has been used for
lekking species; an observer listens for displaying
males along a roadside route, plots the locations of
arenas, and counts the number of males per arena on
subsequent days. The effective width of such a strip
is determined by twice the average distance the
target species can be heard. A similar effort is often
expended in locating all arenas on a block of habitat.
Intensive coverage of a block of land with its redun-
dancy through overlap perhaps gives greater assur-
ance of locating all the arenas per unit by using
listening stops along a belt.

Still another variation of an auditory census is to
employ a recorded sound of the species in question
to stimulate a response. This may be a male call to
elicit a response by a territorial male as suggested by
Kimball (1949) for pheasants; or a female call which
may also cause the male to respond as reported by
Stirling and Bendell (1966) for blue grouse; and
Levy et al. (1966) for Mearn's, Gambel's, and scaled
quail. This variation introduces another potential
variable — the proportion of the population that re-
sponds to the stimulus may vary from year to year
depending on changes in cover conditions and popu-
lation density.

Accuracy and Precision. The auditory census
techniques must be used for indexes only, at least
under the present methods employed. The data gath-
ered do not permit conversion to actual numbers of
birds. Even with the total coverage of a unit for are-
nas and high counts of males attending, data are
available to show the presence in the population of
an unknown number of non-participants (Dorney
et al. 1958; Rippin and Boag 1974).

Certain of the methods, if conducted properly,
will provide trend data of value to the manager
(Dorney et al. 1958 — ruffed grouse; Smith and Galli-
zioli 1965 — Gambel's quail; Jenni and Hartzler
1978 — sage grouse).

Discussion. Auditory counts hold the most
promise as a usable technique for the habitat man-
ager. Most calls can be more easily heard than birds
can be seen. Consequently, at least presence can

Upland Game Birds

423

be determined. Because most calls are related to a
breeding cycle, and this is often an all out seasonal
effort, certain variations are avoided. Because a high
proportion of the male population usually partici-
pates, auditory counts can be used on less dense
populations than methods relying on sight
observations.

The major disadvantage to auditory counts is
that most are restricted to the breeding season and
consequently provide only an index to the adult
population. Indexes to the adult population are valu-
able in comparing populations between years or
between habitats, but frequently are inadequate for
use in predicting a harvestable surplus for the fol-
lowing fall. However, Smith and Gallizioli (1965)
suggested that the index they obtained from calling
quail reflected both numbers of birds and breeding
success, thus providing some predictability of fall
numbers and hunter success. Calling counts for
mourning and white-winged doves (Zenaida asiatica)
were believed to be a valid technique for determin-
ing spring population levels, but only for mourning
doves did the counts provide reasonable predictabil-
ity for fall hunting success (Brown and Smith 1976).

Lek (Arena) Surveys.

Description. Lekking species (prairie chickens,
sharp-tailed grouse, and sage grouse in North
America) are those which conduct group breeding
display on traditional areas (arenas). The arenas may
be located through an auditory census
(triangulation), auditory-visual (systematic ground
coverage), or visual (aerial coverage). Persistent
attendance by some of the males early in the season
permits locating the arenas 2 to 3 weeks before the
seasonal peak attendance. Peak male attendance on
sage grouse arenas occurs in 1 to 3 weeks following
peak female attendance (Eng 1963; Jenni and
Hartzler 1978). Three to four counts of male sage
grouse on each arena during the same period each
year will produce the most reliable trend data (Jenru
and Hartzler 1978; Emmons and Braun 1984).
Counting only the number of arenas (leks) for lesser
and greater prairie chickens provides an index to
density of displaying males (Cannon and Knopf
1981b).

Accuracy and Precision. Over 90% of the 33
radio-marked male sage grouse attended the display
grounds during the period of peak male attendance,
indicating that most of the males of an area can be
found on the arenas (Emmons and Braun 1984).
However, in a sharp-tail population, substantial num-
ber of males (about one-half during one year) were
non-territorial and non-participants in display (Rip-
pin and Boag 1974).

Discussion. Determining trends in populations
of lekking grouse through counts of displaying males
is considered an accepted practice. For sage grouse,
available data suggest that most of the males will
be present if counts are scheduled properly. Sage
grouse arenas are very traditional; arenas are less
likely to be abandoned or new ones started than
with sharp-tailed grouse. Thus for sage grouse,
counts of males on arenas is necessary to reflect
trend data. Counts of arenas only on a unit area, as
recommended for obtaining trend data for prairie
chickens (Cannon and Knopf 1981b), may also suf-
fice for sharptails.

DISCUSSION

Wildlife biologists responsible for habitat man-
agement on public lands frequently find themselves
in the uncomfortable position of having to make
recommendations based on scanty data. Further-
more, the size or diversified nature of many areas
make it difficult to maintain the intimacy necessary
for sound management. Thus an ongoing search is
conducted for yardsticks to be used in determining
suitability of cover for or the status of a particular
population.

Probably as a result of the popular use of key
plant species in evaluating big game habitat, game
bird habitat has been evaluated in much the same
manner. Yet a species of game bird may have a dis-
tribution of sufficient breadth that the plant species
composition may be very different from one end
of the range to the other. It appears, then, that at
least with respect to cover, structure of the vegeta-
tion is probably a better common denominator
throughout the birds' range than plant species
composition.

On the other hand, when food requirements are
considered, plant species are important. Many of
the same plants, at least members of the same gen-
era, are listed as food sources throughout much of
the range of Gambel's quail. The entire distribution
of the sage grouse is governed by the plant genus
Artemisia. Although the sage grouse uses Artemisia
for both food and cover, their dependence on this
plant for winter food dictates this inseparable rela-
tionship. This plant structure (cover) and plant spe-
cies (food) relationship has been clearly depicted for
marsh birds by Weller (1978).

Studies of upland game populations and of habi-
tat change are often conducted simultaneously. Un-
fortunately many habitat studies are not initiated
until a decided decline in bird numbers has been
noted. This situation often results in attempts to
evaluate deteriorated habitat from which key ingredi-
ents may already be missing.

424

Upland Game Birds

This sequence should not be totally unexpected
because land use (habitat) changes can be very eas-
ily overlooked when they occur gradually. Even the
occasional conversion of rangeland to cropland may
seem somewhat innocuous to a manager of grassland
fauna until it is conducted on a scale of some of the
recent sod-busting ventures. Thus it is often the con-
sistent, annual, downward trend in numbers of a
species of game birds that calls one's attention to the
habitat problems. A study is then initiated, and as
Robel (1980) pointed out, the population and habi-
tat studies become inseparable.

Knowledge of a game bird's annual home range
is of prime importance both in assessing its habitat
or determining its presence or relative abundance.
With a species like the gray partridge (and many of
the quail), the home range is sufficiently small so
that most transects established for sampling habitat
or birds would be of sufficient length to include
areas occupied throughout the year. As such, sea-
sonal timing of the sampling effort would be less
critical. Conversely, more mobile species like sage
grouse (Berry and Eng 1984), blue grouse (Mussehl
I960), as well as migratory doves, may seasonably
occupy habitats somewhat different in type and
measurably different in location. Therefore, with
such species, timing of the effort within a particular
habitat is most critical.

Overton (1971) defined a census index as "a
count or ratio which is relative in some sense to the
total number of animals in a specified population."

The problem here, of course, is determining how the
count or ratio relates to the total number, since the
latter figure favors playing the role of the unknown.

Most efforts directed toward measuring the sta-
tus of a game bird population result in data being
compared to similar data collected from the same
area in previous years (population trend data), or to
similar data gathered from different areas (relative
abundance data). Trend data are most frequently
used by agencies concerned with annual population
status as it applies to harvest recommendations, al-
though long-term trends can also provide informa-
tion on the status of the habitat. Relative abundance
data from different areas are probably of greater
use to those agencies primarily concerned with habi-
tat management. Such data enable personnel to as-
sess the role of a given species within a unit of
habitat in relation to other potential uses.

In most instances, when target species and sea-
son are appropriate, a version of the auditory census
will probably provide the best return for the effort.
This would be true even for a simple presence or
absence determination but especially so if relative
abundance data are of concern. As mentioned earlier,
the fact that the birds (at least the males) are en-
gaged in a seasonal activity which will have a meas-
urable seasonal and daily peak permits the observer
to capitalize on this predictable sequence and
thereby reduce considerably the variables encoun-
tered. A reduction in variables will provide a more
usable measure of the population level.

Upland Game Birds

425

LITERATURE CITED

ALLEN, A.W. 1984. Habitat suitability index models: gray
partridge. U.S. Dep. Inter., Fish and Wildl. Serv. Biol.
Rep. 82(1073). 23pp.

ARMBRUSTER, M.H., T.S. BASKETT, W.R. GOFORTH, and
K.C. SADLER. 1978. Evaluating call-count procedures
for measuring local mourning dove populations.
Trans. Missouri Acad. Sci. 12:75-90.

AULT, S.C. and FA. STORMER. 1983. Seasonal food selec-
tion by scaled quail in northwest Texas. J. Wildl.
Manage. 47:222-228.

BAXTER, W.L. and C.W. WOLFE, Jr. 1973. Life history and
ecology of the ring-necked pheasant in Nebraska.
Nebraska Game and Parks Comm., Lincoln. 58pp.

BEASOM, S.L 1970. Turkey productivity in two vegetative
communities in south Texas. J. Wildl. Manage. 34:166-
175.

BECK, T.D.I. 1977. Sage grouse flock characteristics and
habitat selection in winter. J. Wildl. Manage. 41:18-26.

BEER, J. 1943. Food habits of the blue grouse. J. Wildl.
Manage. 7:32-44.

BENDELLJ. and P.W. ELLIOTT. 1966. Habitat selection in
blue grouse. Condor. 68:431-446.

BERRY, J.D. and RL. ENG. 1984. Interseasonal movements
and fidelity to seasonal use areas by female sage
grouse. J. Wildl. Manage. 49:237-240.

BISHOP, R.A. and C.R. HUNGERFORD. 1965. Seasonal

food selection of Arizona Mearn's quail. J. Wildl. Man-
age. 29:813-819.

BOHL, W.H. 1957. Chukars in New Mexico, 1931-1957.
Bull. 6, New Mexico Dep. Game and Fish. Santa Fe.
69pp.

BRAUN, C.E. 1970. Distribution and habitat of white-tailed
ptarmigan in Colorado and New Mexico. Abstract of
paper presented at 46th Annual Meeting, Southwest-
ern and Rocky Mountain Division, A.A.A.S., Las Vegas,
New Mexico.

, D.E. BROWN, J.C. PETERSON, and T.P. ZAPATKA

1975. Results of the Four-Corners Cooperative band-
tailed pigeon investigation. U.S. Dep. Inter., Fish and
Wildl. Serv. Resour. Publ. 126. 20pp.

BROWN, D.E. 1979. Factors influencing reproductive suc-
cess and population densities in Montezuma quail.
J. Wildl. Manage. 43:522-526.

BROWN, D.E. and R.H. SMITH. 1976. Predicting hunter
success from call counts of mourning and white-
winged doves. J. Wildl. Manage. 40:743-749-

BROWN, R.L. 1982. Effects of livestock grazing on Mearn's
quail in southeastern Arizona. J. Range Manage.
35:727-732.

CAMPBELL, H., D.K MARTIN, P.E. FERKOVICH, and B.K
HARRIS. 1973- Effects of hunting and some other
environmental factors on scaled quail in New Mexico.
Wildl. Monogr. 34. 49pp.

CANNON, R.W. and F.L KNOPF. 1981a. Lesser prairie

chicken densities on shinnery oak and sand sagebrush
rangelands in Oklahoma. J. Wildl. Manage. 45:521-
524.

. 1981b. Lek numbers as a trend index to prairie

grouse populations. J. Wildl. Manage. 45:776-778.

CHOATE, T.S. 1963- Habitat and population dynamics of
white-tailed ptarmigan in Montana. J. Wildl. Manage.
27:684-699.

CHRISTENSEN, G.C. 1970. The chukar partridge, its intro-
duction, life history, and management. Biol. Bull. 4.
Nevada Dep. Fish and Game. Reno. 82pp.

CRAWFORD, J.A. and E.G. BOLEN. 1976. Effects of land
use on lesser prairie chickens in Texas. J. Wildl. Man-
age. 40:96-104.

DALKE, P.D., D.B. PYRAH, DC. STANTON, J.E. CRAW-
FORD, and E.F. SCHLATTERER.1963. Ecology, pro-
ductivity and management of sage grouse in Idaho. J.
Wildl. Manage. 27:811-841.

DAVIS, C.A., R.C. BARKLEY, and W.C. HAUSSAMEN. 1975.
Scaled quail foods in southeastern New Mexico. J.
Wildl. Manage. 39:496-502.

, T.Z. RILEY, RA. SMITH, and M.J. WISDOM. 1980.

Spring-summer foods of lesser prairie chickens in
New Mexico. Pages 75-80 in Proc. of the Prairie
Grouse Symposium. Oklahoma State Univ. Stillwater.

DOERR, P.D., LB. KEITH, D.H. RAUSCH, and C.A. FISHER.
1974. Characteristics of winter feeding aggregations
of ruffed grouse in Alberta. J. Wildl. Manage. 38:601-
615.

DOERR, T.B. and F.S. GUTHERY. 1983- Effects of tebuthiu-
ron on lesser prairie-chicken habitat and food. J.
Wildl. Manage. 47:1138-1142.

DORNEY, R.S., DR. THOMPSON, J.B. HALE, and R.F.

WENDT. 1958. An evaluation of ruffed grouse drum-
ming counts. J. Wildl. Manage. 22:35-40.

DUNKS, J.H., RE. TOMLINSON, H.M. DEEVES, D.D. DOL-
TON, C.E. BRAUN, and T.P. ZAPATKA. 1982. Migra-
tion, harvest, and population dynamics of mourning
doves banded in the Central Management Unit, 1967-
77. U.S. Dep. Inter. Fish and Wildl. Serv. Spec. Sci.
Rep— Wildl. 249. 128pp.

EBERHARDT, L.L. 1978. Transect methods for population
studies. J. Wildl. Manage. 42:1-31.

EDMINSTER, F.C. 1954. American game birds of field and
forest. Charles Scribner's Sons. New York, NY.

EMLEN,JT, JR. and B. GLADING. 1945. Increasing valley
quail in California. Univ. Calif. Agric. Exp. Sta. Bull.
695.

EMMONS, S.R. and C.E. BRAUN. 1984. Lek attendance of
male sage grouse. J. Wildl. Manage. 48:1023-1028.

ENG, R.L. 1963- Observations on the breeding biology of
male sage grouse. J. Wildl. Manage. 27:841-846.

and P. SCHIADWEILER 1972. Sage grouse winter

movements and habitat use in central Montana. J.
Wildl. Manage. 36:141-146.

GALBREATH, D.S. and R. MORELAND. 1953- The Chukar
partridge in Washington. Biol. Bull. 1 1 , Washington
State Game Dep. 55pp.

GATES, R.J. 1983. Sage grouse, lagamorph and pronghorn
use of a sagebrush grassland burn site on the Idaho
National Engineering Laboratory. M.S. Thesis, Montana
State University, Bozeman. 1 35pp.

GOODWIN, J.G. and C.R. HUNGERFORD. 1977. Habitat
use by native Gambel's and scaled quail and released
masked bobwhite quail in southern Arizona. U.S. Dep.
Agric. For. Ser. Res. Pap. RM-197. 8pp.

GRINNELL, J.G. and AH. MILLER. 1944. The distribution
of birds of California. Cooper Ornithological Club.
Pacific Coast Avifauna 27.

GULLION, GW. I960. The ecology of Gambel's quail in
Nevada and the arid southwest. Ecology. 41:518-536.

. 1977. Forest manipulation for ruffed grouse. Trans.

North Am. Wildl. Nat. Resour. Conf. 42:449-458.

. 1984. Ruffed grouse management — where do we

stand in the eighties? Pages 169-181 in Robinson,

426

Upland Gome Birds

W.L. ed. Ruffed Grouse Management: State of the Art
in the Early 1980s. North Cent. Sect. Wildl. Soc.

GUTIERREZ, R.J. 1980. Comparative ecology of the moun-
tain and California quail in the Carmel Valley, Califor-
nia. The Living Bird. 71-93-

HAMERSTROM, FN, JR., and F. HAMERSTROM. 1961.
Status and problems of North American grouse. Wil-
son Bull. 73:284-294.

HARPER, H.T., B.H. HARRY, and WD. BAILEY. 1958. The
chukar partridge in California. California Fish and
Game. 44(l):5-50.

HAYNE, D.W. 1949. An examination of the strip census
method for estimating animal populations. J. Wildl.
Manage. 13:145-157.

HIATT, R.W. 1947. The relation of pheasants to agricul-
ture in the Yellowstone and Big Horn River valleys of
Montana. Montana Fish and Game Dep., Wildl. Rest.
Div., Proj. 1-R. 72pp.

HOFFMAN, DM. 1962. The wild turkey in eastern Colo-
rado. Colo. Dep. of Game and Fish. Tech. Publ. 12.
47pp.

. 1965. The scaled quail in Colorado. Colo. Dep. of

Game, Fish and Parks Tech. Publ. 18. 47pp.

HOFFMAN, G.R. and R.R. ALEXANDER. 1980. Forest vege-
tation of the Routt National Forest in northwestern
Colorado: a habitat type classification. U.S. Dep. Agric,
For. Ser. Res. Pap. RM-221. 4lpp.

HUNGERFORD, C.R. 1962. Adaptations shown in selec-
tion of food by Gambel's quail. Condor. 64:213-219.

JACOBS, K.F. 1959. Restoration of the greater prairie

chicken. Oklahoma Dep. of Wildl. Conserv. Oklahoma
City.

JEFFREY, R.G., Chairman. 1977. Band-tailed Pigeon (Col-
umba fasciata). Pages 210-245 in Sanderson, G.C.
ed. Management of Migratory Shore and Upland Game
Birds in North America. International Assoc, of Fish
and Wildl. Agencies. Washington, DC. 358pp.

JENNI, DA. and J.E. HARTZLER. 1978. Attendance at a
sage grouse lek: implications for spring census. J.
Wildl. Manage. 42:46-52.

JOHNSGARD, PA. 1973 Grouse and quails of North
America. Univ. of Nebraska Press. Lincoln. 553pp.

and R.W. WOOD. 1968. Distributional changes and

interactions between prairie chickens and sharp-
tailed grouse in the Midwest. Wilson Bulletin. 80:173-
188.

JONAS, R. 1966. Merriam's turkeys in southeastern Mon-
tana. Montana Fish and Game Dep. Tech. Bull. 3.
36pp.

JONES, RE. 1963- Identification and analysis of lesser and
greater prairie chicken habitat. J. Wildl. Manage.
27:757-778.

KEELER, J.E. 1977. Mourning Dove (Zenaida macroura).
Pages 274-298 in Sanderson, G.C, ed. Management of
Migratory Shore and Upland Game Birds in North
America. International Assoc, of Fish and Wildl. Agen-
cies. Washington, DC. 358pp.

KIMBALL, J.W. 1949. The crowing count pheasant census.
J. Wildl. Manage. 13:101-120.

KING, R.T. 1937. Ruffed grouse management. J. Forestry.
35:523-532.

KOSICKY, E.L., GO. HENDERSON, P.G. HOMEYER, and
E.B. SPEAKER. 1952. The adequacy of the fall road-
side pheasant census in Iowa. Trans. North Am. Wildl.
Nat. Resour. Conf. 17:293-305.

KOTHMANN, H.G. and G.W. LITTON. 1975. Utilization of
man-made roosts by turkey in West Texas. Proc.

Third Natl. Wild Turkey Symp. 159-163-

KUBISIAK, J.F. 1985. Ruffed grouse habitat relationships
in aspen and oak forests of central Wisconsin. Tech.
Bull. 151. Wisconsin Dep. Nat. Resour. Madison.
22pp.

KUCHLER, A.W. 1964. Potential natural vegetation of the
conterminous United States. Am. Geogr. Soc. Spec.
Pub. 36. 39pp.

LEE, L. 1959. The present status of the wild turkey in New
Mexico. Proc. First. Natl. Wild Turkey Manage. Symp.:
11-18.

LEOPOLD, A. 1933- Game Management. Charles Scribner's
Sons. New York, NY. 281pp.

LEOPOLD, AS. and R.A. MCCABE. 1957. Natural history of
the Montezuma quail in Mexico. Condor. 59:3-26.

LEVY, S.H., J.J. LEVY, and R.A. BISHOP 1966. Use of tape
recorded female quail calls during the breeding sea-
son. J. Wildl. Manage. 30:426-428.

LIGON, J.S. 1946. History and management of Merriam's
wild turkey. New Mexico Game and Fish Dep. 84pp.

. 1961. New Mexico birds and where to find them.

Univ. of New Mexico Press. Albuquerque.

MARSHALL, W.H. 1954. Ruffed grouse and snowshoe hare
populations on the Cloquet Experimental Forest,
Minnesota. J. Wildl. Manage. 18:109-112.

MARTINKA, R.R. 1972. Structural characteristics of blue
grouse territories in southwestern Montana. J. Wildl.
Manage. 36:498-510.

MENDEL, G.W. and S.R. PETERSON. 1983. Management
implications of gray partridge habitat use on the Pal-
ouse Prairie, Idaho. Wildl. Soc. Bull. 11:348-356.

MILLER, G.C. and WD. GRAUL 1980. Status of sharp-
tailed grouse in North America. Pages 18-28 in Proc.
Prairie Grouse Symposium. Oklahoma State Univer-
sity. Stillwater.

MOLINI, W.A. 1976. Chukar partridge species manage-
ment plan. PR Proj. W-8-R, W-43-R and W-48-12. Ne-
vada Dep. Fish and Game. Carson City.

MUSSEHL, T.W. I960. Blue grouse production, move-
ments, and populations in the Bridger Mountains,
Montana. J. Wildl. Manage. 24:60-68.

. 1963- Blue grouse brood cover and land-use impli-
cations. J. Wildl. Manage. 27:547-555.

NEWMAN, D.E. 1959. Factors influencing the winter road-
side counts of cottontails. J. Wildl. Manage. 23:290-
294.

NIELSON, L.S. and C.A. YDE. 1981. The effects of rest-
rotation grazing on the distribution of sharp-tailed
grouse. Pages 147-165 in Proc. Wildl. — Livestock
Relationships Symposium. Univ. of Idaho. Moscow.

OVERTON, W.S. 1971. Estimating the numbers of animals
in wildlife populations. Pages 402-455 in Giles, R.H.
Jr., ed. Wildlife Management Techniques, Third Edi-
tion: Revised. The Wildl. Soc, Washington, DC.
633pp.

PATTERSON, R.L. 1952. The sage grouse in Wyoming.
Wyoming Game and Fish Comm., and Sage Books,
Inc. Denver, CO. 34 lpp.

PENDERGAST, B.A. and DA. BOAG. 1970. Seasonal
changes in the diet of spruce grouse in central Al-
berta. J. Wildl. Manage. 34:605-611.

PETRABORG, W.H., E.G. WELLEIN, and V.E. GUNVALD-
SON. 1953- Roadside drumming counts a spring cen-
sus method for ruffed grouse. J. Wildl. Manage.
17:292-295.

POTTS, G.R. 1984. Gray partridge population dynamics:
comparisons between Britain and North America.

Upland Game Birds

427

Pages 7-12 in Dumke, R.T., R.B. Stiehl, and R. Kohl,
eds. Proc. of PERDIX III: Gray Partridge/Ring-necked
Pheasant Workshop. Wisconsin Dep. Nat. Resour.
Madison.

PYRAH, D.B. 1972. Effects of chemical and mechanical
sagebrush control on sage grouse. Pages 16-25 in
Ecological Effects of Chemical and Mechanical Sage-
brush Control. Montana Fish and Game Dep., Job
Prog. Rep. W-105-R6. 79pp.

RATTI, J.T., D.L. MACKAY, and JR. ALDREDGE. 1984.
Analysis of spruce grouse habitat in north-central
Washington. J. Wildl. Manage. 48:1188-1196.

RIPPIN, A.B. and DA. BOAG. 1974. Recruitment to popu-
lations of male sharp-tailed grouse. J. Wildl. Manage.
38:616-621.

ROBEL, R.J. 1980. Current and future research needs for
prairie grouse. Pages 34-41 in Proc. of the Prairie
Grouse Symp. Oklahoma State Univ. Stillwater.

ROBINSON, W.L. 1980. Fool hen— the spruce grouse on
the Yellow Dog Plains. Univ. of Wise. Press. Madison.
221pp.

SCHEMNITZ, S.D. 1961. Ecology of the scaled quail in the
Oklahoma Panhandle. Wildl. Monogr. 8. 47pp.

SCOTT, V.E. and EL. BOEKER. 1977. Responses of Mer-
riam's turkey to pinyon-juniper control. J. Range
Manage. 30:220-223-

SHIELDS, P.W. and DA. Duncan. 1966. Fall and winter
food of California quail in dry years. California Fish
and Game. 52:275-282.

SMITH, L.M., J.W. HUPP, andJ.T. RATTI. 1982. Habitat use
and home range of gray partridge in eastern North
Dakota. J. Wildl. Manage. 46:580-587.

SMITH, R.H. and S. GALLIZIOLI. 1965. Predicting hunter
success by means of a spring call count of Gambel's
quail. J. Wildl. Manage. 29:806-812.

SNYDER, W.D. 1974. Pheasant use of roadsides for nesting
in northeast Colorado. Colo. Div. Wildl. Spec. Rep.
26. 24pp.

SPICER, R.L. 1959. Wild turkey in New Mexico— an evalu-
ation of habitat development. New Mexico Dep.
Game and Fish. Bull. 10. 64pp.

STAUFFER, D.F. 1983. Seasonal habitat relationships of
ruffed and blue grouse in southeastern Idaho. Ph.D.
Dissertation. Univ. of Idaho. Moscow. 108pp.

and S.R. PETERSON. 1985a. Ruffed grouse and blue

grouse habitat use in southeastern Idaho. J. Wildl.
Manage. 49:461-468.

and S.R. PETERSON. 1985b. Seasonal micro-habitat

relationships of ruffed grouse in southeastern Idaho. J.
Wildl. Manage. 49:605-610.

STIEHL, R.B. 1984. Critical habitat components for gray
partridge. Pages 129-134 in Dumke, R.T., R.B. Stiehl,
and T. Kahl, eds. Proc. of PERDIX III: Gray Partridge/
Ring-Necked Pheasant Workshop. Wisconsin Dep. Nat.
Resour. Madison.

STIRLING, I. andJ.F. BENDELL. 1966. Census of blue

grouse with recorded calls of a female. J. Wildl. Man-
age. 30:184-187.

STONEBERG, R.P. 1967. A preliminary study of the breed-
ing biology of the spruce grouse in northwestern
Montana. M.S. Thesis. Univ. of Montana. Missoula.
82pp.

SVOBODA, F.J. and G.W. GULLION. 1972. Preferential use
of aspen by ruffed grouse in northern Minnesota. J.
Wildl. Manage. 36:1166-1180.

SWENSON, J.E. 1985. Seasonal habitat use by sharp-tailed
grouse ( Tympanuchus phasianellus) in mixed grass
prairie in Montana. Can. Field Nat. 99:40-46.

TAUTIN, J. 1980. Assessment of some important factors
affecting the singing-ground survey. Proc. Woodcock
Symp. 7:6-11.

TAYLOR, M.W., C.W. WOLFE, and W.L. BAXTER. 1978.
Land-use change and ring-necked pheasants in Ne-
braska. Wildl. Soc. Bull. 6:226-230.

TOMLINSON, RE. 1972a. Current status of the endan-
gered masked bobwhite quail. Trans. North Am.
Wildl. Nat. Resour. Conf. 37:294-311.

. 1972b. Review of literature on the endangered

masked bobwhite. U.S. Dep. Inter. Fish and Wildl.
Serv. Res. Pub. 108. Washington, DC.

Chairman. 1983. Pacific flyway management plan

for the Pacific coast band-tailed pigeon. Pacific Flyway
Council. U.S. Dep. Inter., Fish and Wildl. Serv. Port-
land, OR. 21pp.

TRIPPENSEE, RE. 1948. Wildlife management. McGraw
Hill, New York. 499pp.

WALLESTAD, R. 1975. Life history and habitat require-
ments of sage grouse in central Montana. Montana
Dep. Fish and Game. 66pp.

and P. SCHLADWEILER. 1974. Breeding season

movements and habitat use of male sage grouse in
central Montana. J. Wildl. Manage. 38:634-637.
-, J.G. PETERSON, and R.L. ENG. 1975. Foods of

adult sage grouse in central Montana. J. Wildl. Man-
age. 39:629-630.

WEEDEN, R.B. 1963. Management of ptarmigan in North
America. J. Wildl. Manage. 27:673-683.

. 1965. Grouse and ptarmigan in Alaska: their ecol-
ogy and management. Mimeo. Alaska Dep. Fish and
Game. 110pp.
-. 1967. Seasonal and geographic variation in foods

of adult white-tailed ptarmigan. Condor. 69:303-309.

WEIGAND, J.P. 1980 Ecology of the Hungarian partridge
in north-central Montana. Wildl. Monogr. 74. 106pp.

and R.G. JANSON. 1976. Montana's ring-necked

pheasants — history, ecology and management. Game
Manage. Div., Montana Dep. of Fish and Game. Hel-
ena. 178pp.

WELLER, M.W. 1978. Management of freshwater wetlands
for wildlife. Pages 267-284 in Good, R., D. Whigham,
and R. Simpson, eds. Freshwater Wetlands. Academic
Press. NY. 378pp.

YEAGER, L.E., WW. SANFORT, and L.J. LYON. 1951. Some
problems of pheasant management on irrigated land.
Trans. North Am. Wildl. Nat. Resour. Conf. 16:351-
367.

428

Upland Game Birds

21

RODENTS AND
INSECTIVORES

Mayo W. Call

P.O. Box 893
Afton, WY 83110

Editor's Note: Although mamma logists have been
studying small mammals for years, only recently
have management agencies begun to include ro-
dent and insectivore data in their wildlife surveys.
These recent attempts have resulted in many ineffi-
cient and unproductive efforts. Surveys of rodents
and insectivores may be important in land man-
agement decisions especially in areas where (I)
threatened or endangered populations are or may
be present, (2) rodents and insectivores indicate
habitat changes, (3) rodents and insectivores are
important as a prey base for animals higher on the
food chain such as raptors, (4) the larger more
economically important species such as beaver and
muskrat are involved, and (5) rodents and insecti-
vores are locally important as pest species. These
situations usually require careful planning and
designing efforts. This chapter provides guidance on
planning and designing studies for these diverse
groups of species.

INTRODUCTION

Rodents and insectivores (shrews and voles)
live in almost every kind of habitat in North Amer-
ica. Some are common and widespread in many
kinds of habitats, such as deer mice (Peromyscus
maniculatus), whereas others live only in certain
parts of the country or in specific, limited habitats.
For example, beaver (Castor canadensis) and mus-
krats (Ondatra zibethicus) are restricted to riparian
or other aquatic zones and their associated
vegetation.

Most rodents are a major source of food for
predators, such as birds of prey, coyotes (Canis la-
trans), badgers (Taxidea taxus), and others. Thus,
they need to be highly prolific and have some type
of defensive, elusive, or other protective behavior to
survive. Some of these protective mechanisms com-
plicate efforts to catch them or estimate their
numbers.

Before commencing to inventory or monitor
any species or population of rodents or insectivores,
the biologist should first define his or her objectives
for the study. The method used and intensity of the
work will be determined by the information that
is needed for making the resource management deci-
sions at hand.

Biologists should recognize the kinds of prob-
lems they will be faced with as they attempt to esti-
mate numbers and to analyze differences in species
or numbers from different habitats ( or the same
habitat and area) in different seasons or years. Many
variables can affect results; even the most intensive
studies, using the most sophisticated methods, may

Rodents and Insectivores

429

not provide as accurate estimates as might be ex-
pected because one or more assumptions are not
met (Schemnitz 1980).

Some of the models used for estimating rodent
numbers require an assumption that the population
is "closed," i.e., there is no recruitment or loss of
animals to the population during the sampling period
(White et al. 1982). This rarely happens in the real
world and the complicated procedures required
to adjust for probable changes in the population may
be difficult to use for most field biologists. Proce-
dures in this chapter will be geared toward simpler
methods; references are provided for those who may
need to justify more intensive studies.

Populations of many species of rodents and in-
sectivores fluctuate greatly in local areas at various
times of the year. High birth rates, high mortality,
invasion from surrounding areas, and emigration to
other areas must be considered in inventory and
monitoring studies. Some small mammal populations
are cyclic; they alternately irrupt and subside in a
more or less uniform manner between high and low
levels of density (Anderson et al. 1977). These popu-
lation fluctuations sometimes follow a general pat-
tern with respect to time (monthly, seasonal, annual)
but other microtine rodents also experience cycles
of 3- to 4-, 4- to 7-, and 9- to 10-year intervals, as
well as annual fluctuations (Speirs 1939; Elton 1942;
Dymond 1947; Keith 1963).

The causal mechanisms of cycles include biotic
as well as abiotic factors (Anderson et al. 1977).
Biotic factors are inherent in the populations them-
selves and in the interrelationships of different spe-
cies. These include disease, predation, food, and
physiological mechanisms. Abiotic factors are physi-
cal and chemical elements of the environment, in-
cluding moisture, winds, and solar radiation. An
example of the effects of abiotic factors on nocturnal
rodents in the Lower Colorado River Valley in Ari-
zona was provided by Anderson et al. (1977). The
authors stated that, although there was a general
decrease in rodent populations during the entire
study period (1974-77), there was significant intra-
specific asynchrony among the populations in differ-
ent vegetation types. There was also a significant
degree of interspecific asynchrony in population fluc-
tuations "which renders the task of evaluating habitat
difficult and subject to error unless carried out for
several years in various vegetation types" (Anderson
et al. 1977).

Numerous studies were reviewed by Jorgensen
and Smith ( 1974) relative to estimating the numbers
and densities of small mammals in unrestricted popu-
lations. They noted that these studies quickly ex-
posed some major difficulties that must be
accounted for to produce a single estimate for one
restricted location and that "the effort required often

challenges the necessity of the data, particularly
when one's interest is frequently focused on popula-
tions in much larger areas."

Biologists must consider these variables and
recognize limitations that may be placed on the data.
For example, what might appear to be a simple den-
sity calculation, based on a series of parallel trap
lines within a designated area of habitat, turns out to
be not so simple because one does not know the
extent of movement into or out of the trapping area
by the various species, factors affecting trappability
of individuals or species, times of night of greatest
activity of the different species, influences of subtle
changes in weather during the trapping period, sizes
of home ranges of the different species for that par-
ticular habitat, and many other factors.

Monitoring studies should be conducted —

( 1 ) at the same time of year as previous studies;

(2) in the same kind of weather (as nearly as
possible) and with approximately the same
amount of moonlight;

(3) in the same habitat; and

(4) with the same kind of data collected each
time.

There are few situations that require informa-
tion other than species occurrence and relative den-
sity of rodents and insectivores for Federal land
management programs. These include —

( 1 ) determination of presence and estimated
numbers of rare, endangered, or sensitive
species (either Federal or State);

(2) determination of species, estimated numbers
(biomass), and trends of small mammals that
have been serving as a prey base for raptors
or carnivores whose populations are critical
or declining;

(3) determination of estimated numbers or
trends of a species of economic or scientific
importance (such as beaver and muskrat);
and

(4) determination of effects of environmental
contaminants on either the animals them-
selves or on dependent predators.

These are only guidelines, however, and the biologist
must ultimately decide and justify what information
should be obtained to help land managers maintain
the ecosystem desired in the area of interest.

In this chapter, I will emphasize a few species
that are important in land management programs

430

Rodents and Insectivores

because of their economic, aesthetic, or scientific
values. I will give beaver special attention because
they have the ability to modify their habitat more
than any other rodent and create changes that affect
many other species of wildlife.

HABITAT FEATURES CORRELATED WITH
SPECIES GROUPS

Because the species included in this chapter
occupy virtually every kind of habitat on the conti-
nent, it will not be possible to discuss individual
species habitats. I will, however, discuss general hab-
itat types occupied by the different families of ro-
dents and insectivores.

Beaver (Family Castoridae)

Beaver are found in most parts of the U.S. In the
West their preferred foods and, therefore, habitat
types include aspen (Populus sp. ), willow (Salix
sp.), cottonwood {Populus sp.), alder (Alnus sp.),
and other deciduous plants (Call 1970). For short
periods of time, beaver may subsist on conifers or
some of the less preferred deciduous species, such as
hackberry (Celtis sp.) and hickory (Carya sp.), but
they will not thrive on them.

These animals are most often found along low
gradient streams where preferred foods are available
within about 80 m (80-90 yd). This permits more
surface area to be impounded with a given amount

of beaver dam construction. Valley grades of 1 to 6%
are considered excellent for beaver; those from 7 to
12% are good, and those beyond 12% are question-
able for occupation (Rutherford 1964). Beaver are
sometimes found occupying small springs, streams,
or seeps on timbered mountain slopes (Call 1970).

Muskrat (Family Cricetidae)

Muskrats are found in almost any kind of flowing
water or in any kind of impoundment, natural lake,
pond, or marsh from sea level to high mountain
areas. Their nests may be in holes in stream banks or
in piles of vegetation and debris constructed in
ponds. Their presence is tied largely to the presence
of suitable aquatic vegetation which they use for
food and nest construction.

Moles (Family Talpidae)

Moles {Neurotrichus sp., Scapanus sp., Parascal-
ops sp., Scalopus sp. and Condylura sp.) live most
of their lives beneath the ground's surface, usually in
the upper dirt layers where digging is easy. Their
presence may be detected by the low ridges they
push up as they move along just beneath the surface;
also, by the earth mounds, each consisting of 0.5 to
2 ft of earth, which they push up from beneath.
There is no indication of an entrance to the burrow
(Burt and Grossenheider 1964).

Beaver dam.

Beaver lodge.

Rodents and Insectivores

431

Shrews (Family Soricidae)

Shrews usually inhabit moist locations, but some
species are found in sagebrush areas of arid western
States. They are common along most streams and are
usually found in moist areas along streams, under
logs, and in thick vegetation of seepages.

Squirrels (Family Sciuridae)

Marmots (Marmota sp.), ground squirrels (Sper-
mophilus sp.), prairie dogs (Cynomys sp. ), and chip-
munks (Tamias sp.) nest in burrows in the ground
or beneath rocks or logs. Marmots are most often
found living in boulder fields on talus slopes but may
also live in small rock outcroppings, under isolated
boulders on hillsides, in piles of logging debris on
hillsides or in gullies, or under abandoned houses or
within bridge abutments.

Tree squirrels (Sciurus sp. ) and flying squirrels
(Glaucomys sp. ) nest in tree cavities or leaf and
grass nests in tree canopies. The red squirrel (Tam-
iasciurus hudsonicus) and northern flying squirrel
(G sabrinus) are the most common tree squirrels in
the West. Red squirrels are found mostly in pine
(Pinus sp.) and spruce (Picea sp.) or mixed hard-
wood forests and in swamps. The presence of red
squirrels in an area is usually noted quickly by an
observer who walks through a forested area by
either the presence of piles of shucked cones or by
the sharp "barking" of the animals when intruders
come near.

Chickarees (T. douglasii) have habits similar to
red squirrels but live almost exclusively in conifer-
ous forests. The Arizona gray squirrel (S. arizonen-
sis) is found chiefly in oak (Quercus sp. ) and pine
forests. Tassel-eared squirrels (S. aberti) are found
mostly in yellow pine (P. ponderosa) forests of

northern Arizona and New Mexico, southeastern
Utah, and southwestern Colorado.

'^^'auA /&<!/ ^jLoOt^y-L^-

432

Rodents and Insectivores

Pocket Gophers (Family Geomyidae)

Pocket gophers (Thomomys sp., Geotnys sp.)
are true earth burrowers, seldom seen aboveground.
Their presence is easily detected by the mounds of
earth they push out as they excavate subterranean
tunnels. The mounds are characteristically fan-shaped
with an indication of the position of the opening, a
round earth plug. The burrow entrance is seldom
left open for long. They prefer soil that is slightly
moist and easy to work (Burt and Grossenheider
1964).

Pocket Mice, Kangaroo Mice, Kangaroo Rats
(Family Heteromyidae)

Pocket mice (Perognathus sp. ), kangaroo mice
(Microdipodops sp. ), and kangaroo rats (Dipodomys
sp.) are all adapted for living in arid or semiarid
conditions and all burrow into the ground for nest
sites. They prefer the more pliable, sandy soils.

Mice, Rats, Voles, and Lemmings (Family
Cricetidae)

Mice (Reitbrodontomys sp., Peromyscus sp.,
Onychomys sp., Sigmodon sp.), New World rats
(Sigmodon sp., Neotoma sp. ), and voles (Cletbriono-
mys sp., Phenacomys sp., Microtus sp., and Lagurus
sp. ) live mostly on the ground, some in rocks, and
a few in trees. Grasshopper mice (Onchomys sp. )
live chiefly on prairies and southwest desert areas.
Harvest mice (Reitbrodontomys sp.) seem to prefer
rather dense, low vegetation. White-footed (P. leuco-
pus) and deer mice are nocturnal, living in woods,
prairies, and around rocks and buildings. Most are
ground dwellers, but some nest in trees and are
largely arboreal.

Woodrats (Neotoma sp. ) normally occupy the
pinyon-juniper (Pinus edulus-Juniperus sp. ) plateau
region, frequenting rocky places and even cliffs
(Hoffmeister and Luis de la Torre I960), but some
forms will be found in Southwest deserts and others
in mountain brush of the Northwest.

Woodrats are active mostly at night. They pile
sticks and rubbish together for nests in cliff holes or
on ledges. In the plains and deserts, they build stick
and cactus nests, often 1 to 2 m (4 ft or more) in
diameter, in clumps of cactus, yucca, or brush. On
the West Coast their nests may be found in live oak
trees.

Various species of lemmings (Synaptomys sp. )
may be found from bogs in the eastern U.S. to arctic
and tundra habitats of the far north. The bog lem-
mings (S. borealis) prefer wet bogs and meadows
where there is a thick mat of ground vegetation
(Burt and Grossenheider 1964).

Deer mouse.

Redback voles (Clethrionomys sp.) are small
rodents of forested areas in Alaska, Canada, and scat-
tered forested areas of the U.S. Some of the meadow
voles and other species may be found in most re-
gions where there is a good grass cover, especially
where it is moist. At least one species, the Townsend
vole (Microtus townsendii), inhabits moist fields
and meadows from tidewater up to alpine meadows.
The presence of narrow runways, 2.5 to 5.0 cm ( 1
to 2 in. ) wide, through the matted grass usually indi-
cate their presence. In a few places, voles are found
among rocks or on forest floors where there is no
grass.

Phenacomys sp. may inhabit grassy areas within
the cold forested regions of Canada and Alaska, the
high mountain tops of western U.S., or dense for-
ested regions of the Northwest. Most are ground-liv-

ing, but the tree Phenacomys feed almost entirely on
leaves of the tree in which they are living — spruce,
hemlock (Tsuga sp), or fir (Abies sp).

Jumping Mice (Family Zapodidae)

Jumping mice (Zapus sp. and Napaeozgapus
sp. ) normally inhabit damp meadows and forests and
hibernate during the winter.

Porcupines (Family Erethizontidae)

Porcupines (Erethizon dorsatum) may be seen
lumbering through the forest, along canyon bottoms,
or hunched into large dark balls high in trees. They
may be found in deciduous or coniferous forests and
strangely, sometimes in shrublands or prairies, miles
from the nearest trees.

Porcupines sometimes hide in trees by curling into a large
dark ball.

Old World Rats and Mice (Family Muridae)

This group includes the Norway rat (Rattus
norvegicus), black rat (R. rattus), and house mouse
(Mus musculus). None of these animals normally
occur far from man-made structures and will not be
considered in this chapter.

POPULATION MEASUREMENT TECHNIQUES

With the variety of rodents and insectivores
present in most habitats, more than one method is
usually required to determine species present, rela-
tive density, or estimated numbers. Space is not
available here to present the many variations in ro-
dent/insectivore behavior as may apply to sampling
techniques. In fact, there remains much to be
learned about small mammal home ranges, behavior
in different seasons, cycles, responses to traps, and
other aspects of their ecology that affect surveys.

Rodents and Insectivores

433

Volumes of material have been written about
statistical procedures that may be used in sampling
wild populations. Depending on their needs for ani-
mal population data, biologists must decide if statisti-
cal procedures are needed and refer to the Literature
Cited section for detailed accounts of the necessary
applications. Not every sampling of rodent or insecti-
vore populations needs to be analyzed by elaborate
statistical procedures or complicated sampling pro-
cedures. Much useful information can be derived on
species, relative density, and trends from the raw
data itself, together with simple statistics such as
means and variances.

General Considerations

Before starting any project, the biologist
should —

( 1 ) determine the objective (what kind and how
much information is necessary);

(2) determine what kinds of inventory methods
or sampling will be required; and

(3) design the study.

For most land management programs, only infor-
mation on relative density or species trends will be
required for rodents and insectivores. At times, how-
ever, the biologist may decide that additional data
are needed on either a specific species or total popu-
lations of several species. Computer programs are
available to assist with the more difficult computa-
tions (Schultz 1961; Snedecor and Cochran 1967;
Sokal and Rohlf 1969; Petrusewicz and MacFayden
1970; Seber 1973; Anderson et al. 1976; Burnham et
al. 1980; White et al. 1982).

For most rodents and insectivores, counts of
individual animals are a practical impossibility.
Therefore, one normally takes a sample from a popu-
lation and attempts to derive a reasonable estimate
of the total population from that sample. The esti-
mate for the total population is then based on the as-
sumption that all animals in the population occur
throughout the defined study area in the same ratio
as they did within the sampled area.

An index may be employed for monitoring a
population. An index is an object (such as a beaver
lodge) that is related to the number of animals in
the area. The object is usually easier to count than
the animal itself. For example, the number of mus-
krat houses or beaver lodges is an index to the num-
ber of animals living in the area. Through intensive
trapping, or by using values derived from other re-
searchers in the general area, one can estimate the
number of animals in the area by multiplying the
number of houses or lodges by the normal number

of occupants per house or lodge. One of the prob-
lems in using an index is the necessary assumption
that the ratio of the index to the population is the
same in the several areas or populations being com-
pared (Schemnitz 1980).

The basic premise used in rodent and insecti-
vore monitoring is that a similar proportion of the
total population is tabulated each time the same
technique is used under similar field conditions with
the same species. No statistical procedure will make
data collected under different conditions compara-
ble. With all techniques for rodents and insectivores
that employ direct counts, the major problem is to
conduct counts under sufficiently similar conditions,
so that comparison of counts is valid (Schemnitz
1980).

Most populations of rodents and insectivores are
not distributed randomly. Each species will tend to
be more abundant within small microhabitats that
commonly occur within a general habitat type, even
when the general habitat appears to be homogene-
ous. Cover and food are not usually distributed ran-
domly, so neither are the animals. Some burrowing
rodents prefer certain soil textures and types for
digging their tunnels, and these soils will vary some
in what appears to be a fairly homogeneous vegeta-
tion type. Thus, using the same transect site for mon-
itoring in subsequent years, or selecting sites for
comparative purposes, which are as nearly alike as
possible, may reduce variances.

Sampling for Large to Medium-Sized
Rodents

Because of the diversity of rodents and insecti-
vores, a number of different survey techniques are
required in any given area to determine all species
present, their relative density (abundance), or esti-
mated numbers. I will discuss some of the general
methods suitable for most species and provide addi-
tional references for more sophisticated techniques.
In some instances the biologist may need to devise
innovative methods to determine presence or esti-
mated numbers of endangered or rare species or
animals that are trap-shy or elusive in other ways.

Beaver.

Presence. The presence of beaver along a
drainage is determined by looking for beaver dams,
lodges, or beaver cuttings (Figure 1). Along desert
streams and larger rivers, beaver live in bank dens
and do not build dams, but their tree- or shrub-
cutting activities will be obvious. Most stream
drainages in western mountains that contain aspen,
willow, or cottonwood will contain beaver.

454

Rodents and Insectivores

Population Estimate. The average number of
beaver per colony for the entire U.S. and Canada is
about five (Denney 1952). Where beaver trapping
occurs over widespread areas, the normal colony
composition is disrupted. Although this makes it less
precise to use a normal average number for popula-
tion estimates, the method described will suffice
for most management purposes.

The suggested population estimate method is to
cruise streams in October or November, counting
the active colonies (as evidenced by food caches and
repairs on dams and visible lodges) and multiplying
the number found by five. Food caches are easily
discernible and are the most obvious indication of a
wintering beaver colony. The size of the food cache
is some indication of the number of beaver in the
colony (Figure 2). However, there is no direct rela-
tionship between the size of the food cache and
the number of beaver in the colony because ( 1 ) dif-
ferent colonies may use different kinds of food mate-
rials in their caches and ( 2 ) some colonies will have
much more readily available food (aspen, willow,
etc. ) adjacent to their primary pond. In a few places,
such as western Oregon, beaver may not use food
caches but will cut the needed food all winter long.
In such areas, the beaver can be surveyed by count-
ing the number of individual cutting areas along
the river which should be concentrated near their
bank dens (C. Maser, pers. commun.).

Even where colonies do not have a normal
makeup because of trapping each year, this method
will be satisfactory for most management purposes. A
few colonies are usually overlooked and a few bank-
dwelling beaver are usually present but not dis-
cerned, which will largely make up for a reduction
in the average number of beaver per colony caused
by trapping.

If the task of determining the number of beaver
involves many miles of streams or relatively inacces-
sible drainages, the best method to use is an aerial
survey, either by fixed-wing or helicopter. The ter-
rain will dictate the type of aircraft to use. It is best
to fly within 90 to 151 m (300 to 500 ft) of the
ground so that all food caches may be seen and
counted. Such food caches can be seen on aerial
surveys, even when ice has formed on the ponds
(Figure 3). This is not difficult in a fixed-wing air-
craft along wide valley bottoms with low stream
gradients, but becomes more challenging in rough
terrain.

Muskrat.

Presence. Muskrats are noted throughout most
of their geographical distribution for their conical
or dome-shaped houses of vegetation set in marshes
or in shallow pond waters. However, in many areas,

Figure 1. A two-lodge beaver pond in southeastern
Wyoming. This pond is about 1 acre and was
about 20 years old at the time this picture was
taken.

Figure 2. The same beaver pond as shown in Fig-
ure 1, as viewed from the air. Food caches, not
present in this photograph, are usually built
adjacent to dens or lodges, and are easy to dis-
cern from this altitude.

Rodents and Insectivores

435

Figure 3. A relatively large food cache in a beaver
primary pond. Eleven beaver were trapped from
this colony.

especially along ditches and rivers, they dig bank
burrows. Some marshes contain dozens of the dome-
shaped houses that leave no doubt as to the
occupants, but detecting the presence of muskrats
that live along streams is more difficult. Burrow
entrances are usually underwater. Their presence in
such areas can best be detected by watching for
cuttings of cattails, rushes, sedges, small willows,
pondweeds, water lilies, and other aquatic and
terrestrial plants on which they feed. They are active
during the day and may occasionally be seen
swimming slowly up or down the stream or sitting
quietly on a sandbar, partly submerged tree trunk, or
other convenient object where they may sun
themselves and still be close to water for a quick
escape from danger.

Population Estimate. Estimating numbers of
muskrats in an area is very difficult. The number
of houses in a marsh is only an index to the number
of muskrats.

One method for monitoring trends of these
animals is to question local trappers concerning their
harvest from local habitats (C. Maser, pers. com-
mun.). For example, if trappers believe they have
harvested about 60% of the existing population of a
marsh or stream and have taken 70 muskrats, then
the population at the beginning of the trapping pe-
riod would have been about 1 17 animals. Regardless
of trappers' estimated percentage of harvest, trends
in the annual harvest would give an indication of
trends in the population level. The basic assumption

is that prices and, therefore, interest in trapping,
make it worthwhile for trappers to conduct about
the same intensity of trapping over the years.

A mark and recapture method, using the Lincoln
Index, could be used to estimate the population but
would be time-consuming. High mortality from pre-
dators could distort the results if too much time
were consumed in live trapping and marking a num-
ber of animals.

Prairie Dog.

Presence. The black-tailed prairie dog
(Cynomys ludovicianus) is especially gregarious,
i.e., it lives in "towns." Its presence is usually
revealed by a group of bare mounds from 8 to 23 m
(25 to 75 ft) apart and each mound from 0.03 to
1 m(l to 3ft) high (Burt and Grossenheider 1964).
Colonies of several hundred animals are not
uncommon on some of the dry upland prairies on
which they live. Populations may vary from 12 to 85
per ha (5 to 34 per a.) or more.

Some white-tailed prairie dogs (C. leucurus) are
widely scattered across sagebrush flats and rolling
grasslands. For family groups that occur in habitats
where they are largely concealed, their presence
is often revealed by their danger signal — a two-sylla-
ble bark, issued at about 40 per minute.

Utah prairie dogs (C. panndens) have been ob-
served in both dense and scattered colonies. I have
counted more than 500 prairie dogs in a pasture
of about 4 ha ( 10 a.) near Cedar City, Utah.

Population Estimate. Prairie dogs live in fam-
ily groups and exhibit some territoriality around
their dens but may use communal feeding areas

436

Rodents and Insectivores

away from the dens. An estimate of numbers, prefer-
ably when their population is somewhat stabilized
in the fall, can be obtained by the mark and recap-
ture method using Lincoln Index calculations (see
small mammal sampling). However, this procedure is
very time-consuming.

I recommend that inventorying and monitoring
of these animals be based on the number of active
dens, rather than on number of animals observed at
any given time. Areas containing prairie dogs should
be mapped out and then a grid established over the
colony area to facilitate the counting of burrows.
Burrows can be plotted on the map for comparison
of numbers in future years. Stakes or prominent
shrubs can be used as corners of the grid and
brightly colored flagging used for easy reference
while walking and counting burrows. Additional flag-
ging should be placed along the edges of the colony
at intervals of about 10 m (30 yd) along each side to
effectively cut the colony area into squares so that
burrows can be counted and plotted on a map. Grid
square numbering can be done the same as that for
the International Bird Plot Census Method, or in any
other convenient manner for future comparisons.

Marmots.

Presence. Presence of these animals is usually
revealed by their sharp chirps when intruders
approach. The den area is rank with the odor of
their feces. By sitting near a rockslide in early
morning and watching for activity, the observer can
usually determine the presence of these animals
without doing any trapping. Watching an area of
interest for a few hours will generally give the
observer a rough idea of how many of the animals
are present. Both young and old animals like to sun
themselves on rocks or logs near their dens.

Population Estimate. Most colonies of these
animals are small. A mark and recapture method can
be used for estimating numbers by marking some
of the animals with easily visible tags and then
spending a period of 3 or 4 days observing the ani-
mals and tallying all marked and unmarked animals.
Lincoln Index procedures can then be used to esti-
mate total numbers present.

Woodrats.

Presence. Presence is revealed by large stick
or cactus houses in clumps of brush, cactus, yucca,
or in crevices in cliffs. They may also take over
abandoned cabins or other old buildings in foothills
and canyons where they build large nests containing
rocks, bones, sticks, leaves, and other debris. Nests
built on ledges of cliffs or rocky outcroppings are
sometimes confused with nests of ravens or birds of

prey. But the inclusion of bones or fragments of
human refuse, including pieces of glass or other
shiny objects, usually identify it as a woodrat nest.
There may be from one to six, or more, rats per
nest. The smelly urine stains on rafters or narrow
cliff ledges near their nests also reveal the identity of
the occupants.

Rocky slopes are home for the yellow-bellied marmot.

Population Estimate. Colonies of woodrats
are usually widely scattered. Rough estimates of their
abundance may be obtained by walking transects
along the desert floor or canyon hillsides while
watching for their large stick nests which are often
1.5 to 2 m (4 to 6 ft) in diameter and 1 m (3 ft) tall.
When located, they can be checked for tracks and
fresh material to determine if they are active. Rela-
tive density of this animal can be determined along
with other small mammals through the periodic use
of trapping transects (see section on small mammal
sampling).

Tree Squirrels. The tree squirrels, including the
red squirrel, western gray squirrel (5. griseus), fox
squirrel (S. niger), tassel-eared squirrels, and
northern flying squirrels, can usually be identified by
a combination of region of country where found
and the signs which they leave.

Presence. They usually have favorite stumps or
racks where shucks from pinecones or nuts may
accumulate in piles of a bushel or more. Their nests
are either in cavities in trees, on outside branches
built of leaves, twigs, and shredded bark, or some-
times in cavities at the bases of trees. Conifer cones
and nuts are stored in caches under logs or in cavi-
ties in the ground or among tree roots. They are
active mostly in daylight hours.

Rodents and Insectivores

437

There are seldom more than two chickarees or
red squirrels per ha (per 5.0 a.) of forest, but there
may be as many as 10 per 0.4 ha (1 a.; Burt and
Grossenheider 1964). Chickarees are fairly quiet dur-
ing spring and summer while raising their young
but become vociferous in fall. Thus, fall is the best
time to check for their presence (Maser et al. 1981).
Nests may be bulky structures built of twigs, mosses,
or lichens, well within the crown of the tree or in
hollows of trees.

Northern flying squirrels are strictly nocturnal.
They usually nest in tree cavities but may also build
outside nests of leaves and twigs. Normally, nests are
round balls of material situated in a fork of the trunk
of a tree or on a whorl of limbs against the trunk
(Maser et al. 1981). In most areas, few people realize
how abundant these night-gliders really are. Maser
et al. (1981) stated that trappers who trap marten
sometimes catch hundreds of these squirrels each
winter in traps. These animals are fairly quiet and
generally must be detected by observations at night
in suspected habitats or caught with traps baited
with fresh, raw, or putrid meat, which they seem to
relish (Jackson 1961). Much of their diet consists
of green vegetation, insects, nuts, seeds, and fruits
and they leave few signs of their feeding.

The Arizona gray squirrel (5. arizonensis) nests
either in tree cavities or in tree branches. It stores
nuts and seeds singly in small holes in the ground
rather than in large food caches. The squirrels will
cling to the opposite sides of trees or lie quietly
along a branch and never reveal themselves as long
as an intruder is moving near them. Visual observa-
tions are the best means for detecting presence of
these animals and for estimating numbers. Two per-
sons working together, but a short distance apart, are

effective in locating the animals. Depending on the
habitat, there may be from 2 to 20 squirrels per 0.4
ha (1 a.; Burt and Grossenheider 1964).

Tassel-eared squirrels build bulky nests high in
the pines. They are usually found from 2,121 to
2,424 m (7,000 to 8,000 ft) in elevation. They may
be heard "barking" when excited but are usually
quiet. Their presence is best detected by listening
for their calls and observing them in appropriate
habitats.

Porcupine.

Presence. The presence of porcupines is most
often determined by watching for areas of bark
removal in coniferous trees and bushes or in winter
by finding their waddling trails in the snow. They are
often observed at great distances in winter as dark
balls in the tops of bushes or deciduous trees as they
feed on the bark. Their brownish droppings,
consisting of woody materials, are very
characteristic. They often have a home den where
they may "hole up" during bad weather or before
the birth of the single young. The den may be a
small cave, crevice in a cliff or among boulders and
talus, or in a hollow log. Their feces accumulate in
such places and reveal their presence.

Population Estimate. Porcupines seem to be
scattered sparsely over most areas. Some estimate of
numbers can be obtained by running a 1.6- to 32-
km (1- to 2-mi.) transect through open forest or
mixed deciduous-conifer forest types while watching
for signs. Being fairly sparse, they would be easy to
miss entirely. Driving unimproved roads through
forested areas at night would probably reveal more
porcupines than daytime observations. Late one
night in the fall of 1950, I counted 21 porcupines
along 32 km (20 mi.) of Forest Service road in west-
ern Wyoming. I doubt I could have found half as
many during daylight hours.

Caution should be used in trying to estimate
numbers of porcupines per unit of land area because
the animals will sometimes concentrate along can-
yon bottoms or other desirable habitats.

Sampling for Small Rodents and
Insectivores

Inventorying of most small rodent and insecti-
vore species can best be accomplished with trapping
transects. Subsequent monitoring of species and pop-
ulations should be done at the same sites and with
the same methods as the original inventory. All trap
lines will provide information on species present and
relative density; if certain designs are used, one can

438

Rodents and Insectivores

also calculate an estimate of the populations in-
volved. However, density is a more complicated
problem and if such data are needed, the biologist
will need to research literature that describes the
required field designs and statistical analyses (Smith
et al. 1971a; Smith et al. 1972; Smith et al. 1975;
O'Farrell et al. 1977; Scott et al. 1978; Burnham et
al. 1980; White et al. 1982). The degree of accuracy
of either population or density estimates will depend
largely on meeting certain assumptions. The follow-
ing procedures can be used for trapping small mam-
mals to obtain species, relative density, and
estimated numbers — but not density.

Species Occurrence and Relative Density. For

this information, I suggest following the procedures
used by Anderson et al. (1977) and Kepner (1978):

( 1 ) Be sure the vegetation communities are de-
scribed so that relationships between plant
and animal communities can be suggested.
Soil types should also be described.

(5) Tabulate and analyze your results. Catches
will normally be expressed as numbers of
each species per 270 trap nights (if the above
system is used). If the same type of trapping
grid is used in different communities, com-
parison between species and relative density
can be made for different communities. It
will soon become obvious which communi-
ties are most important in terms of total spe-
cies and numbers. These communities,
however, may not be the ones where rare or
indicator species are found that will require
special management, and you will need to
sample the preferred habitats used by these
species.

Live traps suitable for taking medium-sized ro-
dents, such as squirrels and woodrats, may be substi-
tuted for the Victor rat trap, if desired (Figure 5).
However, if this substitution is made, the animals
captured should be removed from the area or kept
in captivity until all trapping is completed.

(2) Establish sampling grids within each vegeta-
tion community where data are needed. Each
grid will consist of snap trap stations placed
along two parallel lines of traps 15 m (50 ft)
apart (Figure 4). Each line will consist of
15 stations that are also 15 m apart. At each
station, set two museum special traps and
one Victor rat trap. Bait them with rolled oats
and peanut butter with a chemical mixed in,
such as dimethyl pthalate, to repel insects
(Anderson and Ohmart 1977).

(3) Set the traps in early evening and check them
every hour until midnight, if possible. Set
them again at midnight and leave until morn-
ing, then check them, remove the animals,
and leave the traps sprung until evening.
Checking the traps every hour before mid-
night will remove most of the easily caught
animals, such as deer mice, and leaving the
traps set will increase chances to catch less
abundant or more trap-shy animals for the
remainder of the night.

(4) Run the trap lines for three consecutive
nights. Tabulate all catches by species, sex,
and age group (sub-adult and adult). In moni-
toring a small mammal population, recognize
that animals are caught differentially by spe-
cies, sex, and age, and the catch will depend
largely on season of the year. In any future
monitoring of the same community, be sure
to trap at the same season to obtain compara-
tive results.

25

20

| rock

15

shrub | V3

^r

1p f° in

10

I trap

5

Enlargement of Station 10

1

Figure 4. Diagram of a simple linear trap line with
an enlarged view of one station (No. 10) to
show an example of trap placements.

Rodents and Insectivores

439

L * W "P *»,"

wmmmwzn

-J*. Ji

ncnrwarrzrm

Figure 5. Live traps are efficient for a wide variety of small mammals. This bushy-tailed woodrat (Neotoma
cinerea) was captured in western Wyoming using peanut butter as bait.

Trap placement at each station may be biased
by placing them in obvious runs, near other rodent
signs, or in preferred microhabitats. Small rodents
and insectivores have specific microhabitat prefer-
ences (e.g., washes, certain kinds of bushes or trees,
rocky areas, rotten logs, wet areas). Although all
animals move around an area within certain inherent
limits or home ranges, it is best to place traps within
or adjacent to these preferred sites. The objective is
usually to catch as many as possible of each species
present in each vegetative community during the
predetermined trapping period. Setting traps 2 to 4
m (6 to 13 ft), or more, from the centerline of the
transect to improve the chance of taking additional
specimens is suggested.

Some species of small mammals, such as squir-
rels and some mice and voles, are largely arboreal.
Consequently, they may not be captured easily in
the usual ground surface trap lines. To collect these
species, place live or small traps on platforms in
trees or fasten them to limbs. Traps may also be
placed on the ground around the base of the trees
where evidence of the animals is found. Red tree
voles (Arborimus longicaudus) in Oregon are
strictly nocturnal and cannot be trapped. They must
be evicted from their arboreal nests, often as high
asl5m(50ft) above the ground, and captured by
hand (Maser et al. 1981). Small mammals observed
during the daytime may be shot if the specimens are
needed. Special Conibear traps, which catch the
animals around the neck or chest, killing them al-
most instantly, are useful for taking animals ranging
in size from squirrels to beaver (Collier 1957).

Gophers and moles are almost entirely fossorial.
Gophers may be captured with Macabee gopher
traps which are available at many hardware stores.
The traps should be set in all gopher tunnels leading
from one or more mounds and should be staked
down to prevent gophers from pulling them into
their burrow complexes. Occasionally, gophers are
captured in snap traps on the ground surface.

Moles may be captured with gopher or special
mole traps. A gopher trap set upside down in the
mole's burrow is often effective (Ingles 1965). Spe-
cial commercial mole traps consisting of a box-like
housing, which opens at the ends and has an impal-
ing device actuated by a trigger mechanism, are also
useful in obtaining specimens for identification
(Giles 1969).

Shrews and other insectivores are rarely cap-
tured with traditional trapping methods for small
mammals. Because they are insectivorous, they are
often not readily attracted to bait composed of pea-
nut butter or oatmeal. One of the most effective
methods for collecting them is to use pitfall or can
traps, which are also effective in catching amphibi-
ans, reptiles, and other small mammals. For best re-
sults, the cans or buckets should hold more than
1 gallon and have a round aperture in the center,
which helps reduce the possibility of escape. How-
ever, regular ice cream or pickle buckets or cans
will also serve the purpose well.

Cans or buckets should be buried in the ground
so that the tops are flush with the substrate (Figure
6). For best results, the pitfall traps should be in-

440

Rodents and Insectivores

stalled and then covered with a lid for about 10 days
to allow for accommodation by the resident animals.
Maximum trapping success for shrews does not oc-
cur until the traps have been in place for about 10
days (S. Cross, pers. commun.). Setting up a trapping
array with either three or four arms will capture
shrews in moist areas, as well as small mammals and
reptiles in arid areas.

The trapping arrays shown in Figure 6 are set so
that each can is connected by a drift fence of wire
mesh or tin, about 0.3 m (10 in.) high. The drift
fence should be set about 3 cm (1 to 2 in. ) into the
substrate. If the animals are believed to occur in
very low densities, the size or number of the sample
plots should be expanded and the area of survey
increased.

Because shrews move about very rapidly or
perhaps because they smell meat (or smell and hear
other small animals on which they feed) in the bot-
tom of the pit trap, they frequently drop into the pit

and cannot climb back out. If the pitfalls are baited
with sardines, earthworms, rodent brains or liver, or
other meat, more shrews may be obtained. If you
cannot examine the traps on a daily basis, 10 cm (4
in.) of formalin or AFA (Giles 1969), covered by a
thin layer of mineral oil, should be placed in each
can or bucket to retard decomposition of the speci-
mens. This will also help to prevent the shrews from
destroying or eating other trapped animals in the
pit. Some trappers fill each bucket about one-third
full of water when they are activated to accomplish a
similar purpose and to help keep the shrews from
jumping out (S. Cross, pers. commun.).

Shrews are frequently found near permanent
water sources. Therefore, the biologist should seek
out such areas to check for their presence. Shrew
sign is not easy to detect, but with some experience,
the worker can learn to recognize their runs, bur-
rows, and nests. Leaf litter, matted grass, and moist
ground are good habitat signs to look for.

^ Z3 it.

^— ' drift fenc

O

drift fenc

o

6

6

Figure 6. Configuration of pitfalls with associated drift fences to capture small mammals, lizards, and snakes.
Arrays with either three or four drift fences are effective. Buckets (insert) are placed flush with the
ground surface with a cover supported by rocks to provide cover and protection for the trapped animals.

Rodents and Insectivores

441

Another method for capturing shrews is to dig
small trenches between the pitfalls, instead of using
drift fences (Figure 7). Good success is often
achieved by laying out a grid of pitfalls (coffee cans
or other suitable containers) adjacent to small
streams or other moist areas (Figure 8). Bait is
placed in the bottom of the cans and along the
trenches; the trenches should be moistened so they
are damp. The trap grid should be checked once
during the night and again before sunrise. The traps
should be closed with plastic lids during the day
unless they are going to be visited frequently.

Care should be taken in handling live shrews as
some species secrete a toxin in the saliva that can
make a bite quite painful. I recommend using gloves
when handling them. Unless you intend to retain
them as specimens for identification, they should not
be kept in captivity longer than about 2 hours.

Most researchers seem to prefer a combination
of snap traps and live traps to obtain a representative
sample of all species present. Varying results have
been reported when only one type of trap (snap vs.
live) has been used for sampling. Live traps come
in different sizes. Common sizes for small mammals
are 3 x 3-5 x 9 in. and 2 x 2.5 x 6 in. Sherman live
traps are available from H.B. Sherman Traps, Aenon
Creek Industrial Park, Route 4, Box 529 X-2, Talla-
hassee, Florida 32304; telephone (904) 575-0424.

Permanently molded live traps seem to be pre-
ferred to collapsible ones. However, the number
to be set and the distance they need to be carried
will have a bearing on the type used. Also, rodents
are more apt to chew through aluminum live traps
and escape than they are with tin traps. Triggers on
collapsible traps need to be sensitized each time
they are set to be sure they will operate properly.

In-hand examination of small mammals is often
required for their identification. Certain members
of genera Dipodomys, Peromyscus, Perognathus, and
Neotoma are extremely difficult to identify in the
field and may require several individual animals to
confirm the identifications. Juveniles are particularly
difficult to identify and may require use of compara-
tive skulls and skins. Because of the lack of expertise
of most biologists in identifying subspecies (some
of which may be endangered or threatened forms), I
recommend that arrangements be made to send
specimens to qualified museums or universities for
identification. If this is done, be sure proper meas-
ures are taken to preserve the specimens during
shipment so they will be in good condition upon
arrival at the institution. Obviously, this will not be
necessary if qualified taxonomists are available in-
house or with cooperating agencies.

Traps should be checked as early as possible the
morning following a night's trapping. Many desert
areas heat up rapidly and dead animals will spoil
quickly in the heat. When using live traps, consider
placing a shingle (or board) on top of the trap with
a rock to hold the shingle on. This seems to improve
efficiency and provides some shade in early morning
heat. Also, I recommend using bedding material such
as wool (fleece) in live traps. When sampling for
diurnal rodents, set the traps as early in the day as
possible. Check them before noon and spring them.
Reset them in the afternoon and rerun the line be-
fore dark. Remove animals and spring the traps.

The terminal ends of each trap line should be
marked with stakes, flagging, or other visible mate-
rial. Permanent stakes are desirable for future refer-
ence; the locations of these markers should be
marked on maps so they may be relocated for future
monitoring studies. In forested areas or in heavy

ground level

X

K7

trench

1 - 2' deep

:X coffee can

Figure 7. Diagram of pitfall and associated trenches for catching shrews. The trenches should be slightly
moistened.

442

Rodents and Insectivores

brush, you may wish to run a cord through the habi-
tat and attach tags at appropriate places to facilitate
finding the traps. Strips of flagging material may be
tied to branches for identifying trapping stations.

Provide a thorough description of the site
where the trap line has been placed. Include the
following items:

• abundance of trees, shrubs, and grasses

• canopy coverage and height of the plants

• serai stage of the plant community

• litter depth and quantity; disturbance by live-
stock, fire, or man

• presence of logs, debris, rocky cover

• size of trees and species in forest situations

Include a brief description of topography (e.g., slope,
elevation, relation to drainage system, nature of the
soil, aspect). For future monitoring studies, photo-
graphs of habitat sites are helpful. Describe special

microhabitats where traps might have been placed
within the general habitat type and note differences
in species caught there. Make special note of the
weather conditions at the time the traps were in op-
eration and just preceding the trapping effort.

Variability is considerable in the daily and sea-
sonal behavior of rodents and insectivores, including
the time species are aboveground and available for
trapping. Some hibernate or estivate, whereas others
spend parts of the year underground where they
are active. Some species spend most or all of their
time in tree canopies during spring and summer,
then spend time on the forest floor in late fall and
winter.

Therefore, determine from the literature or local
experts when various species are expected to be
most active aboveground and concentrate on trap-
ping during those periods. Trapping for species oc-
currence and relative density during all four seasons
is desirable to obtain a more complete picture of
total species using an area.

cans

-=?fe

Figure 8. Grid pattern of pitfalls and trenches for collecting shrews along a stream. Trenches are intercon-
nected with pitfalls to guide shrews into the pits.

Rodents and Insectivores

443

Obviously, it will be difficult to obtain a true
relative density of all species present, from gophers
to tree squirrels. They will not have the same trappa-
bility nor be taken by the same kind of traps. For
example, if you do not set gopher or mole traps
where signs of these species are present along the
trapping transect, you may not get any in your sam-
ple, even though they may be fairly numerous. Spe-
cial traps or techniques, then, will be necessary if
you want a true representation of all species present
in a given habitat type.

Population Estimates. Three methods for
estimating numbers of animals will be discussed: ( 1 )
the strip census, (2) Lincoln Index or mark and
recapture, and ( 3 ) probability of capture method.
Any of these methods will give satisfactory results
when used for monitoring the same populations
during some future point in time, providing all
assumptions are met to the same degree as in the
original sampling.

(1) Strip Census. One method for estimating
the number of medium-sized rodents, such as tree
squirrels or ground squirrels, within a habitat type is
to count the number of animals observed within a
strip, transect, and then apply the results to the gen-
eral area. The area censused should be representa-
tive of the entire habitat type. Counts of several
similar areas or several counts of the same area pro-
vide numbers needed for calculation of an estimate
of total numbers for the species being surveyed.
Detailed versions of the strip census method can be
found in Anderson et al. 1976 and Burnham et al.
1980.

The strip census method entails following a
straight transect line or series of straight line seg-
ments (Anderson et al. 1976), counting the animals
observed, and recording the right angle distances
from the transect line to the animals observed. The
requirement to follow a straight line is usually com-
promised in the field because of physical obstacles
and the near-impossibility of traveling a straight line
and searching for animals at the same time. Marking
the transect line in some manner will improve the
ability to maintain a fairly straight line. This would
be done a few days before making a count.

Four assumptions are critical to achieve good
results (Anderson et al. 1976):

( 1 ) Animals directly on the line are seen with
probability of one (all seen).

(2) Animals are pinpointed at the initial sighting
position (they do not move before being
sighted ) and none are counted twice.

(3) There are no measurement errors (distances
are measured exactly).

(4) Animal sightings are independent events.

Obviously, one or more of these assumptions
may not be entirely met. Ground squirrels or other
burrowing animals may duck into a hole on or near
the transect line and not be seen at all, or may run
some distance before being observed. Many animals
respond to intruders by shying away or running
some distance under cover before flushing into the
open. This tends to exaggerate the right angle dis-
tance from the transect line to the animal when first
observed. Some animals run ahead of the observer
and, after being counted once, may be counted again
if they continue to run ahead. Also, the flushing of
one animal is not supposed to cause another animal
to flush, but it sometimes does. All these situations
tend to violate the assumptions. Nevertheless, if the
same procedures are used each time the population
is sampled, comparison of results will provide satis-
factory information on trends. As in all methods, a
replication along the same transect a few days later
or in a similar area will add credibility to the results
and provide a means to calculate standard deviation
and variance.

See Figure 9 for a diagrammatic display of a line
transect and the various measurements required.

(2) Lincoln Index or Mark-Recapture
Method. The first mark-recapture method for terres-
trial animals was known as the Lincoln Index and it
is still widely used (Seber 1973). In using the Lin-
coln Index, part of the animals in a population are
live-trapped, marked in some manner, and then re-
leased back into the population. Samples are next
captured to observe the proportion marked in the
entire population, which now consists of a known
number of marked animals along with an unknown
number of unmarked. An estimate of the total num-
ber is computed by dividing the total number
marked in the population by the proportion marked
in the samples, under the assumption that the sam-
ples will estimate closely the proportion marked
throughout the entire population.

There are three general methods for marking
captured animals: mutilation, tagging, and coloring
(Mosby 1963).

Mutilation. Forms of mutilation include toe-
clipping, ear-cropping, hole-punching, fur-clipping,
tail-notching (beaver), and branding. Any of these
forms of marking may receive public criticism in that
they maim the animal to some extent and possibly
affect its behavior or survival. The main advantages
of mutilation are that marks may be readily applied
with a minimum of equipment and the marks, in

444

Rodents and Insectivores

some instances, may be identified at a distance. Sys-
tems for marking small mammals by toe-clipping and
ear-punching have been described by several work-
ers (Blair 1941; Reynolds 1945; Sanderson 1961).
Two methods for numbering systems used in toe-
clipping are shown in Figure 10. Toe-clipping is a
widely used technique for both small mammals and
reptiles. A small pair of sharp scissors or toenail clip-
pers works very efficiently. Only the distal portion
of the animal's toe is removed. This is done system-
atically and recorded as a unique mark to identify
a specific animal.

Shearing, clipping, or burning off the tips of the
hairs can be used when the hair-base, or underfur,
is of a contrasting color, as it often is, such as in
newborn mice (Svihla 1934) and chipmunks (Euta-
mias sp.; Yerger 1953)- This method is useful only
until the next molt but has three advantages: it is
easy to apply, painless to the animal, and it consti-
tutes a good mark for field identification. On Norway
rats, Chitty and Shorten (1946) used a number
drawn on the pelage with a depilatory. Such marks
last 2 to 3 weeks, normally long enough for sampling
and making population estimates.

In the beaver, the flat, hairless tail has been
marked by branding (Bradt 1938) and by cutting

notches. Bradt found straight-line brands on the up-
per surface of the tail recognizable for at least 4
years. Cuts into the edge of the tail presumably last
for life (Mosby 1963).

Tagging. Tagging involves the attachment to
the animal of a piece of metal or plastic on which is
stamped an individual number. Advantages are that
they are easy to affix and to see. The tag should be
as light as possible and should be placed so that it
does not pinch the animal. It is desirable to place
tags low on the ear, where the cartilage is heavy, and
on the inner edge where there is greater protection
of the tag (Mosby 1963). In some mammals where
the ears are very small or there may be high losses
of tags through freezing or tearing out, tags may
be placed elsewhere.

Muskrats have been tagged by slipping a 5 by 24
mm aluminum tag through two slits in the skin of
the back (Errington and Errington 1937) and is most
successful with muskrats not over 2 months old,
due to the superior regrowth of blood vessels across
the strip of skin between the two incisions in young
animals (Errington 1944). This method might cause
infection if the water they live in is polluted (Takos
1943).

Figure 9. The unit to be censused has area A. The transect length is L and the transect width is 2W. Z is the
position of an observer when an animal or object is detected at X. P is the point on the line which is
perpendicular to the animal. The sighting distance is r„ the sighting angle is Oj, and the perpendicular
distance from the object to the center line of the transect is y;. W can be left unspecified or can be a
fixed, finite constant (adapted from Anderson et al. 1976).

Rodents and Insectivores

445

Figure 10. Illustration of toe-clipping procedures. In the left diagram only the first four toes are considered
for clipping, counting from the outer toe. A numbering system used in toe-clipping squirrels is shown
(from Mosby 1963) in the diagram on the right.

Small bird bands have been used as loose-fitting
bracelets around the hind legs of small mammals
(Chitty 1937). Use of such bands is more time-con-
suming than toe-clipping, however, and on some
species caused irritation, swelling, skin-puncture, and
infection (Takos 1943). Some of these difficulties
may be overcome by encircling the band around the
Achilles tendon rather than the entire leg. Smaller
bands have been clamped loosely around a hind toe
on squirrels and are kept in place by the toe pad
(Cooley 1948).

A different type of marker, used on muskrats
and other medium-sized rodents, consists of a water-
proof plastic button, threaded through the loose
skin of the back with a flexible steel needle and se-
cured with another button. This plastic marker has
the advantage of being visible from a distance on
a swimming muskrat (Hensley and Twining 1946;
Mosby 1963).

Coloring. Pelage coloring with commercial fur
dyes is another successful method of marking ani-
mals and lasts until the next molt. Most of these dyes
are poisonous and must be used with care. Two
common dyes are Nyanzol 4R (reddish-brown) and
Nyanzol D (black). Both may be fixed after applica-
tion with hydrogen peroxide ( Evans and Holdenried

1943) or mixed with hydrogen peroxide before
application (Fitzwater 1943). Dyes are useful for
mammals of light pelage but the range of colors is
small. The series can be increased by applying spots
to different parts of the body (Mosby 1963). Table 1
shows a few applications of these dyes to small
mammals.

(3) Probability of Capture Method (Re-
moval Trapping). Where animals are removed
from a population as captured, it is necessary to use
a different approach from the mark-recapture
method. However, data from the mark and recapture
samples may be examined by either the Lincoln In-
dex method or by the removal trapping method,
by using only the records for animals trapped the
first time.

The number of animals captured during any
trapping period may be viewed as the product of
two quantities, the first being the probability of cap-
ture and the second the number of animals present
at the beginning of the period (Hayne 1949a). The
probability of capture is assumed to be constant,
describing the hazard in which any animal stands in
relation to capture in the set of traps during one
period. The number of animals present at the begin-
ning of each trapping period is assumed to be the

446

Rodents and Insectivores

original population minus the number of animals
previously captured. Any great departure from these
assumptions seems to invalidate this method for
estimating the population.

Monitoring Small Mammal Populations

Monitoring small mammals is simply the check-
ing of the status of species, relative density, or num-
bers at some point before an initial survey or
inventory of the population. Baseline information is
required to determine how species and numbers are
changing over time. The objectives for monitoring
small mammals must be determined by the biologist,
e.g., to determine trends in populations of endan-
gered, threatened, or sensitive species; to determine
trends in species serving as important prey bases;
or to determine trends in numbers of other species
for economic, scientific, or other reasons. This might
involve monitoring only one species, such as beaver,
to maintain their numbers within prescribed limits
to prevent damage to vegetation or scenic resources,
or it might be to monitor several small rodent spe-
cies that are essential to the continued production of
an endangered or sensitive species or raptor. Rea-
sons will vary with locality. Time intervals between
checks of species and populations will vary accord-
ing to productivity, initial status, importance, and
possible decimating factors for the species involved.
Susceptibility to cycles must also be considered.

DISCUSSION AND SUMMARY

Larger rodents may often be surveyed visually
or by their signs, but small mammals, such as mice
and shrews, are difficult to observe directly. Noctur-
nality, protection by overhanging vegetation, and
retirement to nest or burrow upon detection of an
approaching human contribute to this difficulty of

observation. For this reason biologists use a variety
of sampling devices and techniques. In previous sec-
tions I have discussed general kinds of habitats occu-
pied by different groups of small mammals and some
of the methods for determining their presence, rela-
tive density, and estimated numbers. Here I will
discuss some of the problems involved in rodent and
insectivore studies that should be kept in mind when
designing an inventory or monitoring study and
some of the limitations of the data derived. Anyone
can set a trap line, but only a person with some un-
derstanding of rodent or insectivore ecology can put
the data derived in their proper perspective.

Sampling Assumptions

Species Occurrence and Relative Density. In

setting out single trap lines or a series of parallel
lines, it might be assumed that most of the individual
animals and species within easy travel range of the
traps will be equally susceptible to being caught.
This is not true. First, using bait in the traps
immediately biases the catch. Some species (and
probably some individuals) will be attracted to this
bait much more quickly than others. Animals, such
as deer mice, may be drawn from considerable
distance by the smell of the bait, whereas others are
only mildly attracted. Therefore, in many localities,
a preponderance of deer mice are caught during
early evening. In fact, most traps may become
occupied by deer mice, excluding other species for
the remainder of the night unless the traps are
emptied and reset. This is the reason for checking
and resetting traps each hour until midnight. By so
doing, most of the deer mice or other easily caught
species will be caught early and leave the traps
ready to capture more secretive, less numerous, or
more trap-shy animals. By following this procedure, a
more accurate relative density among species can
be obtained. Otherwise, the most easily trapped
species will appear the most numerous in the

Table 1. Techniques for coloring the pelage of mammals (adapted from Mosby 1963).

Species

Coloring Agent

Special Techniques

Duration
of Color

Authority

Pocket
gopher

Human black hair dye with oil
base: apply equal parts
3% hydrogen peroxide plus
granulated soap until liquid is
thick.

Ammonium hydroxide (4%) —
1 part to 2 parts 3% hydrogen
peroxide plus soap as above.

Recommended for light-
pelaged mammals

Recommended for dark-
pelaged mammals

Until
molt

Until
molt

Morejohn and
Howard 1956

Morejohn and
Howard 1956

Squirrel

Picric acid in 5% formalin.

Nyanzol A: 20 gm per L of
water-hydrogen peroxide
mixture in ratio of 2:1.

Use on light pelage

Ring animal's body in broad
bands

Until
molt

Fitzwater 1943
Fitzwater 1943

Rodents and Insectivores

447

composition, whether it is in actuality or not.
Regardless of the procedure you use in checking
traps, do it the same way each time in order to have
comparative information.

When monitoring populations, be sure all envi-
ronmental conditions and biotic factors are about
the same as for prior sampling. For example, species
will vary as to time of reproductive peaks, times of
greatest seasonal behavior, and months of greatest
dispersal. Results may be erroneous if sampling is
done before the reproductive peak one year and
after the peak when sampling two years later.

In spring and early summer, a preponderance of
males of some species may be caught when they
are especially active in seeking mates. Times vary
with different species. There is always a reason for a
preponderance of one sex or age group in a catch.
Some knowledge about rodent or insectivore ecol-
ogy will usually reveal the reason. In making com-
parisons of numbers or species among years, be
aware that many factors may be influencing the re-
sults. The reasons may be more important than the
numbers themselves.

Of all factors affecting capture probability, time
is the most easily controlled. The biologist can select
the season of the year when studies are to be con-
ducted, the length of the trapping period, and the
time of day when trapping is to be done. The objec-
tive is to reduce variation in capture probabilities
over time. For usable comparisons equal effort
should be expended on each sampling occasion. The
number of traps should be the same each time, trap-
ping should be done at the same time of day and, if
bait is used, the type and amount should be the
same on all occasions.

Weather is an extremely important factor in
small mammal trapping (or any kind of sampling).
Animals vary greatly in their trappability, behavior,
and activity depending on such factors as moonlight,
cloudiness, precipitation, high winds, and tempera-
ture. Where seasonal weather conditions have pro-
duced a poor seed crop, animals will likely be more
attracted to baits. It is also important to be aware
of possible damage to habitats caused by storms or
flooding before the sampling period. Many animals in
the study area might have been lost by drowning or
emigrated to safer habitat elsewhere. Drought can
similarly reduce populations by lowering productiv-
ity. On the other hand, unusually good survival con-
ditions may produce much higher than usual
numbers during the sampling period. Severe storms
and low temperatures or dry conditions that inhibit
seed production are among the most important fac-
tors to consider when evaluating species changes,
relative densities, or population fluctuations.

Relative rather than absolute density has been
assumed to suffice for solving many biological prob-
lems, but the area factor in density has been ignored.
If all procedures are maintained constant, including
the time over which samples are taken, it is assumed
that the relative densities obtained will suffice for
comparing different habitats or the same habitat at
different times. However, these relative densities can
be misleading. For example, on the trap lines run
by the North American Census of Small Mammals
from 1948 to 1951, 1,901 male and 1,521 female
deer mice were trapped. The conclusion might be
that males in these areas were 25% more abundant
than females, but this is probably not true. It is more
likely that males have a larger home range than do
females and, therefore, more males than females are
exposed to the trap lines (Stickel 1948).

Similar inaccuracies of assumptions that relative
densities are proportionate to true densities also
apply to comparisons of the density of different gen-
era or of the same genus in different habitats. Until
one knows more about the influence exerted upon
home ranges by species, sex, and habitat, it is well to
use caution in drawing conclusions from relative
densities other than for those densities that concern
a single species and sex from different times in the
same habitat.

Many small mammal species have home ranges,
whereas other species seem to wander continually
with no defined center of activity (Calhoun and
Casby 1958). When using the Lincoln Index method,
one assumes that the marked animals when released
into the population will become distributed through-
out and that further samples will take these marked
animals with no greater or lesser probability of cap-
ture than the unmarked animals (Hayne 1949a). If
the marked animals return to the use of their origi-
nal home ranges, they may or may not present a
random distribution within the population for subse-
quent sampling. In fact, White et al. (1982) stated
frankly that "there is simply no basis for thinking
that samples are drawn randomly in capture studies
of animal populations."

Another important assumption is that no signifi-
cant replacement of the population by unmarked
animals will occur between the marking and the
sampling periods. Mortality does not present a signif-
icant problem, so long as it happens in equal propor-
tion among marked and unmarked animals and does
not result in replacement of the dead animals by
unmarked individuals from other sources (Hayne
1949a). Movement out of a population is treated the
same as mortality; movement into it, as part of re-
cruitment (Schemnitz 1980). If either mortality or
recruitment is significant during the initial inventory
or monitoring period, subsequent sampling will be
difficult to analyze correctly without conducting
time-intensive studies (White et al. 1982).

448

Rodents and Insectivores

Turnover is rapid in many small mammal popu-
lations. Individuals are replaced constantly, either
by dispersal with replacement by individuals from
adjacent areas or by natural mortality with replace-
ment by younger individuals growing to adult status.
If too much time lapses between the preliminary
marking period and the subsequent sampling period,
an overestimation of the size of the population that
was present during the marking period inevitably
results. In either this method or methods for estimat-
ing density (Scott et al. 1978; O'Farrell et al. 1977;
White et al. 1982), the importance of being able
to assume no recruitment or mortality is empha-
sized. It is important not to wait more than a few
days between the marking period and the subse-
quent recapture period.

situation will tend to produce underestimates of the
population (Hayne 1949a). In other situations ani-
mals learn to avoid traps. This may happen to ani-
mals that have a toe clipped or have some other
unpleasant experience during or after capture in a
live trap or after a bare escape from a snap trap.
After first capture, an animal's capture probability on
subsequent capture occasions changes, often greatly
(Getz 1961; Bailey 1969; White et al. 1982). There-
fore, marked animals probably do not have the same
catchability as previously uncaught animals. Some
factors affecting small mammal responses to traps in-
clude weather, season, individual inquisitiveness,
population density, food availability, type of traps
used, social dominance, sex, and activity patterns
(Sheppe 1972; Sarrazin and Bider 1973; Summerlin
and Wolfe 1973).

General Considerations for all Small Mammal
Studies. Sufficient time and money for meaningful
studies will continue to be problems facing most
biologists. Good judgment must be exercised to
obtain the best information possible with limited
funding. Most small mammal studies in the past have
been classified by Hayne ( 1976) as "descriptive
studies" and discuss animals found at the time and
place for each particular study but have little
application to other times and areas because they are
seldom replicated. Hayne (1976) emphasized that a
sample size of even just two is very much better
than a sample size of one. He did not imply that a
sample size of two is adequate for precision.
Obtaining precision in nature can be very costly, but
the value of even one replication of a study can shed
great light on the results of the first study. Usually,
the results of two samples can be lumped or
averaged to give a better picture of the real
situation. Unreplicated studies can lead to
generalizations and unrestrained speculations; even
one replication of a sample in a comparable habitat
type should put some limitations on how the results
are interpreted.

One should not be intimidated by criticism that
two studies in the same kind of habitat in different
areas cannot be compared because they are not
identical habitats. There is no such thing as any two
areas being identical; there will always be some dif-
ferences, but if two similar areas are selected care-
fully, either for replication of an initial study or for
comparison with other years, the results should be
valid for comparisons.

The assumption of equal trappability is probably
false in most situations, even though it is a necessary
assumption in most sophisticated statistical proce-
dures (White et al. 1982). In some studies, a dispro-
portionate number of animals are captured, which
had a high probability of capture; these may be
termed as "repeaters" or described as having a "trap
habit." When using the Lincoln Index method, this

Considerable differences occur in the numbers
of animals trapped in different types of traps; also,
different sizes of the same type of trap may catch dif-
ferent numbers of animals. It would be advisable,
therefore, to use the same kind of traps and in the
same sequence (live vs. snap traps) for obtaining
samples that are to be compared between years, sea-
sons, or habitat types (Schemnitz 1980).

There have been few studies to check popula-
tion estimates against a known population. Most
present-day techniques of estimation have such low
precision that only large changes of the population
or a large influence of a factor can be detected
(Schemnitz 1980). Fortunately, low precision esti-
mates may be adequate for management purposes on
small mammals, because even a method of low preci-
sion can detect a large effect in a population. How-
ever, if better precision is required to meet a specific
objective, procedures may be found in the literature
(Hayne 1949a; DeLury 1951; Leslie 1952; Fredin
1954; Davis 1956; Nixon et al. 1967; Eberhardt
1969; Burnham 1972; Hansson 1974; Jensen 1975;
O'Farrell et al. 1977; Burnham and Overton 1979;
Cormack 1981; Seber 1982; White et al. 1982).

In the typical mark-recapture study, a main ob-
jective is to estimate the population size (N). Nei-
ther the true value of (N) nor the correct
assumptions to make about capture probabilities are
known (White et al. 1982). Delving into the litera-
ture one will find dozens of published estimators.
The biologists should select one and proceed with
the calculations. However, they rarely can test the
assumptions they have made in collecting samples,
nor do they take time to estimate the sampling vari-
ance of the population estimate. In some instances
assumptions hold for one species but not for another
living in the same habitat. Unfortunately, the devel-
opment of the mathematics and statistics has pro-
ceeded far more rapidly than has the testing of the

Rodents and Insectivores

449

V

assumptions usually used for statistical tests (Schem-
nitz 1980). The task of testing assumptions about
small mammal populations is tremendous and be-
yond the scope of most investigations. Because there
are numerous published estimators, biologists using
different estimators can get different estimates with
the same data. It is difficult to decide which esti-
mator will give the most accurate results for any
given situation. Another difficulty in analyzing differ-
ences in species, relative densities, or estimated
numbers during monitoring studies over a sequence
of years is that of determining whether apparent
differences are real or only variations that might be
expected among any set of samples.

At times, the biologist may wish to determine an
estimate of density of small mammals in a particular
habitat type. Density is defined as the number of
animals per unit area. Density estimation extends
population size estimation to include an estimate of
the area to which the population estimate relates.
Density estimates are usually derived with the use of
a trapping grid. However, density estimation is not
as simple as dividing the estimated population of
animals by the area of the trapping grid, because of
the phenomenon known as edge effect (White et al.
1982). For example, animals at the edge of a grid
will not spend all of their time on the grid, because
the grid area contains only a part of their home
ranges. Thus, the effective area trapped is somewhat
larger than the grid area, because the area to which
the population estimate applies includes the entire
home ranges of such animals.

A boundary strip should be included as part of
the area occupied by the population. Dice (1941)
suggested that the width of the boundary strip be
taken as one-half the average diameter of the home
range of the species in question. Other authors have
made additional recommendations (Hayne 1949b;
Stickel 1954; Tanaka 1972; Otis et al. 1978) but all
approaches are subject to difficulties. For example,
the estimates depend on trap spacing and numbers
of recaptures (White et al. 1982). A second ap-
proach involves the use of assessment lines (M.H.
Smith et al. 1971; R. Smith et al. 1975; H.D. Smith et
al. 1972; Scott et al. 1978; O'Farrell et al. 1977).
Although this approach has produced good results,
the method can become complex and is heavily
dependent on the design of the trap layout. How-
ever, it can be demonstrated that the importance of
the boundary strip decreases as the size of the grid

increases. Placing traps in a rectangular configuration
or in a series of parallel lines reduces this edge effect
even when using the simpler methods, such as trap
lines set out to determine relative density or species
occurrence.

Biologists desiring to calculate densities of one
or more species of small mammals should refer to
Literature Cited for needed statistical procedures.
White et al. (1982) stated that:

"computer programs have become essential in
the analysis of capture-recapture and removal stud-
ies. The iterative nature of many estimators of popu-
lation size under closure makes them nearly
impossible to compute with a hand calculator, and
the testing and model selection procedures are tedi-
ous. The notation and algebra for the estimators of
open-population parameters are difficult, and re-
cently developed models do not have closed-form
estimators. In all cases, rounding errors, especially
for the estimates of sampling variances, can be seri-
ous on a calculator. Now and in the future, a com-
prehensive analysis of any set of multiple capture
data will require the use of sophisticated computer
programs. A good computer system and its accompa-
nying software are now mandatory, if biologists are
to benefit from the statistical and theoretical ad-
vances made in the past decade. A general, flexible,
easy-to-use system will be a great help in future re-
search on biological populations."

The foregoing thoughts have been presented to
help the biologist comprehend the magnitude of
the factors that influence results of small mammal
studies. They are not meant to create doubt but,
rather, to create caution in making too broad gener-
alizations based on small samples and to emphasize
the need to design all monitoring studies to be com-
parable with previous studies. Only when the effects
of variables discussed in this section are about equal
can results of sampling be compared with any confi-
dence from year to year.

Most small mammal surveys for land manage-
ment purposes are to determine trends in species
and numbers, not necessarily to determine densities
(Kepner 1978). Even though the Lincoln Index
method or Probability of Capture method may pro-
vide somewhat crude results, data derived can be
effectively used for trends so long as conditions are
comparable at each sampling period.

450

Rodents and Insectivores

/

LITERATURE CITED

ANDERSON, B.W.J. DRAKE, and R.D. OHMART. 1977.
Population fluctuations in nocturnal rodents in the
Lower Colorado River Valley. Pages 183-192 in Proc.
Symp. on the Importance, Preservation, and Manage-
ment of the Riparian Habitat. U.S. Dep. Agric, For.
Serv. Gen. Tech. Report RM-43.

ANDERSON, DR., J.L. LAAKE, BR. CRAIN, and K.P. BURN-
HAM. 1976. Guidelines for line transect sampling of
biological populations. Utah Coop. Wildl. Res. Unit.
Logan. 27pp.

and R.D. OHMART. 1977. Rodent bait additive

which repels insects. J. Mammal. 58:242.

BAILEY, J.A. 1969. Trap response of wild cottontails. J.
Wildl. Manage. 33(l>48-58.

BLAIR, W.F. 1941. Techniques for the study of mammal
populations. J. Mammal. 22(2): 148- 157.

BRADT, G.W. 1938. A study of beaver colonies in Michi-
gan. J. Mammal. 19:139-162.

BURNHAM, K.P. 1972. Estimation of population size in
multiple capture-recapture studies when capture
probabilities vary among animals. Ph.D. dissertation.
Oregon State University, Corvallis.

DR. ANDERSON, and J.L. LAAKE. 1980. Estimation

of density from line transect sampling of biological
populations. Wildl. Monogr. 72:1-202.

and W.S. OVERTON. 1979- Robust estimation of

population size when capture probabilities vary
among animals. Ecology 60(5):927-936.

BURT, W.H. and R.P. GROSSENHEIDER. 1964. A field
guide to the mammals. The Riverside Press. Cam-
bridge, MA. 284pp.

CALHOUN, J.B. and J.V. CASBY. 1958. Calculation of home
range and density of small mammals. Public Health
Service, Public Health Mongr. 55. 24pp.

CALL, M.W. 1970. Beaver pond ecology and beaver-trout
relationships in southeastern Wyoming. Ph.D. disserta-
tion. Univ. of Wyoming, Laramie. 319pp.

CHITTY, D. 1937. A ringing technique for small mammals.
J. Animal Ecol. 6:36-53-

and M. SHORTEN. 1946. Techniques for the study

of the Norway rat (Rattus norvegicus). J. Mammal.
27:63-78.

COLLIER, E. 1957. Revolutionary new trap. Outdoor Life.
Sept. 1957, p. 38-41, 80, and Oct. 1957, p. 70-73,
80, 82.

COOLEY, ME. 1948. Improved toe-tag for marking fox
squirrels. J. Wildl. Manage. 12:213-

CORMACK, R.M. 1981. Log-linear models for capture-
recapture experiments on animal populations. Univ.
of Otago. Dunedin, New Zealand.

DAVIS, D.E. 1956. Manual for analysis of rodent popula-
tions. Edwards Brothers, Inc. Ann Arbor, MI. 82pp.

DELURY, D.B. 1951. On the planning of experiments for
the estimation of fish populations. J. Fish. Res. Board
of Canada. 8:281-307.

DENNEY, R.N. 1952. A summary of North American bea-
ver management. Colo. Game and Fish Dep. Current
Report 28. Denver, CO. 58pp.

DICE, L.R. 1941. Methods for estimating populations of
animals. J. Wildl. Manage. 5(4>398-407.

DYMOND, JR. 1947. Fluctuations in animal populations
with species reference to those of Canada. Trans.
Royal Soc. Canada 4l(5):l-34.

EBERHARDT, L.L. 1969. Population estimates from recap-
ture frequencies. J. Wildl. Manage. 33(l>28-39.

ELTON, C. 1942. Voles, mice and lemmings. Clarendon
Press. Oxford. 496pp.

ERRINGTON, P.L. 1944. Additional studies on tagged
young muskrats. J. Wildl. Manage. 8:300-306.

and C.S. ERRINGTON. 1937. Experimental tagging

of young muskrats for purposes of study. J. Wildl.
Manage. 24:231-260.

EVANS, EC. and R. HOLDENRIED. 1943- A population
study of the Beechey ground squirrel in central Cali-
fornia. J. Mammal. 24:231-260.

FITZWATER, W.D., Jr. 1943. Color marking of mammals,
with special reference to squirrels. J. Wildl. Manage.
7:190-192.

FREDIN, R.A. 1954. Causes of fluctuations in abundance of
Connecticut River shad. U.S. Dep. Inter., Fish and
Wildl. Serv. Fishery Bull. 88. Washington, DC.

GETZ, L.L 1961. Response of small mammals to live-traps
and weather conditions. Am. Midi. Nat. 66( 1 ):160-
169.

GILES, R.H., ed. 1969- Wildlife management techniques.
3rd ed. The Wildl. Soc. Washington, DC. 623pp.

HANSSON, L. 1974. Influence area of trap stations as a
function of number of small mammals exposed per
trap. Acta Theriol. 19:19-25.

HAYNE, D. W. 1 949a. Two methods for estimating popula-
tion from trapping records. J. Mammal. 30:399-41 1.

1949b. Calculation of size of home range. J. Mam-
mal. 30(1):1-18.

1976. Experimental designs and statistical analyses

in small mammal population studies in Populations of
Small Mammals Under Natural Conditions. The Pyma-
tuning Symposia in Ecology, Vol. 5. Special Publ.
Series, Pymatuning Lab. of Ecology, Univ. of Pitts-
burgh. Symposia held at the Pymatuning Lab. of Ecol-
ogy, May 14-16, 1976.

HENSLEY, A.L. and H. TWINING. 1946. Some early sum
mer observations on muskrats in a northeastern Cali-
fornia marsh. Calif. Fish and Game 32:171-181.

HOFFMEISTER, D.F. and L. DE LA TORRE. I960. A revi-
sion of the woodrat Neotoma stephensi. J. Mammal.
4l(4):482.

INGLES, L.G. 1965. Mammals of the Pacific States. Stanford
Univ. Press. Stanford, CA. 506pp.

JACKSON, H.H.T. 1961. Mammals of Wisconsin. Univ.
Wisconsin Press. Madison. 504pp.

JENSEN, T.S. 1975. Trappability of various functional

groups of the forest rodents Clethrionomys glareolus
and Apodemus Jlavicollis, and its application in den-
sity estimations. Oikos 26(2): 196-204.

JORGENSEN, CD. and H.D. SMITH. 1974. Mini-grids and
small mammal estimates in Proceedings of the Utah
Academy of Sciences, Arts, and Letters. Vol. 51, Part I.

KEITH, LB. 1963. Wildlife's ten-year cycle. Univ. Wiscon
sin Press. Oxford. 201pp.

KEPNER, W.G. 1978. Small mammals of the Black Canyon
and Skull Valley planning units, Maricopa and Yavapai
Counties, Arizona. U.S. Dep. Inter., Bur. Land Manage.
Tech Note 350. 37pp.

KREFTING, L.W., J.H. STOECKLER, B.J. BRADLE, and WD.
FITZWATER. 1962. Porcupine-timber relationships in
the Lake States. J. For. 60:325-330.

LESLIE, PH. 1952. The estimation of population parame-
ters from data obtained by means of the capture-
recapture method: II. The estimation of total num-
bers. Biometrika 38(3/4):368-388.

Rodents and Insectivores

451

and D.H.S. DAVIS. 1939. An attempt to determine

the absolute number of rats on a given area. J. Anim.
Ecol. 8:94-113.

MASER, BR. MATE, J.F. FRANKLIN, and C.T. DYRNESS.
1981. Natural history of Oregon coast mammals. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. PNW-133.
496pp.

MOREJOHN, G.V. and WE. HOWARD. 1956. Moult in the
pocket gopher, Thomomys bottae. J. Mammal.
37:201-212.

MOSBY, H.S. ed. 1963. Wildlife investigational techniques,
2nd Edition. The Wildl. Soc. Washington, DC. 4 19pp.

NIXON, C.N., W.R. EDWARDS, and L.L. EBERHARDT.

1967. Estimating squirrel abundance from live trap-
ping data. J. Wildl. Manage. 31(1):96-101.

OTARRELL, M.J., D.W. KAUFMAN, and D.W. LUNDOHL.
1977. Use of live-trapping with the assessment line
method for density estimation. J. Mammal. 58(4):575-
582.

OTIS, D.L., K.P. BURNHAM, G.C. WHITE, and DR. ANDER
SON. 1978. Statistical inference from capture data
on closed animal populations. Wildl. Monogr. 62:1-
135.

PETRUSEWICZ, K. and A. MACFAYDEN. 1970. Productiv-
ity of terrestrial animals — principles and methods.
Inter. Biol. Prog. Handbook 1 3. FA. Davis Co. Phila-
delphia, PA. 190pp.

REYNOLDS, H.C. 1945. Some aspects of the life history
and ecology of the opossum in central Missouri. J.
Mammal. 26:361-379.

RUTHERFORD, W.H. 1964. The beaver in Colorado, its
biology, ecology, management and economics. Colo.
Game, Fish and Parks Dep. Fed. Aid in Wildl. Rest.
Proj. W-83-R Denver, CO. 49pp.

SANDERSON, G.C. 1961. Estimating opossum population
by marking young. J.Wildl. Manage. 25:20-27.

SARRAZLN, J.P.R. and JR. BIDER. 1973- Activity, a ne-
glected parameter in population estimates — the devel-
opment of a new technique. J. Mammal. 54(2 ):369-
382.

SCHEMNITZ, S.D. ed. 1980. Wildlife management tech-
niques. 4th ed. revised. The Wildl. Soc. Washington,
DC. 686pp.

SCHULTZ, V. 1961. An annotated bibliography on the uses
of statistics in ecology — a search of 3 1 periodicals.
Publication TID-3908. U.S. Atomic Energy Comm.
Office of Tech. Info, Environmental Science Branch,
Div. of Bio. and Medicine. Washington, DC. 31 4pp.

SCOTT, D.T., CD. JORGENSEN, and H.D. SMITH. 1978.
Comparison of live and removal methods to estimate
small mammal densities. Acta Therio. 23(8): 173- 193-

SEBER, G.A.F. 1973- The estimation of animal abundance.
Hafner Press, New York, NY. 506pp.

. 1982. Estimation of animal abundance and related

parameters. 2nd. ed. Hafner Press. New York, NY.
654pp.

SHEPPE, W. 1972. The annual cycle of small mammal
populations on a Zambian floodplain. J. Mammal.
53(3>445-460.

SMITH, H.D., C D. JORGENSEN, and H.D. TOLLEY. 1972.
Estimation of small mammals using recapture meth-
ods: partitioning of estimator variables. Acta Therio.
17:57-66.

SMITH, M.H., R. BLESSING, J.G CHELTON, J.B. GENTRY,
G. GOLLEY, andJ.T. MCGINNIS. 1971. Determining
density for small mammal populations using a grid
and assessment lines. Acta Theriol. 16:105-125.

SMITH, R., H. GARDNER, J.B. GENTRY, D.W. DAUFMAN,
and M.J. OTARRELL. 1975. Density estimation of
small animal populations, Pages 25-53 in F.G. Golley,
K. Petrusewicz, and L. Ruszkowski, eds. Small Mam-
mals: Their Production and Population Dynamics,
Intern. Biol. Prog. 5. Cambridge Univ. Press, London.

SNEDECOR, G.W. and W.G. COCHRAN. 1967. Statistical
methods, 6th ed. Iowa State Univ. Press, Ames. 593pp.

SOKAL, R.R. and F.J. ROHLF. 1969. Biometry: the princi-
ples and practice of statistics in biological research.
W.H. Freeman and Co. San Francisco, CA. 776pp.

SPEIRS, J.M. 1939. Fluctuations in numbers of birds in the
Toronto region. Auk 56:411-419.

STICKEL, L.F. 1948. The trap line as a measure of small
mammal populations. J. Wildl. Manage. 12:153-161.

1954. A comparison of certain methods of measur-
ing ranges of small mammals. J. Mammal. 35(1):1-15.

SUMMERLIN, C.T andJ.L. WOLFE. 1973. Social influences
on trap response of the cotton rat, (Sigmodon hispi-
dus). Ecology 54(5): 11 56- 11 59.

SVIHLA, A. 1934. Development and growth of deer mice
(Peromyscus maniculatus artemisiae), J. Mammal.
15:99-104.

TAKOS, M.J. 1943- Trapping and banding muskrats. J.
Wildl. Manage.7:400-407.

TANAKA, R 1972. Investigation into the edge effect by
use of capture-recapture method in a vole population.
Res. Pop. Ecol. 12(1 ):1 11-125.

WHITE, G.C, DR. ANDERSON, K.P. BURNHAM, and D.L.
OTIS. 1982. Capture-recapture and removal methods
for sampling closed populations. Los Alamos National
Laboratory. Los Alamos, NM. 235pp.

YERGER, R.W. 1953- Home range, territoriality,

and populations of the chipmunk in central New
York. J. Mammal. 34:448-458.

452

Rodents and Insectivores

22

LAGOMORPHS

Joseph A. Chapman

Utah State University
Logan, UT 84322

Editor's Note: Lagomorphs, among other midsize
mammals, have often been neglected in wildlife
surveys. Although often locally important as a
game or pest species, they have not received the
same amount of attention as other more conspicu-
ous or economically important species. Further-
more, typical inventorying techniques and
equipment used for other species are not useful in
studying lagomorphs. This chapter identifies some
possible techniques that can be used for inventory-
ing and monitoring this often overlooked group
of species.

Gale R. Willner

The Ecology Center
Utah State University
Logan, UT 84322

INTRODUCTION

Lagomorphs are important components of the
world's ecosystems, particularly in the western U.S.
and Alaska where they make the base of many carni-
vore food chains. In addition, they are important
indicators of habitat quality, sport animals, depreda-
tors, and models for ecological research.

The order Lagomorpha occurs worldwide and
contains two families, Leporidae and Ochotonidae,
both found in the U.S. The North American Lepori-
dae includes hares and jackrabbits (Lepus sp.), cot-
tontails (Sylvilagus sp. ), and volcano rabbit
(Romerolagus sp.), found only in Central Mexico.
Five species of Sylvilagus and four species of jackrab-
bits occur in the western half of the U.S. Snowshoe
hares (L. americanus) are widely distributed in
Alaska and the western U.S., but the Alaskan hare (L.
othus) occurs only in Alaska. The pikas (Ochotona
princeps) are the only members of the family Ocho-
tonidae and they live on talus slopes throughout
most of the higher mountains in the western U.S. and
southeastern Alaska.

All Leporidae are adapted for quick movement
and flight from danger. Their hind legs are long and
adapted to cursorial locomotion. Their ears are large
and moveable to permit detection of approaching
enemies. Unlike pikas, their eyes are large and suited
to their crepuscular and nocturnal habits. These
features are most developed in the hares. Many cot-
tontails rely on thick brush or even burrows to avoid
predators. Pikas are rodent-like in appearance and
are well-suited to living among the rocks and boul-
ders of their talus slope environment.

In general, Lagomorphs are highly fecund and
short-lived. Population levels are regulated by mor-
tality (survival) and dispersal. Indirectly, the compo-
sition and structure of the plant community within
their habitat plays an important part in the survival
of these mammals. They are usually associated with
disturbed or subclimax plant communities.

Lagomorphs

453

The purpose of this chapter is to summarize
techniques that are most useful in determining Lago-
morph population levels as well as examine habitat
features that are important to this order. This should
enable the land manager to evaluate the relationship
of density and habitat and thus determine the status
of Lagomorphs in a given area.

In this chapter, the Lagomorphs are divided into
four species groups on the basis of taxonomy, repro-
ductive characteristics, and their most common habi-
tat requirements.

HABITAT FEATURES CORRELATED WITH
SPECIES GROUPS

Jackrabbits

Jackrabbits are true hares, primarily associated
with arid regions of the 1 1 western States. They are
found at elevations from below sea level in Death
Valley to nearly 4,000 m (13,000 ft). Jackrabbits will
drink water, but researchers do not consider it a
requirement of their habitat.

60c

^^

WjSS*.

o

vo r

Miles

400 800
1

400 800

Kilometers

120;

Figure 1. Distribution of the black-tailed jackrabbit in the western United States (adapted from Dunn et al.
1982).

454

Lagomorphs

White-tailed jackrabbits are often found in prai-
ries and in open flats and alpine zones above tree-
line. Black-tailed jackrabbits are the most abundant
hare in the western U.S.; white-tailed jackrabbits are
locally abundant. The antelope and white-sided jack-
rabbits are uncommon to rare throughout their
rather restricted ranges. (See Figures 1,2, and 3 for
distribution. )

White-tailed jackrabbits usually occur in low-
growing scrub, grasslands, or prairie areas. In the
western States they are often found at higher eleva-
tions, even above treeline where they inhabit tundra-
like vegetation (Braun and Streeter 1968). In moun-
tainous regions they also occur in areas vegetated
with disjunct clumps of sagebrush. See Tables 1 and
2 for cover and food plants found in areas occupied
by jackrabbits.

White-tailed jackrabbit.

Black-tailed jackrabbits are often associated with
shortgrass-sagebrush areas in arid regions. They also
inhabit agricultural areas. White-tailed jackrabbits are
also found in sagebrush areas in higher mountain
regions, but are most often found in prairie grass-
lands. Black-tailed jackrabbits appear to displace the
white-tailed jackrabbit in disturbed or cultivated
areas where the two occur sympatrically.

Although black-tailed jackrabbits seem to prefer
diverse plant communities made up of disjunct
cover, there are little quantitative data available.
Black-tailed jackrabbits are found in a wide array of
habitat types, all of which are characterized or domi-
nated by low-growing trees or shrubs and well-
drained soils. High populations occur in sagebrush
{Artemisia sp.) interspersed with sparse grass. In
general, only open or semi-open areas are occupied
by this species. No area appears to be too open as
long as it provides adequate hiding, resting, and
birthing sites. Black-tailed jackrabbits are usually
associated with vegetation that is no more than
about 1 m (3 ft) high.

The antelope jackrabbit occurs exclusively in
the lower Sonoran life zone in Arizona where its
preferred habitat is grassy slopes. However, it is also
found in the cactus belt in areas dominated by mes-
quite (Prosopis sp. ), catclaw (Acacia greggi), and
various grasses. Smaller populations occupy creosote
bush (Larrea sp. ) deserts and valley bottoms (Figure
2). It occurs at elevations from sea level to 1,200
m (3,960 ft) (Vorheis and Taylor 1933). Similarly,
the white-sided jackrabbit is found exclusively in the
shortgrass habitats of New Mexico (Figure 2). The
preferred habitat is composed of 65% or more
grasses, 25% or less forbs, and 1% or less shrubs
(Bednarz 1977).

Black-tailed jackrabbit.

Antelope jackrabbit.

Lagomorphs

455

Snow shoe and Alaskan Hares

Alaskan hares are among the least known Lago-
morphs in North America. Recent work indicates
that L. othus, L. arcticus, and L. timidus are perhaps
the same species with a Holarctic distribution
(Dixon et al. 1983). There are two distinct ecotypes
of Alaskan hares: L. o. othus, associated with tundras
and alluvial plains, and L. o. poudromus, associated
with coastal lowland areas of the Aleutian Island
chain (Bittner and Rongstand 1982).

Snowshoe hares are found in most higher moun-
tains of the western States, except Arizona, as well
as throughout Alaska (Figure 3).

Snowshoe hares inhabit subclimax forests and
transition zones adjacent to heavy cover. High densi-
ties of snowshoe hares will not be found at distances
greater than 200 to 400 m (660 to 1,320 ft) from
conifer stands. High densities are unlikely in areas
with a solid canopy (Conroy et al. 1979). Through-
out their range, these hares appear partial to dense
thickets and windfalls where they seek cover. In
Alaska, they are found in a mosaic of plant communi-
ties that include upland black spruce/ledum/moss,
willow-alder thickets, and white spruce/birch/aspen
habitat types (Wolff 1978).

Snowshoe hares are also found in dense, second-
growth forests, often associated with aspen (Populus

60

□ White-tailed
| Antelope

^QtfcC?^ □ White-sided

Miles

400
1 —

Kilometers

Figure 2. Distribution of the white-tailed, antelope, and white-sided jackrabbits in the western United States
(adapted from Dunn et al. 1982.)

456

Lagomorphs

tremuloides) thickets and coniferous swamps. In
the central Rocky Mountains, snowshoe hares use
areas of spruce (Picea sp.), fir {Abies sp. and Pseu-
dotsuga sp.), and lodgepole pine (Pinus contorta). In
those habitats they prefer conifer stands with dense
brushy understories (Dolbeer and Clark 1975). In
Utah, prime snowshoe hare habitat occurs in areas
with high sapling numbers, high basal area, and
dense canopy cover. Optical cover densities of less
than 40% above the snowline limit winter use of
a habitat by snowshoe hares (Wolff et al. 1982).
Snowshoe hare numbers are greatest in areas sub-
jected to periodic burning (Grange 1965).

Little is known about the habitat of the Alaskan
hare except that one subspecies prefers tundra or
alluvial plain, and the other prefers coastal lowlands
(Bittner and Rongstand 1982).

In the western States snowshoe hares are invari-
ably linked with conifers at higher elevations. How-
ever they may range away from coniferous timber at
high elevations where streams are thickly bordered
with vegetation (Dice 1926; Orr 1940). These hares
also utilize logged-over areas adjacent to standing
conifers. See Tables 1 and 3 for detailed list of cover
and food plants found on areas occupied by
snowshoe hares.

60

| Snowshoe

Miles

400 800

Kilometers

120

Figure 3. Distribution of the snowshoe and Alaskan hares in the western United States and Alaska (adapted
from Bittner and Rongstand 1982).

Lagomorphs

457

Table 1. Some plants occurring in habitats occupied by various Lagomorph species.

Lagomorph and Plant Species

Area

Black-tailed jackrabbit

Rabbitbrush (Chrysothamnus sp.)
Sagebrush (Artemisia tridentata)
Saltbush {Atriplex sp.)
Horsebrush {Tetradymia sp.)

California

Creosotebush (Larrea tridentata)

Southern Nevada and
southeastern California

Joshua trees (Yucca brevifolia)

Creosotebush

Brig ham tea (Ephedra sp.)

Southern Utah

Shadscale (Atriplex confertifolia)

Greasewood (Scarcobatus vermiculatus)

Rabbitbrush

Sagebrush

Juniper (Juniperus sp.)

Pinyon pines (Pinus edulis)

Salt deserts

of
Great Basin

Sagebrush

Rabbitbrush

Mesquite (Prosopis sp.)

Snakeweed (Gutierrezia sarothrae)

Soaptree yucca (Yucca sp.)

Agricultural crops — alfalfa hay

Central and
southern Utah
New Mexico

White-tailed jackrabbit

Juniper (J. occidentalis)
Lodgepole pine (Pinus contorta)
Dwarf juniper (J. communis)
Creambush (Holodiscus discolor)
Granite/gilia (Leptodactylon purges)

California
(Lake Tahoe)

Sagebrush

Bitterbrush (Purshia tridentata)

Giant wild rye (Elymus condensatus)

Chokecherry (Prunus melanocarpa)

Snowberry (Symphoricarpos rotundifolius)

Mountain myrtle (Pachystima myrsinites)

Northern Utah

Deerbrush (Ceanothus velotinus)
Aspen (Populus tremuloides)

Utah and Idaho

Sagebrush

Greasewood

Shadscale

Saltbush

Dogwood (Cornus canadensis)

Wild rhubarb (Polygonum alaskanum)

Horsetail (Equisetum sp.)

Uintah Basin

Audubon's cottontail

Willow (Salix sp.)

Buttonwillow (Cephalanthus occidentalis)

Wild grape (Vitis californica)

Arroyo willow (S. lasiolepis)

Brushland (Adenostoma fasilutatuium)

(Eriodictyeon sp.)

(Eriogonum sp.)

California

Arrow-weed (Pluchea sericea)
Screw-bean mesquite (Prosopis pubesans)
Catclaw (Acacia greggii)

Southern Nevada

Joshua trees
Cactus

Southern Utah

458

Lagomorphs

Table 1. Some plants occurring in habitats occupied by various Lagomorph species (continued).

Lagomorph and Plant Species (continued)

Area

Audubon's cottontail (continued)

Brigham tea

Creosotebush

Junipers

Pinyon

Evergreen oaks (Quercus turbinella)

Gumine bush (Garrya flaresceni)

Saltbush

Rabbitbrush (Chrysothamnus nauseous)

Snakeweed

Winterfat (Eurotia sp.)

Western Utah

Nuttall's cottontail

Sagebrush

Rabbitbrush

Wild rye (Elymus condensatus)

Bitterbrush

Chokecherry (Prunus virginiana)

Northern Utah

Brush rabbit

Bramble {Rubus sp.)

Oregon

Buckbrush (Ceanothus sp.)

Willows

California rosebay {Rhododendron californicum)

Chaparral broom {Baccharis pilularis)

Wild rose (Rosa californica)

Poison oak {Rhus diversiloba)

Scrub oak (0. dumosa)

Snowberry (Symphoricarpus albus)

Yellow pine (Pinus ponderosa)

Manzanita (Arctostaphylos sp.)

California

Antelope jackrabbit

Catsclaw {Acacia greggi)

Mesquite

Grasses (Graminea)

Creosotebush (Larrea sp.)

Arizona

White-sided jackrabbit

Grasses

New Mexico

Grama (Bouteloua gracilis)
Black grama (B. eriopoda)
Buffalo grass (Buchloe dacyloides)
Wolftail (Lycurus phleoides)

Snowshoe hares (Continental U.S.)

Willows

Douglas fir (Pseudotsuga menziesii)

Snowberry

Snowbrush

Mountain myrtle

Mountain ash (Sorbus scopulina)

Aspen

Northern Utah

Lagotnorphs

459

Table 1. Some plants occurring in habitats occupied by various Lagomorph species (concluded).

Lagomorph and Plant Species (Concluded)

Area

Snowshoe hares (Continental U.S.) (continued)

Alpine fir (Abies)

Engleman spruce (P. englemanii)

Lodgepole pine

Alder (Alnus sp.)

Willow

Conifers

Chaparral {Ceanothus sp.)

Manzanita

Buckbrush (C. velutinus)

Douglas fir

Red cedar (Thuja plicata)

Uintah Mountains
of Utah
California

Southern California

Snowshow hares (Alaska)

White spruce (Picea glauca)

Black spruce (P. mariana)

Paper birch (Betula papyrifera)

Aspen

Balsam poplar (Populus balsamifera)

Alder (Alnus crispa)

Dwarf birch (Betula gladulosa)

Willows

Labrador tea (Ledum groenlandlcum)

Blueberry (Vaccinium uliginosum)

Lowbush cranberry (Vaccinium vitis-idaea)

Rose (Rosa acicularis)

Grasses (Calamagrostis sp.)

Fireweed (Epilobium angustifolium)

Pikas

Rabbitbrush

Sagebrush

Bitterbrush

Columbine (Aquilegia pubescens)

California

Sources: (Orr 1940, Janson 1946; Davis et al. 1975; Bednarz 1977; Dice 1926; Wolff 1978; Chapman 1971; Smith 1974a).

Cottontails

Cottontails are widely scattered throughout 1 1
western States, but are not present in Alaska. The
Nuttall's cottontail inhabits the intermountain region,
whereas the Audubon's cottontail is found in the
arid Southwest and high deserts north to Montana.
Ranges of these two species broadly overlap in Colo-
rado, Montana, Utah, and Wyoming (Figure 4). The
brush rabbit is confined to the coasts of California
and Oregon, whereas the pygmy rabbit is found in
the Great Basin with a small disjunct population
in eastern Washington (Figure 5). The eastern cot-
tontail occurs primarily from the Midwest east, with

the exception of Arizona and New Mexico, which
have large eastern cottontail populations. This spe-
cies occurs peripherally in eastern Colorado, Wyo-
ming, and Montana and has been introduced into
Oregon and Washington (Figure 5).

The Audubon's cottontail occurs in the lower
Sonoran life zone and often prefers stream bottom
vegetation. Shrubs interspersed with pinyon (Pinus
edulis) and juniper (Juniperus sp.) provide excellent
cover for this species. Downed trees or shrubs piled
170 to 220 per ha (425 to 1,210 per a.) create ex-
cellent habitat for this species (Kundaeli and Rey-
nolds 1972).

460

Lagomorphs

The Nuttall's cottontail is found in rocky,
wooded, or brushy areas. It is often observed in
rocky ravines next to sagebrush-covered hills. When
Audubon's and Nuttall's cottontails occupy the same
region, the former is found on desert valleys and
the latter on higher rocky, sagebrush-covered slopes.
Where cover is scarce, Nuttall's cottontails will use
burrows dug by badgers ( Taxidea taxus) or crevices
under rocks (Dice 1926; Orr 1940).

Pygmy rabbits are always found in sagebrush.
They prefer tall sagebrush (1 to 1.5 in [3 to 5 ft])
often in ravines. They dig their own burrows which
are found at the base of sagebrush clumps (Chapman
et al. 1982).

The brush rabbit gets its name from its prefer-
ence for dense, brushy cover. Clumps of brush
smaller than 460 m (14,900 ft ) are probably not
permanently occupied (Chapman 1971). Brush rab-
bits have been reported to use burrows, but they do
not dig their own.

In Arizona and New Mexico, eastern cottontails
are found in open areas and along stream bottoms or
near farm buildings. Where they have been intro-
duced in Washington and Oregon, they have become
established on old, overgrown farmsteads, in fence
rows, and around buildings. Eastern cottontails in
Minnesota have a distinct preference for shrubby
vegetation (Swihart and Yahner 1984). They are

60

□ Nuttall's
| Audubon's

Miles

400

1 —

Ki lometers

120

Figure 4. Distribution of the Nuttall's and Audubon's cottontails in the western United States (adapted from
Chapman et al. 1982).

Lagomorphs

461

most abundant in disturbed successional areas
(Chapman et al. 1982). See Tables 1 and 4 for de-
tailed list of cover and food plants found in areas
occupied by cottontails.

Pikas are found throughout the higher mountain
ranges of the western States except Arizona and New
Mexico. There is also a population of collared pikas
in the mountains of southern Alaska (Figure 6).

Pikas

Pikas are guinea-pig-shaped Lagomorphs, having
short, rounded ears and no visible tail. Pikas move
with a running gait rather than the typical hopping
or leaping of other Lagomorphs. They are the only
North American Lagomorphs that store or cache
food. They are active during the day and are ex-
tremely vocal.

Pikas inhabit talus slopes where they feed al-
most exclusively on grasses, forbs, and low bushes
adjacent to these talus slopes. Pikas require a boul-
der-strewn habitat on high mountain slopes with
a nearby source of food. Pikas require an abundance
of plant species which they can collect and store
in "hay" piles for later use in winter. Most often they
prefer alpine meadows with abundant grass and low
shrubs adjacent to talus slopes.

60c

□ Brush

^s^M} D Py9my

*4^0«fcC3 □ Eastern

Miles

400

1
400

Kilometers

Figure 5. Distribution of the brush rabbit, pygmy rabbit, and eastern cottontail in the western United States
(adapted from Chapman et al. 1982).

462

Lagomorphs

Pikas are noticeably absent from the Olympic
Mountains in Washington even though suitable habi-
tat appears to be available. Researchers attribute
this to the lack of talus occurring between the Olym-
pics and the Cascade Range where pikas are abun-
dant (Dalquist 1948). Barriers to dispersal, such
as climatic conditions, prevent juveniles from occu-
pying suitable habitats and, therefore, limit popula-
tion densities (Smith 1974b). Table 1 lists plant
species found on areas occupied by pikas.

POPULATION MEASUREMENT TECHNIQUES

Standard census techniques may be applied to
Lagomorph populations. Those techniques are sum-

marized in the Wildlife Management Techniques
Manual (Davis and Winstead 1980) and in The
Handbook of Census Methods for Terrestrial
Vertebrates (Davis 1982). These census techniques
may be divided into three broad categories: ( 1 ) pres-
ence or absence, ( 2 ) relative abundance, and ( 3 )
density estimates.

Absolute density and relative abundance are
somewhat misleading terms. In reality it is difficult to
meet the requirements for estimates of absolute den-
sity because of problems associated with delineating
boundary areas. Drive counts and mark-recapture
techniques are often thought of as measures of abso-
lute density, but it is probably better to consider

60

SfrKSSP.

Miles

0 400

I

400 800

K ilometers

Figure 6. Distribution of the pika in the western United States and Alaska (adapted from Chapman 1979).

Lagotnorphs 463

Table 2. Food of jackrabbits as reported in the literature.

Black-tailed Jackrabbit

White-tailed Jackrabbit

SHRUBS

SHRUBS

Shadscale

Winterfat

Fringed sage {Artemisia frigida)

Greasewood

Sagebrush

Hopsage {Grayia spinosa)

Four-winged saltbush (Atriplex

canescens)

Horsebrush

Snakeweed

Creosote

FORBS

FORBS

Globemallow {Sphaeralcea

Aster (Casfer sp.)

cocci ni a)

Vetch (Astragalus sp.)

Winterfat (Eurotia lanata)

Saltbush

Larkspur (Delphinium sp.)

Indian paintbrush (Castilleja integra)

Spiderwort (Tradescantia sp.)

Golden aster (Chrysopsis villosa)

Alfalfa (Medicago sativa)

Goosefoot (Chenopodium sp.)

Mallow plants (Malvaceae sp.)

Corydalis (Corydalis aurea)

Curly dock (Rumex crispu)

Daisy (Erigeron sp.)

Yucca (Yucca glauca)

Gilia (Gilia aggregate)

Prickly pear (Opuntia sp.)

Gaura (Gaura coccinea)

Sand dropseed (Sporobolus

Sunflower (Helianthus annuus)

cyptandrus)

Summer cypress (Kochia sp.)

Pigweed (Amaranthus retroflexus)

Blazing star (Liatris punctata)

Russian thistle (Salsola kali)

Lupine (Lupinus sp.)

Buffalo burr (Solanum rostratum)

Alfalfa

Lily (Lilium sp.)

Bluebells (Mertensia lanceolata)

Sedge (Carex sp.)

Four-o'clocks (Mirabilis linearis)

Cactus (Opuntia sp.)

Musineon (Musineon divaricatum)

Prickly pear cactus

Locoweed (Oxytropis sp.)

Beardtongue (Penstemon sp.)

Plaintain (Plantago purshii)

Cinquefoil (Potentilla sp.)

Psoralea (Psoralea tenluiflora)

Rabbitbrush

Russian thistle

Groundsel (Senecio sp.)

Scarlet globe mallow (Sphaeralcea

cocci ni a)

Dandelion (Taraxacum officinale)

Spreading thermopsis (Thermopsis

divaricarpa)

Verbena (Verbena bracteata)

Yucca

GRASSES

GRASSES

Needle-and-thread (Stipa comata)

Crested wheatgrass (Agropyron

Downy brome (Bromus tectorum)

cristatus)

Western wheatgrass (Agropyron

Western wheatgrass

smithii)

Wheatgrass (A. trachycaulum)

Barley (Hordeum sp.)

Red three awn (Aristida longiseta)

Buffalo grass (Buchloe

Oats (Avena sativa)

dactyloides)

Blue grama

Winter wheat (Triticum aestivum)

Brome (Bromus anomalus)

Blue grama

Smooth brome (B. inermis)

Indian ricegrass (Oryzopsis

Downy brome

hymenoicJes)

Sedges

Lovegrass (Eragrostis sp.)

Hairgrass (Deschampsia caespitosa)

464

Lagomorphs

Table 2. Food of jackrabbits as reported in the literature (concluded).

Black-tailed Jackrabbit

White-tailed Jackrabbit

GRASSES

Barnyard grass (Echinochloa sp.)
Panicum {Panicum sp.)
Dropseed (Sporobolus sp.)
Trichloris (Trichloris sp.)

GRASSES

Love grass

Arizona fescue (Festuca arizonica)

Foxtail barley

Junegrass (Koeleria cristada)

Ringgrass

Bluegrass (Poa sp.)

Squirreltail (Sitanion hystrix)

Sand dropseed

Needle-and-thread

Letterman needlegrass (S. lettermanii)

Wheatgrass

White-sided Jackrabbit

Antelope Jackrabbit

GRASSES

Buffalo grass

Tabossa grass

Fiddleneck (Amsinickia sp.)

Wolftail

Blue grama

Vine Mesquite (Pancium

obstosum)
Ring muhly (Muhlenbergia

torreyi)
Wooly Indian wheat (Plantago

purshii)
Wright buckwheat (Eriogonum

wrightii)

FORBS

Cactus (Echinocactus sp.)

GRASSES
Three awn
Grama
Love grass
Spike grass

Sandbar (Cenchrus Sp.)
Drop Seed

Red top (Agrostis sp.)
Barnyard grass
Finger grass (Chloris sp.)

Sources: (Hanson and Flinders 1969; Bear and Hansen 1966, Flinders and Hansen 1972, Bednarz 1977; Vorheis and Taylor 1933).

these techniques as being more accurate estimates
rather than measures of absolute density.

Presence

The presence of Lagomorphs can be determined
by direct observation of the animal or by "activity
signs." Activity signs include fecal pellets, active
runways and trails, tracks, skeletal remains, and feed-
ing sites (Figure 7). For example, jackrabbits can
usually be observed in the field because of their
rather large size. Snowshoe hares use runways, espe-
cially in the snow. Hay piles along talus slopes are
good indicators of pikas. The presence of these signs
does not necessarily provide an estimate of numbers
of Lagomorphs in a given area, but will help deter-
mine sites where they occur or avoid.

Relative Abundance

The most commonly used technique for census-
ing several species of Leporids is the roadside count.
Roadside census of Leporids is best in the evenings

or early morning when Lagomorphs' eyes can be
reflected from a spotlight. Jackrabbits and desert cot-
tontails have been censused in the evenings along
unpaved roads in shortgrass prairie habitat (Flinders
and Hansen 1973). The "eye shine" provided by
the spotlight and subsequent "freeze" of the Lago-
morphs permitted easy identification of species and
counting. Roadside counts may be used as long as
factors such as time of day, time of year, and
weather conditions remain constant (equal). This
type of transect count provides an index to relative
abundance but does not give a density estimate
(Chapman et al. 1982).

Density Estimates

Density estimates can be obtained by drive
counts. Using this method, several people spread
equally across an area, walk abreast along parallel
transects and count Leporids as they are flushed.
Flushing angles and flushing distances should be re-
corded. The flushing angle is obtained by pinpointing

Lagomorphs

465

Table 3. Food of the snowshoe hare in western
North America as reported in the literature.

Alaska

Birch

Spruce

Alder

Willow

Viburnum (Viburnum edule)

Labrador tea (Ledum groenlandicum)

Aspen

Rose

Bramble (Rubus ideaus)

Grass

Dogwood (Cornus canadensis)

Fireweed

Horsetail

Wild rhubarb

Ontario

Pines (Pinus strobus and P. resinosa)
Aspen
Alder

Hazel (Corylus cornuta)
(Ostrya virginiana)
Juneberry (Amelanchier sp.)
Willow

Poplar (P. balsamifera)
White birch
Grasses

Washington

Spotted catsear (Hypochoeris radicata)

Sources: (Radwan and Campbell 1968, Wolff 1978; de Vos
1964).

Snowshoe hare with winter coat.

Figure 7. Cottontail fecal pellet group.

the location of the animal and noting its compass
direction by using the transect line as a point of
reference. The flushing distance is the distance be-
tween the center of the transect and point where
the Leporid is first observed. Flushing distance and
angle will vary according to the density and height
of the cover in which the Leporid is living. In gen-
eral, the greater the density or thickness of the
cover, the shorter the flushing distance. The flushing
distance and angle will be determining factors in
the width of the transect used. The greater the ob-
servable flushing distance and angle, the wider the
transect will be and, thus, the greater the area cen-
sused (Webb 1942).

Behavioral patterns and density of vegetation
tend to preclude drive counts as a suitable technique
for cottontails and snowshoe hares. Cottontails tend
to freeze rather than flee, making a head count
difficult.

Snowshoe hares inhabit conifer stands with
thick understories, making them difficult to see and
count. However, drive counts have been successfully
used in jackrabbit studies where sagebrush was the
dominant vegetation type (Gross et al. 1974). Biases
in census data may result from a change in flushing
behavior, observer behavior, and duplicate counting
of hares. Gates (1969) used an equation to adjust for

466

Lagomorphs

flushing behavior changes in jackrabbits due to sea-
son and density. Jackrabbits in sagebrush habitat are
easily visible and thus can be censused by using
the drive count technique (F.H. Wagner, pers. com-
mun. )• Broadbooks (1965) successfully used a 30-m
(100-ft ) wide transect to census pikas on a rock
formation.

Lagomorph density can also be estimated by
mark-recapture techniques and fecal pellet group
counts. Taylor ( 1930) was one of the first investiga-
tors to use fecal pellets as an indicator of Leporid
abundance in the West. Pellet counts were also used
to evaluate Audubon's cottontail population density
in a pinyon- juniper forest in New Mexico (Kundaeli
and Reynolds 1972). However, pellet counts are
more useful for determining the habitat preference
of Lagomorphs than for determining density. The
presence of pellets was used to determine the use of
montane forest types by snowshoe hares in Utah
(Wolff et al. 1982). These methods are useful for
comparing Lagomorph populations in different years,
different habitat types, and different seasons. When
using pellet counts to estimate populations, adjust-
ments for varying defecation rates may be necessary
(See Lord 1963).

Mark-recapture techniques involve sampling a
segment of a population, marking and releasing
them, and at a later date recapturing a percentage of
the marked individuals. Mark and recapture tech-
niques are frequently used because they not only
give census data but also specific information on
factors such as sex and age, survival, and movements
of individuals. These techniques have been exten-
sively used to assess Lagomorph population densities.

The Wildlife Management Techniques Man-
ual describes several models such as the Lincoln
Index (Lincoln 1930) or Jolly Index (Jolly 1963) to
estimate the size of a population by the mark-recap-
ture method (Davis and Winstead 1980). Population
levels based on mark-recapture techniques often
underestimate the actual number of animals present
in an area (Eberhardt 1969). Otis et al. (1978) dis-
cuss the appropriateness of different models available
for estimating the population size from capture and
recapture data.

Some problems associated with mark-recapture
procedures include the loss of marks or the tend-
ency of some individuals to become "trap-happy" or
"trap-shy," all resulting in inaccurate population esti-
mates. Loss of ear tags can result in inflated popula-
tion estimates. Some problems associated with this
technique include variation in recapture rates and
the behavioral responses to traps exhibited by indi-
vidual animals (Otis et al. 1978).

Mark-recapture techniques have been used to
census brush rabbits in bush lupine ( Lupinus arbus-

tus) in Humboldt Bay, California (Shields 1958),
Audubon's cottontails in mesquite-grassland habitat
associations (Asserson 1967), Nuttall's cottontail
in shrub- juniper scrubland in Oregon (McKay and
Verts 1978), and jackrabbits in Utah (Clark 1972).
Factors affecting the trap responses of eastern cot-
tontails in Oregon included barometric pressure and
temperature, but not precipitation (Chapman and
Tretheway 1972). Methods to capture, handle, and
mark Lagomorphs are outlined in The Handbook of
Census Methods for Terrestrial Vertebrates (Feist
1982; Smith 1982; and Wolff 1982) and Wild Mam-
mals of North America (Chapman and Feldhamer
1982).

Jackrabbits can be captured in winter by using
large Havahart traps baited with alfalfa or apples (G.
Smith, pers. commun.). Cottontails can be caught
in either baited or unbaited wooden box traps (Fig-
ure 8; Edwards 1975). Leporids can also be captured
from vehicles by using a spotlight and a hand-held
net (Labisky 1959, 1968).

Lagomorphs can be marked with size 3 monel
ear tags, hair dye, or toe-clipping to identify individ-
uals (Wolff 1982). Depending on the size of the
Lagomorph, other types of markers include commer-
cially available colored discs, colored discs cut from
surveyors flagging, an alcohol-picric-acid solution,
and notching or punching holes in the ears (Figure
9). Approaching and handling Lagomorphs requires
care to avoid the flight response and subsequent
injury of the rabbit or hare while the animal is still
in the trap. Placing the Lagomorph in a cloth bag
tends to reduce stress on the animal while handling
it for marking or weighing.

Figure 8. Typical wooden box trap used for
capturing cottontails.

Lagomorphs

467

Figure 9. Black-tailed jackrabbit with ear tag
in place.

POPULATION LEVELS AND DENSITY

Although there are little specific data on the
absolute densities of Lagomorphs associated with
specific habitat types, some general guidelines can be
obtained from the literature. These guidelines will
be useful to the manager because they can compare
densities on their study areas with existing informa-
tion in the literature. Thus, the manager can gain
insight into the status of Lagomorph populations in a
particular area.

In general, Leporid populations do not reach
extremely high densities. There are of course excep-
tions as noted in localized situations with black-
tailed jackrabbits and snowshoe hares. Uniform den-
sities of more than 3 or 4 Leporids per ha ( 1 to 3
per a. ) would be considered high for most North
American species.

Population estimates for jackrabbits in the
Southwest vary between 0.2 and 1.2 per ha (0.1 to
0.5 per a.); in California they were reported at 3 per
ha ( 1 per a.), and at 35 per ha ( 14 per a. ) in culti-
vated areas of Kansas (See Dunn et al. 1982 for
references).

Audubon's cottontail in shortgrass-prairie re-
gions of Colorado reached densities of 0.02 per ha
(0.08 per a.) during the winter. Densities of Nuttall's
cottontails in shrub-juniper scrubland in Oregon
varied between 0.06 to 2.5 per ha (0.25 to 6.25 per
a.). Island populations of eastern cottontails in Mary-
land reached maximum density levels of 10.2 per
ha (25.5 per a.) in December (See Chapman et al.
1982 for references).

Snowshoe hare populations exhibit a marked
variation in density levels. During low population
cycles, densities may vary from 0.13 to 0.26 per ha
(0.32 to 0.65 per a.) to 5.9 to 11.8 per ha (14.7
to 19. 5 per a.) in a high cycle in Alberta (Keith and
Windberg 1978). Continent-wide, peaks in hare pop-
ulations have been reported as high as 38.6 per ha
(96.5 per a.) to lows of 0.12 per ha (0.3 per a.) (See
Bittner and Rongstand 1982 for references).

Pika densities fluctuate greatly, ranging from 2
to 3 per ha ( 5 to 7 per a. ) to as high as 70 to 80 per
ha (175 to 200 per a.) (Kawamichi 1984).

Pika.

DISCUSSION

In this chapter we have defined the major plant
communities that are associated with the various
genera and species of Lagomorphs found in the con-
tinental U.S. and Alaska. We have also discussed
some of the methods used to estimate Lagomorph
population sizes. However, to understand the status
of a Lagomorph population in a particular location,
season, or year, one must know a considerable
amount about the relationship between population
size and the habitat.

The relationship between Lagomorphs and their
habitats is complex. The factors involved include
those inherent in the physiology of the individual
species and those edaphic factors external to the

468

Lagomorphs

species such as weather, topography, and vegetation.
The factors which we will consider in this discussion
include weather, vegetation, reproduction, food hab-
its, predation, and their interrelations with Lago-
morph population dynamics and habitat use.

The population dynamics of Lagomorphs has
been the subject of considerable research and nu-
merous publications, particularly for the snowshoe
hare (Keith 1979) and eastern cottontail (Edwards
et al. 1979). The population dynamics of a species

has been defined as "numbers, structural and distri-
butional changes in population with respect to
changes in the rate of reproduction, survival, move-
ments, and environmental interactions, which in part
determine these rates" (Keith 1979). Much has been
written on the cyclic nature of snowshoe hare popu-
lations and the random fluctuations in cottontail
and jackrabbit populations. We define cyclic as a reg-
ular predictable phenomenon, such as the 10-year
cycle reported in snowshoe hares at northern lati-
tudes (Keith 1974). Population fluctuations are ran-
dom sporadic increases and decreases occurring

Table 4. Food of cottontail rabbits (Sylvilagus) in Utah.

Audubon's Cottontail

Nuttall's Cottontail

Eastern Utah

Saltgrass {Distichlis stricta)

Galleta grass (Hilaria sp.)

Indian ricegrass

Sand dropseed

Alkali sacaton (S. airoides)

Blue grama

Downy brome

Western wheatgrass

Spike rush {Eleocharis sp.)

Sedge

Wire grass (Juncus balticus)

Pigweed (Chenopodium sp.)

Saltbush (A. nuttallii,

A. corrugata, A. wolfii,

A. powelli)
Greasewood
Seepweed (Dondia sp.)
Russian thistle

Tansymustard (Descuraiana sp.)
Yellow cleome (Cleome lutea)
Wild rose

Sand bur (Glycyrrhiza lopidota)
Globe mallow
Prickly pear

Dogbane (Apocynum cannabinum)
Rabbitbrush
Sagebrush
Horsebrush

Northeastern Utah

Wheat grass

Willows

Knotweed (Polygonum sp.)

Spiny hopsage

Oregon grape (Berberis repens)

Bitterbrush

Wild rose

Chokecherry

Snakeweed

Rabbitbrush

Sagebrush

Pygmy Rabbit

Northeastern Utah

Rabbitbrush

Little rabbitbrush (C.

viscidiflorus)
Horsebrush
Russian thistle
Pigweed
Smooth wheatgrass (Agropyron

inerme)
Spiked wheatgrass (A. spicatum)
Giant wild rye (Elymus condensatus)
Small scale (Atriplex pusilla)
Four-winged saltbush
Western wheatgrass
Sagebrush
Globe mallow
Snakeweed

Western Utah

Galleta grass

Alkali sacaton

Saltgrass

Four-winged saltbush

Mound saltbush

Winterfat

Finchook bassia (Bassia

hyssopi : folia)
Pigweed (Amaranthus sp.)
Purple cleome
Snakeweed
Sagebrush

Source: (Janson 1946)

Lagomorphs

469

on an irregular basis, caused by man-induced
changes in habitat or uncontrolled changes in envi-
ronmental factors such as weather (Edwards et al.
1979). Fluctuations in Lagomorph population levels
may occur annually or seasonally. Several models
have been developed to explain the cause of cycles
in Lagomorph populations, but most are based on
Keith's (1974) model of habitat-hare-predation inter-
action (Bryant 1979, 1981; Wolff 1980a). Keith's
(1974) model is based on regular increases and de-
creases in population numbers of snowshoe hares
with respect to changes in survival, reproduction,
and movement of this species (Keith 1979). Simi-
larly, regular cycles have been reported in Utah jack-
rabbits (Gross et al. 1974).

The effects of environmental factors on Lago-
morph populations are difficult to assess and quan-
tify. However, environmental factors such as weather
are believed to have an impact on the regularity of
Lagomorph population fluctuations. Weather varia-
bles such as snowfall and rainfall may have a major
impact on the early survival of young cottontails
(Applegate and Trout 1976; Havera 1973). In
snowshoe hares, the depth of the snowfall can deter-
mine the suitability of winter habitat. Optimal habitat
for snowshoe hares in the montane forests of north-
ern Utah had an understory cover between 1.0 and
2.5 m (3 and 8.25 ft) high (Wolff et al. 1982).

Temperature and snow depth are correlated
with the size of snowshoe hare litters (Meslow and
Keith 1971). Similarly, cold, inclement weather af-
fects the onset of reproduction and fetal survival in
Leporids (Chapman et al. 1977).

The onset and duration of the breeding season
may be regulated by certain weather factors. The
length of photoperiod is associated with earlier onset
of the breeding season in snowshoe hares ( Meslow
and Keith 1971 ) and in cottontails (Bissonnette and
Csech 1939). The length of the breeding season
for Nuttall's cottontails was shortened when precipi-
tation was reduced. Consequently, population densi-
ties were lower in years of low precipitation.
Densities in late summer reached a high of 2.54 cot-
tontails per ha (6.35 per a.) when precipitation was
11.54 cm (4.6 in.); in contrast, population levels
dropped to 0.66 per ha (1.65 per a.) when precipita-
tion was 5.21 cm (208 in.) the following year (Verts
et al. 1984).

Reproduction is an important factor affecting
population cycles. Keith's (1974) model attributed
decreases in fecundity from small litter sizes, shorter
breeding seasons, and lower pregnancy rates, due
at least in part to the regulation of hare populations.
The onset of the breeding season has been corre-
lated with midwinter photoperiod (Meslow and

Keith 1971). Litter sizes are larger in areas where
days are longer; hence latitude is a factor in repro-
ductive output (Keith 1979). Litter sizes were also
larger in hares when over-winter weight loss did not
occur (Keith and Windberg 1978), pointing to the
importance of adequate high quality browse. There
is also considerable variation in the size of first and
subsequent litters in several species of Leporids
(Chapman et al. 1977; Bittner and Rongstand 1982).
Leporid litter sizes also vary with latitude (Conaway
et al. 1974; Dunn et al. 1982). Similarly, there is a
correlation with latitude and length of gestation pe-
riod in North American rabbits. Those species occur-
ring at more northern latitudes have shorter
gestation periods (Chapman 1984).

Lagomorphs are important in many food chains,
being prey items for many avian and terrestrial
predators (Bittner and Rongstand 1982; Chapman et
al. 1982; Dunn et al. 1982). The selection of refugia
or escape cover from predators during snowshoe
hare population "lows" was considered of major
importance in their survival. Refugia for snowshoe
hares in Alaska were dense thickets of alder or wil-
low (Wolff 1980a). Hare vulnerability to predation
increases as populations rise, forcing individuals to
occupy suboptimal areas. In Utah, these areas were
generally open, where the vegetation density profile
was less then 40% above the snowline (Wolff et al.
1982). Coyote predation on jackrabbits was greatest
when jackrabbit densities were at low or moderate
levels (Wagner and Stoddart 1972). Even in areas
with abundant plant food resources, cover is the
most important habitat component (Wolff 1980a).
Keith's (1974) model includes predation as a factor
in the population regulation of snowshoe hares, par-
ticularly in northern latitudes. Thus escape cover is
an essential component of Lagomorph habitat.

Lagomorphs eat a wide variety of plants. How-
ever, plant resins have been identified as a deterrent
in Lagomorph herbivory (Bryant 1979, 1981). These
resins have been suggested as a key to hare popula-
tion regulation (Bryant 1979). This idea is new and
requires further verification before it will be widely
accepted.

The diameter of the twigs consumed and brows-
ing intensity has been used as an indicator of "food
stress" in habitat utilized in excess of carrying capac-
ity (Wolff 1980a). Hares will starve in winter if
forced into habitats where the diameter of twigs
used for browse was greater than 3 mm (0.12 in.)
(Pease et al. 1979). Clark and Wagner (1984) ob-
served that only 20% of a browse plant "brown
sage" (Kochla americana) was consumed when jack-
rabbit populations were at their highest level. Heavy
grazing by sheep in addition to the limited foraging
by hares caused the plants to decline. Food is appar-
ently an important factor in hare population regula-
tion (Keith 1974). As hare densities increase, the

470

Lagomorphs

biomass of available food is less than that needed by
the population, which ultimately leaves a food short-
age and thus causes mortality in hare populations.

The relationship of Lagomorph populations to
their habitat is evident from studies on snowshoe
hares, jackrabbits, cottontails, and pikas. Periods of
drought forced jackrabbits to occupy cultivated
fields, which provided an abundant food source and
supported higher jackrabbit population levels (Bron-
son and Tiemeier 1959). Similarly, overgrazing by
cattle in conjunction with periods of drought cre-
ated more open, weedy areas, a habitat which also
supported higher jackrabbit populations (Taylor
et al. 1935). In the Mojave desert, jackrabbit popula-
tion densities were higher around water sources in
winter, declining substantially from spring to sum-
mer (Hayden 1966). Thus, changes in jackrabbit
populations were affected by food availability and
changes in environmental conditions brought on by
factors such as drought. Changes in cottontail popu-
lation levels have also been attributed to changing
land-use patterns (Edwards et al. 1979). In pikas, size
of territories is determined at least in part by vegeta-
tion quality (Smith 1979). Thus, the higher the qual-
ity of vegetation, the more pikas an area will
support.

Habitat patchiness or the mosaic pattern of veg-
etation within a Lagomorph's environment plays a
key role in the survival of snowshoe hares during a
population low (Wolff 1980a). All habitat types from
open suboptimal areas to optimal habitat types of
dense cover were occupied by snowshoe hares dur-
ing population highs. However, during lows, the
remaining individuals in a population tended to seek
refuge in dense thickets of spruce, willow, or alder.
In winters of high densities, hares in Alaska utilized
more open stands and consumed all of the available
browse (Wolff 1979, 1980b).

It is apparent that the inventory of Lagomorph
habitat must take into account many factors. Most of
these factors are poorly understood or known only
for a particular genus or species. This is especially
true for the pika and Alaskan hare which remain
among the least known Lagomorphs. In addition,
there are little quantitative data on habitat variables
available for most North American Lagomorphs. This
lack of quantitative data makes it difficult to assess
changes in Lagomorph abundance in relation to habi-
tat quality.

Table 5. Lagomorph species groups, reproductive characteristics, and common habitat requirements.

Reproductive

Family

Species

Characteristics

Habitat

Leporidae

Black-tailed jackrabbit (Lepus

Precocial

Desert scrub and

californicus)

dry grasslands

White-tailed jackrabbit

(L townsendii)

Antelope jackrabbit

(L alleni)

White-sided jackrabbit

(L. callotis)

Leporidae

Snowshoe hare (L americanus)

Precocial

Coniferous forests,

Alaskan hare (L. othus)

subalpine and

(Some workers feel L. othus is a

climax forest

synonym for L. timidus [Dixon et al.

1983]).

Leporidae

Audubon's cottontail {Sylvilagus

Altricial

Disturbed and

audubonii)

successional plant

Nuttall's cottontail (S. nuttallii)

communities

Brush rabbit (S. bachmani)

Eastern cottontail (S. floridanus)

Pygmy rabbit (S. idahoenis)

(Some workers consider the pigmy

rabbit to be in the genus Brachylagus

[Chapman et al. 1982]).

Ochotonidae

Southern pika {Ochotonidae princeps)

Altricial

Rocky talus at high

(Some workers consider the collared

altitudes

pika [0. p. collaris] of Alaska a

separate species [Chapman 1979; Hall

1981])

Lagomorphs

471

LITERATURE CITED

— , J.G. HOCKMAN, and W.R. EDWARDS. 1982. Cot-
tontails. Pages 83-123 in Chapman, J.A. and G.A. Feld-
hamer eds., Wild Mammals of North America. Johns
Hopkins Univ. Press. Baltimore, MD.

and MM. OJEDA. 1980. Sylvilagus flori-

ADAMS, L. 1959. An analysis of a population of snowshoe
hares in northwestern Montana. Ecol. Monogr.
29:141-170.

APPLEGATE, J.E. and JR. TROUT. 1976. Weather and the
harvest of cottontails in New Jersey. J. Wildl. Manage.
40:658-662.

ASSERSON, W.C. 1967. Upland Game Investigations: Cot-
tontail rabbit investigations. California PR Rep. W-47-
R 15. 12pp.

BEAR, G.D. and R.M. HANSEN. 1966. Food habits, growth,
and reproduction of white-tailed jackrabbits in south-
ern Colorado. Colorado State Univ., Agric. Exp. Sta.
Tech. Bull. 90. 39pp.

BEDNARZ, J. 1977. The white-sided jackrabbit in New
Mexico: distribution, numbers and biology in the
grasslands of Hidalgo County. Res. Rep. New Mexico
Dep. Game and Fish, Endang. Sp. Program, Santa Fe.
33pp.

BISSONNETTE, T.H. and AG. CSECH. 1939. Modified
sexual periodicity in cottontail rabbits. Biol. Bull.
17:364-367.

BITTNER, S.L. and O.J. RONGSTAND. 1982. Snowshoe
hare and allies. Pages 146-163 in Chapman, JA. and
GA. Feldhamer eds., Wild Mammals of North Amer-
ica. Johns Hopkins Univ. Press. Baltimore, MD.
1147pp.

BRAUN, C.E. and R.G. STREETER. 1968. Observations on
the occurrence of the white-tailed jackrabbit in the
alpine zone. J. Mammal. 49:160-161.

BROADBOOKS, HE. 1965. Ecology and distribution of
the pikas of Washington and Alaska. Am. Midi. Nature.
73:299-335.

BRONSON, F.H. and O.W. TIEMEIER. 1959. The relation-
ship of precipitation and black-tailed jackrabbit popu-
lations in Kansas. Ecology. 40:194-198.

BRYANT, J.P. 1979. The regulation of snowshoe hare

feeding behavior during winter by plant antiherbivore
chemistry. Pages 720-731 in Proc. World Lagomorph
Conf.

. 1 98 1 . Phytochemical deterrence of snowshoe hare

browsing by adventitious shoots of four Alaskan trees.
Science. 213:889-890.

CHAPMAN, JA. 1971. Orientation and homing of the
brush rabbit (Sylvilagus bachmani). J. Mammal.
52:686-699.

. 1 979. Rabbits, hares, and pikas. Pages 8 1 -98 in

Wild Animals of North America. National Geographic
Soc, Washington, DC. 406pp.

. 1984. Latitude and gestation period in New World

rabbits (Leporidae: Sylvilagus and Romerolagus).
Am. Nature. 124:442-445.

and G.A. FELDHAMER eds. 1982. Wild mammals of

North America. Johns Hopkins Univ. Press. Baltimore,
MD. 1147pp.

and DEC. TRETHEWAY. 1972. Factors affecting

trap responses of introduced eastern cottontail rab-
bits. J. Wildl. Manage. 36:1221-1226.

, A.L. HARMAN, and D.E. SAMUEL. 1977. Reproduc-
tive and physiological cycles in the cottontail com-
plex in western Maryland and nearby West Virginia.
Wildl. Monogr. 56. 73pp.

danns. Mammalian Species 136:1-8.

CLARK, F.W. 1972. Influence of jackrabbit density on
coyote population change. J. Wildl. Manage. 36:343-
356.

CLARK, W.R. and F.H. WAGNER. 1984. Role of livestock
and black-tailed jackrabbits in changing abundance of
Kochia americana Great Basin Naturalist. 44:635-
645.

CONAWAY, C.H., KC. SADLER, and D.H. HAZELWOOD.
1974. Geographic variation in litter size and onset of
breeding in cottontails. J. Wildl. Manage. 38:473-481.

CONROY, M.J., L.W. GYSEL, and G.R. DUDDERAR. 1979.
Habitat components of clear-cut areas for snowshoe
hares in Michigan. J. Wildl. Manage. 43:680-690.

DALQUIST, W.W. 1948. Mammals of Washington. Univ. of
Kansas Publ. Mus. Nat. Hist. 2:1-444.

DAVIS, C.A., JA. MEDLIN, and J.P. GRJFFING. 1975. Abun-
dance of black-tailed jackrabbits, desert cottontail
rabbits, and coyotes in southeastern New Mexico.
New Mexico State Univ. Agric. Exp. Sta. Res. Rep.
293.

DAVIS, D.E. ed., 1982. Handbook of census methods for
terrestrial vertebrates. CRC Press, Boca Raton, FL.
397pp.

and R.L. WINSTEAD. 1980. Estimating the numbers

of wildlife populations; pages 221-245 in Schemnitz,
S.D. ed., Wildl. Manage. Techniques Manual. Wildl.
Soc, Washington, DC.

DE VOS, A. 1964. Food utilization of snowshoe hares on
Manitoulin Island, Ontario. J. For. 62:238-244.

DICE, L.R. 1926. Notes of Pacific Coast rabbits and pikas.
Occ. Papers Mus. Zool., Univ. of Michigan, Ann Arbor.
166:1-28.

DIXON, K.R, JA. CHAPMAN, G.R. WILLNER, D.E. WIL-
SON, and W. LOPEZ-FORMENT. 1983. The New
World jackrabbits and hares (genus Lepus L. Part 2) —
Numerical taxonomic analysis. Acta. Zool. Fenica.
174:53-56.

DOLBEER, R.A. and W.R. CLARK 1975. Population ecol-
ogy of snowshoe hares in the central Rocky Moun-
tains. J. Wildl. Manage. 39:535-549.

DOUTT, J.K, C.A. HEPPENSTALL, and J.E. GUILDAY. 1967.
Mammals of Pennsylvania. Pennsylvania Game Comm.
Harrisburg. 238pp.

DUNN, J.P, JA. CHAPMAN, and RE. MARSH. 1982. Jack-
rabbits. Pages 124-145 in Chapman, JA. and GA.
Feldhamer eds. Wild Mammals of North America.
Johns Hopkins Univ. Press, Baltimore, MD.

EBERHARDT, L.L. 1969. Population estimation from re-
capture frequencies. J. Wildl. Manage. 33:28-39.

EDWARDS, W.R. 1975. Rabbit Trap. Illinois Nat. Hist.
Survey, Urbana, IL. 1pp.

, SP. HA VERA, R.F. LABISKY, JA. EILLES, and RE.

WARNER. 1979. The abundance of cottontails in rela-
tion to agriculture land use in Illinois (U.S.A.) 1956-
1978, with comments on mechanism and regulation.
Pages 761-789 in Proc. World Lagomorph Conf.

FEIST, D.D. 1982. Snowshoe hare (Alaska). Page 139 in
Davis, D.E. ed. Handbook of Census Methods for
Terrestrial Vertebrates. CRC Press, Boca Raton, FL.

FLINDERS, J.T and R.M. HANSEN. 1972. Diets and habi-
tats of jackrabbits in northeastern Colorado. Colorado

472

Lagomorphs

State Univ., Range Sci. Dep. Sci. Ser., Fort Collins.

12:1-29.

— and — — . 1973- Abundance and dispersion of

leporids within a shortgrass ecosystem. J. Mammal.
54:287-291.

GATES, C.E. 1969. Simulation study of estimates for the
line transect sampling method. Biometrics. 25:317-
328.

GRANGE, W.B. 1965. Fire and tree growth relationships to
snowshoe rabbits. Proc. Tall Timbers Fire Ecol. Conf.
4:111-125.

GROSS, J.E., L.C. STODDART, and F.H. WAGNER. 1974.
Demographic analysis of a northern Utah jackrabbit
population. Wildl. Monogr. 40. 68pp.

HALL, E.R. 1981. The mammals of North America, Vol. I.
John Wiley and Sons, New York, NY. 600pp.

HANSEN, R.M. and JT. FLINDERS. 1969. Food habits of
North American hares. Colorado State Univ., Range
Sci. Dep. Sci. Serv. Fort Collins. 31:1- 18..

HAVERA, S.P. 1973. The relationship of Illinois weather
and agriculture to the eastern cottontail rabbit. Illi-
nois State Water Surv. Tech. Rep. 4. 92pp. (Mimeo.).

HAYDEN, P. 1 966. Seasonal occurrence of jackrabbits

on Jackass Flat, Nevada. J. Wildl. Manage. 30:835-838.

JANSON, R.G. 1946. A survey of the native rabbits of Utah
with reference to their classification, distribution,
life histories and ecology. Unpubl. M.S. Thesis, Utah
State Agric. College. 103pp.

JOLLY, G.M. 1963. Estimates of population parameters

from multiple recapture data with both death and di-
lution-deterministic model. Biometrika. 50:113-128.

KAWAMICHI, T. 1984. Pikas. Pages 726-727 in Macdon-
ald, D. ed. The Encyclopedia of Mammals. Facts on
File Publ. New York, NY.

KEITH, LB. 1974. Some features of population dynamics
in mammals. Proc. Int. Congr. Game Biol. 11:17-58.

. 1979- Population dynamics of hares. Pages 395-

440 in Proc. World Lagomorph Conf.

and LA. WINDBERG. 1978. A demographic analysis

of the snowshoe hare cycle. Wildl. Monogr. 58. 70pp.

KUNDAELI, J.N. and H.G. REYNOLDS. 1972. Desert cot-
tontail use of natural and modified pinyon-juniper
woodland. J. Range Manage. 25:1 16-1 18.

LABISKY, R.F. 1959. Nightlighting: a technique for captur-
ing birds and mammals. Illinois Nat. History Survey
Div. Biol. Notes. 11pp.

. 1968. Nightlighting: Its use in capturing pheasants,

prairie chickens, bobwhites, and cottontails. Illinois
Nat. History Survey Div. Biol. Notes. 62. 12pp.

LLNCOLN, F.C 1930. Calculating waterfowl abundance on
the basis of banding returns. U.S. Dep. Agric. Cir.
118:1-4.

LORD, R.D. 1963- The cottontail rabbit in Illinois. Illinois
Dep. Conserv. Tech. Bull. 3:1-94.

McKAY, DO. and B.J. VERTS. 1978. Estimates of some
attributes of a population of Nuttall's cottontails. J.
Wildl. Manage. 42:159-168.

MESLOW, EC. and LB. KEITH. 1971. A correlation analy-
sis of weather versus snowshoe hare population pa-
rameters. J. Wildl. Manage. 35:1-14.

ORR, R.T. 1940. The rabbits of California. Occ. Papers
California Acad. Sci. 19. 227pp.

OTIS, D.L., KP. BURNHAM, G.C WHITE, and DR. ANDER

SON. 1978. Statistical inference from capture data
on closed animal populations. Wildl. Monogr. 63:1-
135.

PEASE, J.L., R.H. VOWLES, and LB. KEITH. 1979. Interac
tion of snowshoe hares and woody vegetation. J.
Wildl. Manage. 43:43-60.

RADWAN, MA. and D.L. CAMPBELL. 1968. Snowshoe hare
preference for spotted catsear flowers in western
Washington. J. Wildl. Manage. 32:104-108.

SHIELDS, P.W. 1958. Ecology and population dynamics of
the brush rabbit of the north spit of Humboldt
County, California. Unpublished M.S. Thesis. Hum-
boldt State College. 98pp.

SMITH, A.T. 1974a. The distribution and dispersal of pikas:
Influences of behavior and climate. Ecology. 55:1368-
1376.

. 1974b. The distribution and dispersal of pikas:

Consequences of insular population structure. Ecol-
ogy. 55:1112-1119.

. 1979. Population dynamics of pikas. Pages 572-586

in Proc. World Lagomorph Conf.

1982. Pika(Ochotona). Pages 131-133 in Davis,

D.E. ed. Handbook of Census Methods for Terrestrial
Vertebrates. CRC Press, Boca Raton, FL

SWIHART, RK and R.H. YAHNER. 1984. Winter use of
insular habitat patches by the eastern cottontail. Acta
Theriologica. 29:45-56.

TAYLOR, W.P. 1930. Methods of determining rodent pres-
sure on the range. Ecology. 11:523-542.

, C.T. VORHEIS, and P.B. LISTER. 1935. The relation

of jackrabbits to grazing in southern Arizona. J. For.
33:490-498.

VERTS, B.J., S.D. GEHMAN, and KJ. HUNDERTMARK

1984. Sylvilagus nuttallii: a semiarboreal Lagomorph.
J. Mammal. 65( 1 ):1 31-135.

VORHEIS, C.J. and W.P. TAYLOR. 1933. The life histories
and ecology of the jackrabbits ( Lepus alleni and L
californicus ) in relation to grazing in Arizona. Univ. of
Arizona Agric. Exper. Stat. Tech. Bull. 49:1-117.

WAGNER, F.H., and L.C. STODDART. 1972. Influence of
coyote predation on black-tailed jackrabbit popula-
tions in Utah. J. Wildl. Manage. 25:242-329-

WEBB, W.L. 1942. Notes on a method for censusing

snowshoe hare populations. J. Wildl. Manage. 42:159-
168.

WOLFF, ML, N.V. DeBYLE, C.S. WINCHELL, and T.R.

McCABE. 1982. Snowshoe hare cover relationships in
northern Utah. J. Wildl. Manage. 46:552-670.

WOLFF, J.O. 1978. Food habits of snowshoe hares in inte-
rior Alaska. J. Wildl. Manage. 42:148-153-

. 1979. Refugia, dispersal, predation, and geographic

variation in snowshoe hare cycles. Pages 441-449 in
Proc. World Lagomorph Conf.

. 1980a. The role of habitat patchiness in the popu-
lation dynamics of snowshoe hares. Ecol. Monogr.
50:111-130.

. 1980b. Moose-snowshoe hare competition during

peak hare densities. Proc. N. Amer. Moose Conf. &
Workshop. 16:238-254.

. 1982. Snowshoe hare. Pages 141-149 in Davis, D.E.

ed. Handbook of Census Methods for Terrestrial Ver-
tebrates. CRC Press, Boca Raton, FL.

Lagomorphs

473

23

CARNIVORES

Richard A. Spowart and Fred B. Samson

Colorado Cooperative Wildlife Research Unit
Colorado State University
Fort Collins, CO 80523

Editor's Note: Carnivores are interesting, elusive,
and controversial. They are also one of the most
difficult groups to inventory or monitor. Many
studies have focused on prey bases for carnivores,
but other habitat requirements have not been well
defined.

This chapter brings together much of the exist-
ing literature on carnivore techniques, and should
assist in study design. However, the biologist will
find plenty of opportunity for innovation when
working with these species.

INTRODUCTION

The order Carnivora is a diverse group of preda-
tory mammals. Trenchant characteristics of the order
are found in their dentition, which indicates diet.
However, many carnivores are omnivorous or even
herbivorous. The variability in diet is reflected by
the large array of carnivore habitat types and life-
styles. Some species are habitat and food specialists,
distributed within only a few plant communities;
others are habitat and food generalists, widely dis-
tributed in natural and disturbed ecosystems.

In contrast, the very adaptable and successful
coyote (Canis latrans) occupies almost every habi-
tat type in North America, whereas the endangered
black-footed ferret (Mustela nigripes) depends on
undisturbed prairie. The smallest living carnivore,
the least weasel (M. nivalis), weighs about 45 g ( 1.6
oz); the largest living carnivore, a subspecies of the
grizzly or brown bear ( Ursus arctos ), weighs more
than 700 kg ( 1,550 lbs). The hyperactive least wea-
sel feeds almost exclusively on mice and requires
about a third of its body weight in food per day; the
brown bear, however, is omnivorous and hibernates
for about half the year.

Because of the diverse ways carnivores have
evolved to exploit their environments, we chose to
treat most species separately. When important habi-
tat features are similar for several species, we com-
bined their descriptions. Within these combined
descriptions, we attempted to contrast unique habi-
tat preferences of individual species.

We have not attempted to describe all physical
features and vegetation types that occur in all habi-
tats of species that are generalists; only habitats that
support the densest populations or are considered
optimum are discussed. Carnivore species that are
rare in western North America, either because they
have been extirpated from much of their original
range or because their range extends only into the

'Current Address: Colorado Division of Wildlife. Kremmling, CO
80459
"Current Address: U.S. Forest Service, Olympia, WA 98195

Carnivores

475

periphery of that region, are also not discussed. Cer-
tain of these species were excluded largely because
of their threatened or endangered status. Species not
included are arctic fox (Alopex lagopus), black-
footed ferret, coati (Nasua nasua), jaguar {Felis
onca), jaguarundi (F. yagouaroundi), margay cat (F.
wiedi), ocelot (F. pardalis), polar bear (Ursus mari-
timiis), and red wolf (Cants rufus)..

>

mm"

fit

■H m .

>- -- — ~~* m ■ \ ±- ■'TriiniB

Arctic fox.

Important considerations in evaluating habitat
for any species are food supply, habitat size, and
interspecific competition. Food supply, particularly
the availability of prey species, often determines the
density of carnivore populations. Habitat manipula-
tions can often alter prey population size. Since habi-
tat manipulations indirectly influence the dynamics
of carnivore populations, we included a brief de-
scription of food habits of each species or species
group. Unfortunately, there is little information on
food requirements for most carnivores, but when
available, we referenced this information. Table 1
summarizes major foods and habitats of carnivores.

Small populations are more likely to become
extinct than large populations. A major influence in
population size is the availability of habitat. The size
or area of habitat must be large enough to support
self-sustaining populations of carnivores. The number
of prey species in all likelihood will also increase as
habitat size increases. This habitat feature is likely
most critical to those carnivores that are territorial,
have large home ranges, and exploit a patchily dis-
tributed food supply.

boreal and mixed-hardwood forests and tundra. In all
habitats occupied by wolves, either past or present,
availability of their principal prey, large ungulates,
likely has most influenced wolf distribution and
density where wolf control programs have been
absent. The primary objective of many studies of
wolf ecology in Alaska, Canada, Michigan, and
Minnesota has been to determine the
interrelationships of wolves and large ungulates
(Murie 1944; Mech 1966; Pimlott 1967; Kolenosky
1972; Kuyt 1972; Van Ballenberghe et al. 1975;
Peterson 1977; Fuller and Keith 1980; Hollemann
and Stephenson 1981; Gasaway et al. 1983).
Difficulties in measuring wolf and prey densities and
predation rates, as well as the influence of other
forms of wolf and prey mortality, have left many
questions unanswered. Disparate results have been
reported for daily food consumption rates of wolves
and, more importantly, the rates at which wolves kill
prey (Mech 1966; Pimlott 1967; Kolenosky 1972).
In northeastern North America, where simple
predator-prey systems exist, wolves are a major
mortality factor for prey populations (Mech 1966;
Jordan et al. 1967). In addition, declining wolf
populations have been attributed primarily to
decreasing numbers of their primary prey (Jordan et
al. 1967; Mech 1977; Mech and Karns 1977).
However, the role wolf predation plays in more
complex predator-prey systems in Alaska and the
Rocky Mountain national parks of western Canada is
unclear (Murie 1944; Cowan 1947; Gasaway et al.
1983).

Inventorying and monitoring habitat in areas of
the continental U.S., where wolves have been extir-
pated, are important to feasibility and evaluation
studies of wolf reintroductions. Most attempts to re-
establish wolves in areas of their original range have,
at best, met limited success (Mech 1966; Henshaw
and Stephenson 1974; Henshaw et al. 1975; Weise et
al. 1975). Many believe that large areas, having no
human habitation, are necessary to reestablish
wolves because of their large home ranges and social
structure (reviewed by Henshaw 1975; Henshaw et
al. 1975; Mech 1975; Brown 1983). However, in
Europe and some areas of Alaska, wolves live near
humans (Peterson 1975; Pulliainen 1975; Zimen and
Boitani 1975; Gasaway et al. 1983). The density of
prey populations over a large area may be the most
important habitat requisite determining feasibility
of reintroduction.

HABITAT FEATURES CORRELATED WITH
SPECIES GROUPS

Canidae

Wolves. Today, substantial populations of gray or
timber wolves (Cants lupus) occur in wilderness
areas and large national parks in Alaska, Canada, and
northern Minnesota and Michigan. Here they inhabit

Coyotes. The coyote has survived relentless
predator-control programs that extirpated the wolf
in most of the continental U.S. Moreover, this
ultimate opportunist has expanded its range (Gier
1975; Andrews and Boggess 1978; Berg and
Chesness 1978; Hilton 1978). Coyotes evolved in a
plains environment and were once most numerous
in western grasslands where large ungulate
populations were densest. They flourished in the

476

Carnivores

Table 1. Summary of major foods and habitats for carnivores.

1

1

Carnivore
Species

Food

Habitat

c
o

CO

O

<f>
0)
CO
3
O)

c
3

en
Q.
O

E
o

D)
CO

_l

CO

c

CD
"O
O

rr

"O en
c O)
CO CD
co *>
"D "p

CO £3

E<2

« CO

£5

= sz
Q.Q-
CD E
DC CO

CO
o

CD

en

C

03

*^ CO

2 ■-

- CD

CD "O
CD C
CO CO

CO

Li-

"O
CO

E

CO

LL

C
CO
CO
CO
CO

o

SZ
CO

Z)

CD

O)

CO
C/)

to
CD

o

CD

o

CD

tf>

o

§^

CO
O

is

So

CO

c

CO

CO
Q.
CO

sz
O

CO

CD
CO
CD

Q

CO

"O

c

O co

Q- q3
- >

co -c

CO c

_l CO

Wolf

•

•

•

•

•

Coyote

•

•

•

•

•

•

•

•

•

•

•

•

Gray fox

•

•

•

•

•

Red fox

•

•

•

•

•

•

•

•

•

Kit fox

•

•

Swift fox

•

•

Mountain lion

•

•

•

Bobcat

•

•

•

•

•

•

•

Lynx

•

•

Black bear

•

•

•

Grizzly bear

•

•

•

•

Fisher

•

•

•

•

Marten

•

Wolverine

•

•

•

Mink

•

•

•

•

River otter

•

•

•

Badger

•

•

•

Raccoon

•

•

•

•

Ringtail

•

Striped skunk

•

•

•

•

•

•

•

•

•

Hooded skunk

•

•

•

•

•

Hognosed skunk

•

•

•

•

•

•

•

Spotted skunk

•

•

•

•

•

•

Least weasel

•

•

•

Short-tailed
weasel

•

•

Long-tailed
weasel

•

•

shortgrass-steppe, semiarid sagebrush (Artemisia sp. )
grasslands, and deserts. They ranged from deserts
and plains to alpine areas of adjacent mountains.
Today, range extensions indicate that coyotes can be
successful in broken forests, from the tropics of
Guatemala to the eastern U.S., up to northern Alaska.
Altitude, latitude, and vegetation type do not seem
to restrict their survival. Range expansions and
population increases were likely caused by
elimination of gray wolves, clearing of forests, and
agricultural practices (Carbyn 1982).

The coyote's ability to expand its range and
adapt to so many habitat types is partly due to its
versatile food habits. Coyotes are primarily carnivo-
rous, but their diets depend on the food resources
most available. They easily adapt to being omnivo-
rous and, in this respect, their food habits resemble
those of foxes (Green and Flinders 1981). An exten-
sive food habits study (Sperry 1941 ) conducted in
17 western states showed that major diet items were
lagomorphs (33% ), carrion (25% ), rodents ( 18% ),
and domestic livestock (135%).

Carnivores

477

Studies conducted locally in the western U.S.
revealed large variances in coyote diets from one
area to another. Murie (1951) studied coyote food
habits on Arizona cattle ranges. The ranges consisted
primarily of open grasslands, oak (Quercus sp. ), juni-
per (Juniperus sp. ), prickly pear and cholla (Opun-
tia sp. ), and ponderosa pine (Pinus ponderosa).
In these habitats, coyote diets were complex but
contained high percentages of plant material. Juniper
berries were particularly important, followed by
prickly pear fruits. Berries of manzanita (Arctosta-
phylos pungens and A pringlei) and pods of velvet
mesquite {Prosopsis juliflora) were rare but pre-
ferred foods. Other important diet items were car-
rion, grasshoppers, and rodents. Lagomorphs were
found in only 3% of coyote scats but were rare in
this area.

Short ( 1979) found similar results in another
area of Arizona, described as a semidesert, grass-
shrub habitat, dominated by velvet mesquite, cholla,
and prickly pear. Although coyote food habits were
complex, the incidence of some foods in diets fol-
lowed seasonal patterns. Coyote food habits also
exhibited seasonal trends in a ponderosa pine forest
of Arizona (Turkowski 1980). In contrast, coyote
diets in other western states have been less varied
and contained primarily mammals. Clark ( 1972)
reported that coyotes in sagebrush habitat of north-
eastern Utah and south-central Idaho ate about 90%
animal matter; black-tailed jackrabbits (Lepus califor-
nicus) approached 75% of their year-round diet.
This dependence on a single prey species influences
coyote density in this region (Clark 1972; Wagner
and Stoddart 1972; Johnson and Hansen 1979).

In northeastern California, meadow voles (Mi-
crotus sp.) occurred in about half of all coyote scats
analyzed (Hawthorne 1972). Other important diet
items were mule deer and cattle, probably eaten as

carrion. Mule deer were also important in coyote
diets in two areas of southern Utah ( Pederson and
Tuckneld 1983). In central Wyoming, mule deer,
pronghorn antelope (Antilocapra americana),
white-tailed jackrabbits (L. townsendii), and desert
cottontail (Sylvilagus audubonii) were present in
63% of coyote scats (Springer and Smith 1981).

Although coyotes are extremely adaptable to
whatever foods are available, like most species of
carnivores, their populations are regulated largely by
food abundance (Clark 1972; Nellis and Keith 1976;
Todd and Keith 1983). Population densities of coy-
otes and their principal prey are strongly correlated.
Food availability appears to have a strong influence
on litter size and survival but, as yet, the carrying
capacity of coyote habitat in terms of prey abun-
dance cannot be evaluated on an absolute basis. Of
the few studies conducted to estimate food require-
ments of coyotes (Clark 1972; Wagner and Stoddart
1972), disparate results have been obtained (re-
viewed by Gier 1975). Furthermore, the decline in
prime habitat may also cause decreases in coyote
density (Andelt and Andelt 1981 ).

Foxes. Like the coyote, the gray fox (Urocyon
cinereoargenteus) and red fox (Vulpes vulpes) have
adapted to a wide range of habitat types and foods.
Most of the gray fox habitat consists of shrublands
and brushy woodlands on hilly, rough, rocky, or
broken terrain. In the western U.S., the gray fox

478

Carnivores

favors chaparral, woodlands of pinyon-juniper
(Pinusjuniperus sp. ) or oak, along rocky hillsides,
mountainsides, and washes. Chaparral vegetation
of the foothills of the Pacific states includes
ceanothus {Ceanothus sp. ), chamise (Adenostoma
fasciculatum), manzanita (Arctostaphylos sp. ),
mountain mahogony (Cercocarpus sp. ), and oak.
Here the gray fox supplements its primary diet of
rodents with manzanita berries (Ingles 1965).

In Zion National Park, gray foxes are abundant
in blackbrush (Coleogyne ramosissima), brushy
meadows, open meadows, and pinyon-juniper and
ponderosa pine at lower elevations (Trapp and Hall-
berg 1975). Here the gray fox is more a herbivore,
insectivore, or scavenger than a carnivore. Juniper
berries are an important food during winter, as is
mule deer carrion during late winter and prickly
pear fruits during autumn. Similar results were ob-
tained from a food habits study in an Arizona ponde-
rosa pine forest (Turkowski 1980).

The red fox's unspecialized way of life has al-
lowed it to adapt and thrive in many habitat types.
While it ranges from deep forests to the most ex-
posed tundra, it prefers a mixture of forest and
meadows. In the continental U.S., the red fox is
widespread except in the Great Plains and the ex-
treme Southeast and Southwest. One of the densest
populations of red foxes in North America inhabits
southwestern Wisconsin (Richards and Hine 1953).
This country is a patchwork of woodlots, cropland,
pasture, and stream bottoms. By following fox trails
in the snow, Schofield ( I960) found that red foxes
in Michigan, at least during winter, preferred low-
land brush and oak woods but avoided swamps. In-
gles (1965) believed that sheep grazing in meadows
adversely affected red fox populations in the western
mountains by reducing their rodent food supply.
The overall implication is that although the red fox
can survive in many habitat types, it prefers those
with a mixture of plant communities. Major factors
determining habitat selection are probably cover
requirements and food availability.

Red foxes are opportunistic feeders, eating foods
in proportion to their availability (Errington 1935;
Scott 1955). Thus, red fox diets often show seasonal
changes in composition (Korschgen 1959). Most
red fox food habit studies reviewed by Abies ( 1 975 )
showed major foods to be small rodents, rabbits
(Sylvilagus sp. ), wild fruits and berries, and insects.

The kit fox (Vulpes macrotis) and swift fox
{V. velox) inhabit the deserts and Great Plains of
North America, respectively — areas not occupied by
the red fox. An exception is in western Kansas
where the red fox is usually found near towns, and
the swift fox, in open, often grassland areas. The

Kit fox.

kit fox is the most specialized of the North American
canids for desert existence. It prefers open, level,
sandy ground with low desert vegetation.

Egoscue ( 1962) studied kit foxes in a western
Utah desert that contained three plant communities:
vegetated or stabilized dunes, greasewood (Sarcoba-
tiis vermiculatus) flats, and shadscale (Atriplex con-
fertifolia) flats. Vegetated dunes have a diversified
plant cover dominated by desert shrubs, including
fourwinged saltbush {Atriplex canescens), rabbit-
brush (Chrysothamnus sp. ), greasewood, horsebrush
(Tetradymia sp. ), and shrubby buckwheat (Erio-
gonum dubium ). With a single exception, no kit fox
dens were discovered in this plant community.
Greasewood flats are dominated by widely spaced
greasewood shrubs, ltol.5m(3to5ft) high.
Ground cover between shrubs is sparse and consists
primarily of grey molly (Kockia americana), seep-
weed (Suaeda torryand), and shadscale. Numerous
kit fox dens were located in this community, which
also was a favorite foraging area. Shadscale flats con-
sist of sparsely growing vegetation averaging =g60cm
(=524 in.) high. Dominant species were shadscale,
inkweed, and grey molly. This community was pri-
marily important as a hunting ground for black-tailed
jackrabbits, which composed 90 to 95% of kit fox
diets during the pup-rearing season. Although jack-
rabbit populations were not especially dense in
shadscale flats, their numbers did not seem to fluc-
tuate as sharply as in other communities. Thus, this
community appeared to provide a stable food
source.

The swift fox is considered by some to be a
subspecies of the kit fox (Stains 1975). The swift
fox, once common and widespread on the high
plains of the central U.S. and Canada, now has a
much more restricted distribution (Kilgore 1969;
Moore and Martin 1980). The original swift fox habi-
tat, the shortgrass plains of the Oklahoma Panhandle,

Carnivores

479

was dominated by buffalograss {Buchloe dacty-
toides) and blue grama (Bouteloua gracilis) (Kil-
gore 1969). Other grass species were bluestem
(Andropogon scoparius), wiregrass (Aristida sp.),
and sideoats grama (B. curtipendula). Much of this
region is now extensively cultivated; as a result, the
original vegetation has largely been replaced by
grain crops and weeds. In Wyoming, recent records
of swift foxes have come from areas of gently rolling,
shortgrass prairie, dominated by buffalograss and
blue grama (Floyd and Stromberg 1981 ). In this area,
the native shortgrass prairie is interspersed with
winter wheat, alfalfa, and fallow fields. This habitat is
very similar to that described for swift foxes in the
Oklahoma Panhandle by Kilgore ( 1969).

Felidae

Mountain Lions. In western North America,
mountain lions (Felis concolor) are generally
associated with mountainous terrain, canyons, and
rimrock. Here they feed primarily on mule deer
(Robinette et al. 1959; Barnes I960; Spalding and
Lesowski 1971; Toweill and Meslow 1977; reviewed
by Anderson 1983 and Ackerman et al. 1984).

In southern Utah, mountain lion habitat is vege-
tatively and topographically diverse (Barnes I960;
Ackerman 1982; Hemker 1982). Desert shrub and
sagebrush-grassland communities occur at lower
elevations (1,350 to 1,800 m [4,445 to 5,940 ft]).
Pinyon-juniper woodlands, oakbrush (Quercus gam-
belii), and ponderosa pine forests dominate mid-
elevations (1,800 to 2,700 m [5,940 to 8,910 ft]).
Stands of quaking aspen (Populns tremidoides),
Engelmann spruce (Picea engelmann ii), and white
fir (Abies concolor) interspersed with subalpine
meadows occur above 2,700 m (8,910 ft). Deep,
rocky, vertical-walled river canyons within these
communities contain riparian vegetation, including
Fremont Cottonwood (P. fremontii) and willow
(Salix sp.).

Vegetation and topography of the Idaho Primi-
tive Area, where Hornocker ( 1969, 1970) and Sei-
densticker et al. (1973) conducted extensive
mountain lion research, are also diverse. Plant com-
munities are interspersed, depending on site charac-
teristics including elevation, slope, and aspect.
Engelmann spruce-subalpine fir (A lasiocarpa) and
ponderosa pine-Douglas-fir (Pseudotsuga menziesii)
associations grow at higher elevations and in pro-
tected drainages. At lower elevations and on exposed
slopes, curlleaf mountain mahogany (Cerococarpus
ledifolius), bitterbrush (Purshia tridentata), and big
sagebrush (A tridenta ta)-bunchgrass associations
occur.

In California, mountain lions occur primarily
between 600 and 1,800 m ( 1,980 and 5,940 ft) ele-
vations in mixed conifer and brush habitats (Koford

1977). At higher elevations, in pure stands of coni-
fers, and at lower elevations, in pure stands of cha-
mise brush (Adenostoma fasciculata), mountain
lions are rare.

Within these western habitats, stream courses
and ridgetops are frequently used as travel lanes and
hunting routes by mountain lions (Barnes I960).
Riparian vegetation along streams provides cover for
lions traveling in open areas, and ridgetops with
cover allow undetected surveillance.

From these studies, mountain lions appear to
prefer habitats that are vegetatively and topographi-
cally complex. This habitat characteristic and prey
availability determine the amount of space a resident
mountain lion requires. Because of large home
ranges, territoriality, and population social structure,
large areas of wilderness or remote habitat are essen-
tial to sustain viable mountain lion populations (Hor-
nocker 1969, 1970; Koford 1977; Hemker 1982;
reviewed by Anderson 1983). Mountain lions wan-
der close to civilization (Barnes I960), and some
populations live near humans (Koford 1977). This is
especially true in California, where there is a morato-
rium on mountain lion hunting. It is questionable,
however, whether mountain lion populations can be
sustained in this increasingly modified habitat
(Figure 1A).

Bobcats. Bobcat (Felis rafus ) and mountain lion
habitat preferences are similar in that both use a
variety of habitats. Bobcats still inhabit much of their
former range in the eastern and southern U.S.
(Figure IB), where the mountain lion has been
extirpated. Western bobcats prefer rocky canyons at
elevations between 1,400 and 2,100 m (4,620 and
6,930 ft ) with ledges and areas of dense vegetation
(Young 1958; Gashwiler et al. 1961; McCord and
Cardoza 1982). Common tree and shrub species of
western habitats are manzanita, mountain mahogany,
pinyon pine, sagebrush, and juniper. In the
southwestern and western U.S., bobcats have adapted
to living in even the driest deserts (Young 1958;
Jones and Smith 1979).

Prey abundance, topography, and vegetation
structure affect bobcat social structure and home
range size (Bailey 1974; Beason and Moore 1977). In
turn, social organization profoundly affects popula-
tion density and the ways bobcats use their environ-
ment. Competitive interaction with coyotes also
likely affects bobcat population density and behavior
(Linhart and Robinson 1972).

Lynx. Lynx (F. lynx) are more restrictive in habitat
and food selection than bobcats, making them more
vulnerable to a changing environment. Where both
these species are sympatric, competition, should
it occur, may be more detrimental to the lynx
(Parker et al. 1983). Dominant tree species of boreal

480

Carnivores

PAST

PRESENT

Figure 1A. Historical distribution of the mountain lion.

PAST

PRESENT

Figure IB. Historical distribution of the bobcat.

Carnivores

481

forests inhabited by lynx are balsam fir
(A balsamea), black spruce {P. mariana), white
spruce (P. glauca), and paper birch (Betnal
papyrifera) (Brand et al. 1976; Parker et al. 1983).

In Alberta, lynx inhabit a variety of vegetation
types, ranging from undisturbed forest to 90% culti-
vated farmland (Nellis and Keith 1968; Nellis et al.
1972). In Montana, lynx preferred dense stands of
lodgepole pine (P. contorta) (Koehler et al. 1979).
Snowshoe hares were also most abundant in this
habitat type. Throughout their range, lynx depend
on snowshoe hares for most of their diet (Saunders
1963; van Zyll de Jong 1966; Nellis and Keith 1968;
Nellis et al. 1972; Brand et al. 1976; More 1976;
Koehler et al. 1979; Parker et al. 1983). This de-
pendence is reflected in cyclical fluctuations of lynx
populations with changing snowshoe hare densities
(Figure 2) (Keith 1963; Brand et al. 1976; Brand and
Keith 1979). Because of this phenomenon, forest
management plans that incorporate habitat require-
ments of snowshoe hares are, in effect, management
plans for lynx (Parker et al. 1983). Clear-cutting
and burning of mature conifer forests are likely to
initially decrease hare and lynx populations, but
in the long term, an increase in food and cover for
hares should increase densities of both hares and
lynx (Koehler et al. 1979; Parker et al. 1983).

Ursidae

Black Bear. Bray and Barnes ( 1967) noted that
forested environments and food availability were
emphasized in almost all habitat descriptions of black
bear (Ursus americanus). Forests provide escape
cover, thermal cover, and den sites. However, in
western North America, black bear habitat ranges
from desert chaparral to closed coniferous forest

(Jonkel and Cowan 1971; Amstrup and Beecham
1976; Lindzey and Meslow 1977; Kelleyhouse 1980;
LeCount 1980; Graber 1982; Novick and Stewart
1982). The physical features and vegetation of black
bear habitat are exceedingly variable, reflecting
differences in latitude, elevation, slope, aspect,
precipitation, and land-use patterns. Because black
bears have adapted to so many habitat types, there is
no clear consensus on which habitat components
are most important (Treadwell 1979). Determining
local habitat use may be needed for sound manage-
ment of this species.

In whatever ecosystem black bears occur, food
availability greatly influences their abundance and
distribution. Jonkel and Cowan ( 1971 ) found that
seasonal food production dictated habitat use by
black bears in spruce-fir {Picea-Abies) forests in
northwestern Montana. The spruce-fir/mountain
lover (Pachistima myrsinites) association was an
important component of black bear habitat during all
seasons. Other habitat components were important
seasonally: dry meadows in early spring, snowslides
and stream bottoms in early and midsummer, and
spruce-fir/rustyleaf (Menziesia fetrnginea) and
spruce-fir/bcargrass (Xerophylum tenax) associations
in fall. Recent clearcuts were avoided. The home
range of bears included the seasonal food sources
and, when a seasonal food was abundant, bears with
adjacent home ranges congregated in these habitats.

Lindzey and Meslow ( 1977) studied home range
and habitat use by black bears in a western hemlock-
Sitka spruce (7: heterophylla-P. stichensis) forest in
southwestern Washington. They found that black
bears used vegetation types disproportionately to
their availability, depending on availability of food

250-

200

Figure 2. Cyclical fluctuations
in lynx and snowshoe hare
populations, showing
10-year cycle phenomenon,
(after Keith 1963).

o
o
o

*. 150--

uT 100 -

50--

Hare

Lynx

1 1 1 1 n~i frTTTn 1 1 1 1 h 1 1 1 1 1 1 1 m 1 1 iii

1850

1860

1870

1880

1890

1900

482

Carnivores

and cover. The plant communities that followed
clear-cutting provided concentrations of foods, espe-
cially huckleberries (Vaccinium sp.).

In a mixed coniferous forest of Douglas-fir, pon-
derosa pine, subalpine fir, and Engelmann spruce in
Idaho, seven forage species (wild onions [Allium
sp.], waterleaf [Hydrophyllum capitatwn], biscuit
root [Lomatium dissectum], bitter cherry [Prunus
emarginata], chokecherry [P. virginiana], mountain
ash [Sorbus scopulina], and huckleberry [Vaccinium
globulare]) were highly correlated with locations
of radio-collared black bears (Amstrup and Beecham
1976). Locations of black bears corresponded to
areas of maximum food availability, which depended
on phenological stages of these seven key food
plants. Researchers in California also correlated habi-
tat use by black bears with seasonal food availability
(Piekielek and Burton 1975; Kelleyhouse 1980; Nov-
ick et al. 1981; Novick and Stewart 1982; Grenfell
and Brody 1983). Habitat types ranged from chapar-
ral, oak woodlands, and coniferous forest.

Grizzly Bears. In contrast to the importance of
forested habitats to black bears, the importance of
nonforested habitats to grizzly bears has been
emphasized by several researchers. Herrero (1972,
1978) hypothesized that grizzly/brown bears evolved
to exploit nonforested, periglacial environments.
Martinka (1976) stressed the importance of fire in
reducing or eliminating forest canopies to improve
habitat for grizzlies. He stated that without natural
fire, successional advance towards mature forest
creates conditions more favorable for black bears
than grizzlies. Particularly important open habitats
are wet meadows and alluvial plains that supply
grizzlies with key forbs and rhizomatous grasses
(Singer 1978).

Craighead et al. ( 1982) included extensive tim-
ber cover, as well as open grasslands and meadows,
in their descriptions of optimum grizzly habitat in
the continental U.S. Grizzly/brown bear populations
in Alaska and Canada apparently have no intrinsic
need for extensive timber cover, as they thrive in
open areas. However, grizzly populations in the con-
tinental U.S. may require forested habitat for isola-
tion from human activities. Here a combination of a
wide range of vegetation types characterizes prime
grizzly bear habitat; mountain parks, grasslands inter-
spersed with timber, alpine meadows, and talus
slopes are necessary for feeding and social activities.
Alder {Alnus tenuifolia) thickets, lodgepole pine
"downfalls," and other dense vegetation are pre-
ferred bedding sites.

Large geographical areas are needed to provide
a variety of alternative foods. Major foods are car-
rion, ungulates, rodents, insects, berries, pine nuts,
green vegetation, bulbs and tubers and, in some

areas, fish (Mealy 1980). Habitat use by grizzlies is
greatly influenced by the seasonal availability of
these foods. In spring, the location of carcasses of
ungulates that died during the winter or individuals
that emerged from the winter in poor condition
are important (Cole 1972; Houston 1978). In spring
and summer, areas with high population densities
of rodents or insects, fish-spawning sites, or succu-
lent herbaceous vegetation or bulbs and tubers are
important. In fall, berries, pine nuts, and ungulates
crippled during hunting seasons are important food
sources. Because of the conflict between grizzlies
and humans, along with human land uses, wilderness
areas are essential to separate them. Grizzly habitat
must be isolated from developed areas and should
receive only light recreational, logging, and livestock
use. Except for humans, the grizzly has no enemies
that restrict its use of habitat.

Arboreal, Aquatic, and Terrestrial
Furbearers

Fisher, Marten, and Wolverine.

Fisher (Martes pennanti). These mustelids are
generally associated with climax coniferous forests
(de Vos 1952; Hornocker and Hash 1981; Powell
1982; Allen 1982, 1983; Douglass et al. 1983;
Spencer et al. 1983; Hargis and McCullough 1984)
and have large home range sizes (Koehler and
Hornocker 1977; Buck et al. 1978; Powell 1982;
Hornocker and Hash 1981). Fishers prefer
continuous coniferous and mixed coniferous/
deciduous forests with closed canopies ( reviewed by
Powell 1982; Allen 1983). Dense overhead cover is
the one characteristic common to all habitats
preferred by fishers. Canopy closure of 80 to 100%
is selected and canopy closure of less than 50% is
avoided. Within these habitats, fishers eat a variety of
foods, including small mammals, carrion, birds and
their eggs, insects, reptiles, amphibians, and various
fruits and nuts (reviewed by Strickland et al. 1982).

Grinnell et al. (1937) and Schempf and White
( 1977) reported that fishers in California are most
commonly found in Douglas -fir and mixed conifer
forests between 610 and 2,440 m (2,000 and 8,000
ft) elevations. Here an unusual but important diet
item is false truffles (Rhizopogon sp.; Grenfell and
Fasenfest 1979).

Marten {Martes americana ). The effects of fire
and timber harvest have generally been considered
harmful to fisher and marten habitat (de Vos 1952;
reviewed by Powell 1982). However, whether this is
true for marten habitat is subject to disagreement,
because marten appear to be less dependent than
fisher on old growth forest. Marten habitat use varies
with the season.

Carnivores

483

During winter, marten prefer dense overhead
cover of mature coniferous forests (Koehler and
Hornocker 1977). In Idaho, marten activity was
highest in stands having an Engelmann spruce-subal-
pine fir overstory, 100 years old or more, and a can-
opy cover greater than 30% ; mesic sites also had
high marten activity. Similarly, marten in northern
California preferred stands of lodgepole pine or red
fir (A magnified) with closed canopies (Spencer
et al. 1983). Within both these habitat types, marten
selected sites with Douglas squirrel (Tamiasciurus
douglasii) feeding sign. Old-growth red fir with
large snags, stumps, and logs was an especially pre-
ferred habitat. In Yosemite National Park, marten
selected dense cover less than 3 m (9 ft) above
snow level in lodgepole pine and western juniper
forests (Hargis and McCullough 1984). Tree trunks
allowed access to the subnivean (below the snow
surface) zone, and logs served as den sites and hunt-
ing areas. Open areas were avoided. Preference for
closed canopies may also be due to less snow accu-
mulation, which would permit marten to conserve
energy of movement (Raine 1982). Another means
to conserve energy is use of red squirrel (Tamias-
ciurns hudsoniens) middens as roosting sites
(Buskirk 1984).

Soutiere (1979) hypothesized that prey availa-
bility in dense forests rather than density of over-
head cover, per se, explained winter habitat use by
marten. Small mammals, especially the red-backed
vole (Clethrionomys gapperi), make up the major
portion of marten diets during winter (Cowan and
MacKay 1959; Weckwerth and Hawley 1962; Sou-
tiere 1979). Red-backed voles are associated with
downed logs and mesic conditions of uncut conifer-
ous forests (Koehler and Hornocker 1977).

In summer, effects of timber harvest and fire
may be beneficial to marten. In Idaho, Koehler and
Hornocker (1977) reported that plant communities
created by fire provide marten with important sum-
mer foods. Streeter and Braun ( 1968), working in
Colorado, also noted that marten do not require the
cover of climax forests during summer. Contrary
to these reports, Spencer et al. (1983) found that
marten in the northern Sierra Nevada avoided open
areas during all seasons. Soutiere ( 1979) concluded
that marten in the Northeast were not limited to
climax forests, and partial-cutting methods of timber
harvest were compatible with the preservation of
marten habitat. In California, Hargis and McCullough
( 1984) found that mixed-aged forests were impor-
tant in providing protective cover and subnivean
access over a wide range of snow conditions.

Wolverine (Gulo gulo). Like marten and fisher,
the wolverine is associated with boreal forests. In
Montana, wolverines selected subalpine fir cover
types throughout the year; this selection was strong-
est during summer (Hornocker and Hash 1981 ).

Serai lodgepole pine and western larch (Larix occi-
dentalism sites were also used by wolverines. Large
areas of mature timber were preferred, followed
by ecotonal areas and rocky areas of timbered
benches. Areas of dense, young timber; recent burns;
and wet meadows were avoided.

In the Pacific states, wolverines are associated
with coniferous forests containing Shasta red fir (A
magnifica), lodgepole pine, and western hemlock
(T. heterophylla); at higher elevations, subalpine fir,
Alaska cedar (P. albicanlis), and western juniper
(J. occidentalis) (Ingles 1965). In Canada, wolver-
ines inhabit tundra between the northern tree line
and arctic coasts as well as boreal forests (Banfield
1974).

Wilderness is important in protecting wolverine
populations because they are highly susceptible to
human-caused mortality (van Zyll de Jong 1975).
Hornocker and Hash ( 1981 ) found no difference in
wolverine density between wilderness and nonwild-
erness areas, but concluded that wilderness was es-
sential as a refuge and reservoir for wolverine
populations. Furthermore, large areas of habitat are
needed to support wolverine populations because of
the wolverine's scavenging life-style, which requires
a solitary existence and a large home range. Carrion
is a mainstay in wolverine diets, especially during
winter (Rausch and Pearson 1972; van Zyll de Jong
1975; Hornocker and Hash 1981 ). Van Zyll de Jong
( 1975) hypothesized that wolverine abundance is
correlated with biomass and turnover of large ungu-
late populations.

Mink and River Otter.

Mink (Mustela vison ). These amphibious
mustelids are widespread, but uncommon, and have
irregular distributions in western North America.
With the exception of otters, mink lead a more
aquatic existence than any other mustelid. They are
often found near water, where a diet of muskrats
(Ondatra zibethicus), fish, crayfish, frogs, and
aquatic invertebrates are available (Haley 1975;
Gilbert and Nancekivell 1982). In Canada and Alaska,
habitat types with some of the densest mink
populations are tidal flats and ocean beaches (Burns
1964; Banfield 1974). Unlike otters, mink are also
common on land where they primarily hunt rodents
and lagomorphs. In inland Canada, snowshoe hares
are often their principal prey.

River Otter (Lutra canadensis). The more
aquatic river otter depends on fish for a larger per-
centage of its diet than does the mink, especially
in stream habitats (Melquist et al. 1981 ). Mink once
concentrated on mammals and birds in stream habi-
tats but switched to eating mostly fish at lakes (Gil-
bert and Nancekivell 1982). River otters, however,

484

Carnivores

ate primarily fish in both habitats. Besides fish, im-
portant river otter foods are crayfish, frogs, turtles,
and aquatic insects. Muskrat, beaver {Castor cana-
densis), and water birds occur infrequently in river
otter diets, but in some areas they may be seasonally
important (Greer 1955).

Throughout their range, river otters inhabit
streams, rivers, lakes, marshes, and ocean bays where
fish populations are dense. In Canada, they occur
north beyond the tree line in tundra lakes and
streams (Banfield 1974). In west-central Idaho, prey
availability had the greatest influence on habitat use
by river otters, but adequate shelter was necessary
for extensive habitat use (Melquist and Hornocker
1983). Rather than excavate their own dens, river
otters use dens constructed by other animals, often
those of beavers (Haley 1975; Melquist and Hor-
nocker 1983). They also use natural or man-made
structures as dens. Habitat use in Idaho, determined
by biotelemetry, indicated that river otters preferred
valley to mountain habitats and stream-associated
habitats to ponds, lakes, and reservoirs (Melquist and
Hornocker 1983). This was believed to be true be-
cause stream habitats provided more adequate es-
cape cover and shelter and less human disturbance.
However, if food and shelter were adequate, river
otters exhibited high tolerance toward human
activity.

During winter, habitat use by mink and river
otters is complicated by ice-covered waterways
(Banfield 1974; Melquist and Hornocker 1983). Hab-
itat is limited to open water during early winter.
Outflows from lakes are particularly favored haunts
at this time. In late winter, water levels usually drop
below ice levels in rivers and lakes, leaving a layer of
air that allows mink and river otters to travel and
hunt under the ice.

Water quality and quantity are habitat variables
that directly and indirectly affect mink and river
otters. As secondary or higher-order consumers,
these carnivores concentrate water pollutants (Cum-
bie 1975; Henry et al. 1981). Polluted or inadequate
water quantity also adversely affects prey popula-
tions, thereby diminishing the carrying capacity of
mink and river otter habitat.

Badgers, Raccoons, Ringtails, Skunks, and
Weasels. The geographic distribution of most of
these small- to medium-sized carnivores overlap
broad areas in western North America. Many of these
species also have similar food habits. Being
opportunists, they eat foods locally or seasonally
available and, except for the weasels (Mustela sp.),
all are omnivorous. Because of these similarities, we
have grouped these species so important habitat
features can be compared and constructed.

River otters.

Carnivores

485

Badger (Taxidea taxus). Badgers are wide-
spread in the western U.S., occurring in all plant
communities except the rain forests of Washington
and Oregon. They occur from sea level to alpine
meadows and from deserts to dense coniferous for-
ests (Long 1973; Long and Killingley 1983). Conse-
quently, their general food habits are not too
surprising. Badgers eat almost any species of verte-
brate or invertebrate they can capture, but especially
rodents captured underground (Messick and Hor-
nocker 1981; Lampe 1982). They are most numer-
ous in sagebrush-grasslands where pocket gophers
(Thomomys sp.) and ground squirrels (Spermophi-
his sp.) abound. Their home range size varies dra-
matically with the season (Sargent and Warner 1972;
Lindzey 1978; Lampe and Sovada 1981 ), being small-
est during winter when they can enter a state of
torpor and reduce their time aboveground by 93%
(Harlow 1981).

Badgers.

Raccoon (Procyon lotor). Raccoons are com-
mon in forests of the western U.S. at low to middle
elevations, where they greatly depend on riparian
areas (Lotze and Anderson 1979; Hart 1982). They
prefer deciduous stands over coniferous forests, but
even inhabit basaltic outcroppings in sagebrush eco-
systems. They are absent from deserts, high eleva-
tions, and the boreal forests of Canada and Alaska.
Their habitat must include den sites; hollow
branches are preferred, but hollow trunks and logs,
caves or crevices, and attics of abandoned farm
buildings are also used (Berner and Gysel 1967;
Lynch 1974). Raccoons are probably the most omni-
vorous native carnivores in the western U.S., feeding
on whatever foods are available. They forage exten-
sively in farmyards and wetlands (Ellis 1964; Urban
1970; Fritzell 1978; Greenwood 1982) and have
adapted to living in suburban residential communi-
ties (Hoffman and Gottschang 1977).

Ringtail (Bassariscus astutas). Depending on
habitat, ringtails may also be omnivorous. They typi-
cally inhabit rocky and brushy canyons and talus
area. They are also common in shinnery oak (Q. har-
varclii) woodlands and occur in live oak (Q. virgi-
niana) savannas.

In Colorado, ringtails are usually associated with
arid canyon and mesa country in the pinyon-juniper
vegetation type. However, they have been reported
in a wide variety of habitats, including an Engelmann
spruce-lodgepole, pine-aspen community (Rutherford
1954). Similarly, in southern Nevada, Bradley and
Hansen (1965) reported ringtails in many habitat
types. They concluded that ringtails were present in
all habitats except the most arid lower deserts.

Skunk (Conepatus sp., Mephitis sp., Spilogale
sp.). The striped skunk (M mephitis) occurs in a
wider range of habitats and is more omnivorous than
the ringtail (Wood 1954; Verts 1967; Ewer 1973;
Haley 1975; Schowalter and Gunson 1982). This, the
most common mustelid, is absent from colder
climes, likely due to its inability to remain dormant
for such long winters (Haley 1975). Within its range,
the striped skunk flourishes from city parks to wil-
derness areas because of its generalized habits. Pre-
ferred habitats are difficult to delineate for a species
so adaptable and widely distributed.

Verts (1967) concluded that relative abundance
of striped skunks determined their preferred habitat.
Using this criterion, he found that the extent and
location of woodlots in farmlands of northeastern Il-
linois were important habitat factors. Presumably,
woodlots provided needed cover and den sites. Con-
trary to earlier reports, striped skunks were more
abundant in intensively cultivated areas than where
forest, brushland, and farmland were intermixed
(Verts 1967; Smith and Verts 1982). Storm (1972)
found that striped skunks used pasture and hay crops
more frequently than expected because they were
conducive to night foraging and retreats, day or
night. Other crops and uncultivated areas were used
less than expected. In the open, dry country of
Idaho, the importance of riparian growth along
streams was deemed important for the same reasons
(Larrison and Johnson 1981). Along the Pacific
Coast, sand dunes covered with sufficient vegetation
to permit digging of dens are excellent habitat (In-
gles 1965). Here striped skunks scavenge along
beaches at night for washed-up, dead marine animals.
Similarly, in the Mississippi Lowland Region of Mis-
souri, available denning habitat determines the distri-
bution and population density of striped skunks
(Verts 1967). The high water table makes ground
dens unsuitable in many areas of this region.

486

Carnivores

Life-styles of the hooded skunk (M. macroura)
and hognosed skunk are similar to those of the
striped skunk. The hooded skunk prefers rocky and
brushy canyons and riparian areas. Two or more
species of hognosed skunks have been recognized
(Hall and Kelson 1959) in these areas, but may only
be geographical races of the same species (Haley
1975). This skunk inhabits desert valleys, grasslands,
and partly wooded mountain foothills.

Spotted skunks (some authors recognize a west-
ern species, S. gracilis, and an eastern species, S.
putorius) are less common than the striped skunk
and less widely distributed. They generally prefer
more open country than striped skunks, but still
require some form of cover such as a fence row or
gully vegetation between their den and foraging
areas (Crabb 1948). In agricultural areas, they often
den under old farm buildings or in abandoned
ground squirrel burrows. In the southwestern U.S.,
spotted skunks inhabit rough, broken country below
2,400 m (8,000 ft). They can survive in deserts and
drier regions of Nevada, New Mexico, and Arizona,
where the striped skunk cannot (Haley 1975). How-
ever, they depend on riparian areas in arid country
(Larrison and Johnson 1981). Although spotted
skunks are highly omnivorous, as are striped skunks,
rodents are relatively more important and arthro-
pods correspondingly less so in their diets.

Weasel (M us tela nivalis, M. frenata, M. ermi-
nea). Unlike the rest of the species in this group,
weasels are primarily carnivorous, specializing in
killing rodents, lagomorphs, and birds. Like other
species in this group, weasels occupy a broad range
of habitats including boreal and deciduous forests,
shrub-steppe regions, grasslands, and agricultural
areas (Hall 1951). Their distributions are not closely
linked to those of any certain prey species, but may
be limited by competition with other weasel species
(Simms 1979).

Short-tailed weasel.

areas around farmsteads and the edges of urban
areas. However, Ingles ( 1965) reported that short-
tailed weasels in California, Oregon, and Washington
inhabit high mountain red fir, lodgepole pine, and
subalpine forests. Rodents, especially Microtus sp.,
are important in the diet of ermine, but shrews and
birds are also eaten (Ewer 1973; Simms 1979).

The long-tailed weasel is more of a generalist in
food habits and habitat use than the ermine or least
weasel (Ewer 1973; Haley 1975; Gamble 1981 ). It
also has a more southerly distribution than these
smaller species. Simms (1979) hypothesized that the
northward distribution of the long-tailed weasel is
limited by snow cover, while the southward distribu-
tions of the two smaller species are limited by inter-
actions with the long-tailed weasel. In the
southwestern U.S., it is absent from deserts and de-
sert grasslands.

The least weasel feeds almost entirely on small
rodents, primarily Microtus sp. (Ewer 1973; Haley
1975), and prefers meadows, brushy areas, and open
woods. Ermine or short-tailed weasels are found in
habitats similar to those of least weasels, but their
distribution extends farther north and south.

Working in southern Ontario, Simms ( 1979) and
King ( 1983) noted that short-tailed weasels selected
early successional communities and avoided forested
habitats, whereas long-tailed weasels showed no
habitat preferences. Similarly, Larrison and Johnson
(1981 ) reported that the short-tailed weasel in Idaho
shuns the more rugged mountains and higher foothill
ridges and occupies stream bottoms, rocky slides,
fence rows, and shrub growth near water, as well as

POPULATION MEASUREMENT TECHNIQUES

Low densities and nonrandom distributions of
carnivore populations create sampling problems
when estimating population sizes. Carnivores are
often distributed in a nonrandom manner because
they prefer some specific type of habitat or because
of social interactions between individuals or groups.
Sampling problems are often compounded by the
high mobility and wariness of most carnivore spe-
cies. No single or best technique has yet been devel-
oped that can suitably census all carnivore
populations. Basic sampling schemes and census
methods indicate either presence or absence of a
species, relative abundance, or absolute abundance
(reviewed by Lewis 1970; Seber 1973; Sen 1982;
Barrett 1983).

Carnivores

487

How and when a biologist should measure a
carnivore population or prey base depends on the
type of information needed. If information is re-
quired to assess the range or local distribution of the
species, then techniques to determine presence are
satisfactory. Techniques to obtain relative density
estimates should be implemented after strongly de-
clining harvests by trappers and hunters, substantial
animal damage, habitat loss, or public concern for
the species. Such techniques may require modifica-
tions as well as trial and error to achieve the desired
results. They can be applied to most species within
a reasonable time frame. An estimate of absolute
density for most species is virtually impossible. The
techniques are extremely time-consuming and ex-
pensive but may be appropriate when a carnivore is
faced with extinction. These techniques are normally
outside the scope of most management agencies.

The need to assess the prey base is most appro-
priate when carnivores are being censused for abso-
lute density estimates and where knowledge of the
prey is crucial to preserving a species (e.g., the prai-
rie dog and black-footed ferret). Often a relative
density index can be obtained more readily for prey
species than for carnivores. Such estimates are par-
ticularly important when data on carnivores reveal
areas of low and high abundance. Importantly, many
other ecological factors (e.g., denning sites, various
aspects of cover, access to subnivially active prey, or
interactions with other species) may be of equal or
greater importance. Therefore, sampling the prey
base may only be appropriate when other ecological
factors are also being considered. One also presup-
poses a knowledge of the principal prey, and rela-
tively few carnivores depend on one or a few
species of prey for survival.

Presence

Sightings or sign left by carnivores indicate their
presence in a particular area. Yocom ( 1973, 1974)
and Yocom and McCollum ( 1973) used reports of
sightings and sign to indicate the presence and range
of wolverine, marten, and fisher in California. Tracks,
scats, and landing sites of river otter were used by

Melquist and Hornocker (1979) to identify areas
frequented by otters. Harvest records indicate geo-
graphic areas where species had been present. Re-
sponses to man-made stimuli, such as signs left at
scent stations or howling by wolves or coyotes to si-
rens, indicate recent or immediate presence of
carnivores.

Some states provide harvest records of carnivores.

Relative Density

Aerial surveys, bounty and harvest records,
questionnaires, number of road kills, and indexes
developed from counts of sightings, sign, and rate of
capture or catch per unit effort data are widely used
to estimate relative abundance of carnivore species.
These types of data taken over time reveal popula-
tion trends, are less costly than censuses, and are
usually adequate for most management needs. Nellis
and Keith (1976) and Parker (1973) used aerial
surveys to estimate coyote and wolf densities, re-
spectively. However, this technique is not always re-
liable for wolves (Miller and Russell 1977).

Cahalane (1964) used responses from mail ques-
tionnaires, based mostly on bounty and harvest rec-
ords, to estimate population trends and relative
densities of mountain lion, grizzly bears, and wolves
throughout North America. Fuller and Keith (1980)
obtained trends in wolf numbers in northeastern
Alberta from records of trapping and poisoning. Simi-
larly, coyote densities were estimated from fur and
hunter harvest estimates obtained from annual re-
ports and mail and telephone contacts. In Oklahoma,
Hatcher and Shaw ( 1981 ) obtained slightly over a
50% response rate on 3,000 questionnaires sent to
farmers. They compared the results of this mail sur-
vey with those of two scent-station surveys to evalu-
ate bobcat, coyote, and gray and red fox populations.

488

Carnivores

Their results suggested that mail surveys are more
effective than scent-station surveys in estimating
carnivore populations. Furthermore, mail surveys are
more cost-effective. Berg et al. (1983) mailed ques-
tionnaires to trappers and mountain lion hunters,
outfitters, biologists and foresters, and ranchers and
farmers to obtain data on distribution and relative
densities of mountain lions in Wyoming.

Allen and Sargent ( 1975) used sightings of red
fox by rural mail carriers as an index of population
trends in North Dakota. This technique correlated
highly (r = 0.958) with population estimates de-
rived by aerial searches for fox families in the spring.
This index appears best-suited to prairie regions,
where foxes are relatively conspicuous. Its accuracy,
timing, and low cost make this technique especially
useful in management programs.

The relative abundance of coyotes has been
estimated by scat deposition rates (reviewed by An-
delt and Andelt 1984). Scat deposition rates are de-
termined from the number of scats found along a
given length of road during a known period. Before
coyote density indexes and trends in abundance can
be reliably estimated from scat deposition rates, the
effect of diet on deposition rates must be deter-
mined. Numbers of interstate road-kills have also
been used to determine the relative density of wild-
life, including the coyote (Case 1978).

Elicited responses to man-made stimuli are often
used to estimate relative densities of carnivore popu-
lations, especially those of coyotes. Scent-station
indexes have recently gained popularity as a means
of estimating seasonal and annual trends in relative
abundance of carnivores (reviewed by Conner et al.
1983). Richards and Hine (1953) and Wood (1959)
originally developed this method for determining the
relative abundance of red and gray foxes. Humphrey
and Zinn (1982) modified it for river otter and mink.
Lindzey and Knowlton ( 1975) used it extensively
for coyotes in 17 western states. Several hundred
scent-station lines, each consisting of 50 scent sta-
tions located at 0.48-km (0.3-mi. ) intervals along
a 23.5-km (14.7-mi.) route, were checked for five
consecutive days in September. Each station con-
tained a perforated plastic capsule containing a fer-
mented-egg attractant placed in the center of a 1 -
m (1-yd) circle of sifted dirt. Animal visits (based on
tracks ) were recorded to provide an index by which
coyote population trends could be compared be-
tween regions, states, and years. The index of rela-
tive abundance was calculated as follows:

Index = 1,000 x

total animal visits

total operative station nights

The presence of many other carnivore species
was recorded at scent stations, but because survey
lines were located primarily in coyote habitat, these

records are not likely a measure of relative abun-
dance of these species. With modifications, similar
scent stations could be used to estimate relative
densities of these carnivore species. Lindzey and
Meslow ( 1977) modified this technique to make it
more selective for black bears, and Clark and Camp-
bell ( 1983) modified it for small carnivores. Conner
et al. (1983) ran permanent scent-station transects in
Florida and evaluated the indexes to determine sea-
sonal and annual trends in abundance of bobcats,
raccoons, and gray foxes. They found this technique
provided a reliable index of species abundance when
compared to estimates based on trapping, radioiso-
tope tagging, and radiotelemetry. Despite the utility
of this technique, Roughton and Sweeny (1982) and
Conner et al. ( 1983) recognized it needed to be
standardized and verified before between-area com-
parisons could be made. Furthermore, between-spe-
cies comparisons of population densities could not
be made directly with scent-station indexes because
of species-specific visitation rates.

Siren-elicited vocalization has also gained inter-
est as a possible method for estimating relative abun-
dance of coyotes ( reviewed by Okoniewski and
Chambers 1984). Wenger and Cringan (1978) evalu-
ated the technique in northeastern Colorado by us-
ing six radio-instrumented coyotes. Three of the
coyotes readily responded to a siren, but the other
three rarely responded. These results substantiated
the response rate assumed by Alcorn ( 1971 ) for
coyotes in Nevada. However, further research is
needed to determine if an average response rate ex-
ists for all areas. With yearly sampling on a given
area, this technique should provide good indexes of
locally specific, annual coyote fluctuations. Andrews
and Boggess (1978) reported two advantages of
using this technique for coyotes over the more pop-
ular scent-station survey: ( 1 ) coyote and dog vocali-
zations are more easily distinguished than coyote
and dog tracks and (2) the elicited-vocalization tech-
nique is not complicated by regional variation in
coyote use or avoidance of roadways. Okoniewski
and Chambers ( 1 984 ) considered the advantages of
the elicited-vocalization technique to be especially

Carnivores

489

important in areas where dog densities are high and
habitat types diverse. On the other hand, specialized
equipment is needed, and a substantial amount of
field effort and time is required.

Catch per unit effort has also been used to esti-
mate the relative densities of coyotes (Clark 1972;
Knowlton 1972). Clark ( 1972) used the success of a
summer trapping program, expressed as coyotes
captured per 1,000 trap days, as one relative index
of a postbreeding coyote population. Knowlton
(1972) used humane "coyote-getters," set in stand-
ard lines, as a similar technique to quantify relative
coyote abundance. Live-trapping techniques can also
be used to estimate carnivore relative densities.
Simms ( 1979) live-trapped species of weasels to de-
termine their habitat preferences and population
densities. Trapping success was measured as the
number of captures per 100 trap days. A trap day
was 24 hours, from one morning check to the next.

Barrett (1983) used smoked aluminum tracking
stations to sample carnivores in California. The ad-
vantages of the technique are the ease in transport-
ing and the ability to identify tracks. The major
disadvantage is similar to that of other bait-based
techniques ( variation in attractability of many
species).

den, or other sign. A complete count of all individu-
als in a population is a census in the true meaning of
the word. This is very difficult for most carnivore
populations. An exception may be locating wolf
packs by radiotelemetry and then counting entire
packs from aircraft (Fuller 1982). However, counting
packs of wolves may not account for the number of
transients, which in wolves make up 8 to 20% of the
population (Mech and Frenzel 1971).

The Lincoln or Petersen Index method is a cap-
ture-recapture technique commonly used to estimate
population sizes of many species of wildlife, includ-
ing carnivores (Lewis 1970; Seber 1973; Davis and
Winstead 1980). The word "index" is a misnomer
because this technique provides an estimate of total
population. Accuracy of this technique depends on
the proportion of the population sampled. The
Schnabel Method is a variation of the Lincoln Index.
With this method, samples are taken over time so
population estimates can be averaged. With the Lin-
coln or Petersen Index method, M individuals from a
population are caught, marked, and released. On a
second occasion, a sample of n individuals is cap-
tured. If m is the number of marked animals in this
sample, an estimate N2 of the population size N and
an estimate v(N?) of its variance are calculated as
follows:

Absolute Density

Direct counts and capture-recapture techniques
are most commonly used to determine absolute den-
sity of wildlife species. Ilie term "direct count" re-
fers to counting each animal rather than its tracks,

490

Carnivores

$ = n_(M)

2 m

v(N2)

Nr(N2 - Af)(N2 - n)
Mn ( N_> - 1 )

!■

Sign left by carnivores has also been used to
determine absolute density. Brand et al. (1976) be-
lieved the number of lynx and their movements
could be accurately determined on a study area by
track observations. After snowfalls, they drove along
all roads in their study areas each day and recorded
lynx tracks across roads until conditions prevented
identifying recent tracks.

Koford ( 1977) derived estimates of mountain
lion populations for several study areas in California,
based on the number of different sets of tracks de-
tected per 160 km (100 mi.) of road and trail track-
ing. By comparing these results with the total
number of mountain lions detected by intensive
searches of each study area, track counts were cali-
brated to indicate numbers of mountain lions per
260 km2 (100 mi.2). Koford ( 1977) and Shaw
(1979) believed all mountain lions on a study area
could be distinguished by the shape and size of their
footprints. Both authors reported that the width of
track heel pads was the most useful measurement to
be taken.

Sightings of fox dens from the ground ( Scott
and Seiko 1939) and from the air (Sargent et al.
1975) have been used to census red fox populations
in Iowa and North Dakota, respectively. Sargent et
al. (1975) flew systematic aerial searches to locate
red foxes and their dens. The number of individual
fox families was used as a census of the population.

DISCUSSION

Carnivores are a diverse group of mammals.
Their respective habitats and food requirements vary

greatly (Table 1 ). However, to accurately relate habi-
tat to any species or animal community, biologists
must be able to identify environmental features that
accurately reflect a species' or community's require-
ments. Moreover, changes in these environmental
features should allow biologists to make accurate
predictions of a species' response to habitat altera-
tion (Anderson 1982), either directly or indirectly,
through a change in prey base. Usually dominant
soils, plant species, and landforms have been listed
and qualitatively described. Although these descrip-
tions convey much information about an individual
habitat type, they do not allow correlations between
specific environmental features and a species' re-
quirements. Quantitatively measuring many habitat
features allows statistical correlation between spe-
cific habitat components and presence and abun-
dance of a wildlife species.

Habitat requirements or preferences of carni-
vore species can be quantified with current technol-
ogy. Soil, vegetation, and landforms can be
quantitatively evaluated on the ground by techniques
of Nudds ( 1977) and Hays et al. ( 1981 ) and those
reviewed by Anderson (1982). Large-scale aerial
photography (Aldrich 1979; Ulliman et al. 1979) and
satellite imagery (Varney 1973; Brabander and Bar-
clay 1977; Aldrich 1979; Cannon et al. 1982; Craig-
head et al. 1982) can be used to classify, inventory,
and monitor general vegetative cover types (e.g.,
alpine meadows, forest, grassland, shrubland). Devel-
opment of these remote-sensing techniques to evalu-
ate vegetation types and land uses has greatly
reduced the logistical effort and the cost of survey-
ing and monitoring extensive areas by conventional
methods.

Carnivores

491

LITERATURE CITED

ABLES, E.D. 1975. Ecology of the red fox in America.

Pages 216-236 in Fox, M.W. ed. The Wild Canids. Van
Nostrand Reinhold Co. New York, NY.

ACKERMAN, B.B. 1982. Cougar predation and ecological
energetics in southern Utah. M.S. Thesis. Utah State
Univ., Logan. 103pp.

, F.G. LINDZEY, and T.P. HEMKER. 1984. Cougar

food habits in southern Utah. J. Wildl. Manage.
48:147-155.

ALCORN, JR. 1971. A discussion of coyote censusing

techniques. U.S. Bur. Sport Fish, and Wildl., Div. Wildl.
Serv. 8pp.

ALDRICH, R.C. 1979. Remote sensing of wildland re-
sources: A state-of-the-art review. U.S. Dep. Agric, For.
Serv. Gen Tech. Rep. RM-71. 56pp.

ALLEN, AW. 1982. Habitat suitability index models: Mar-
ten. U.S. Dep. Inter., Fish and Wildl. Serv. FWS/OBS
82/10.11. 9pp.

. 1983. Habitat suitability index models: Fisher. U.S.

Dep. Inter., Fish and Wildl. Serv. FWS/OBS-82/0.45.
20pp.

ALLEN, S.H. and A.B. SARGENT. 1975. A rural mail-carrier
index of North Dakota red foxes. Wildl. Soc. Bull.
3:74-77.

AMSTRUP, S.C. and J. BEECHAM. 1976. Activity patterns
of radio-collared black bears in Idaho. J. Wildl. Man-
age. 40:340-348.

ANDELT, W.F. and S.H. ANDELT. 1981. Habitat use by
coyote in southeastern Nebraska. J. Wildl. Manage.
45(4): 1001-1005.

and . 1984. Diet bias in scat deposition-rate

surveys of coyote density. Wildl. Soc. Bull. 12:74-77.

ANDERSON, A.E. 1983- A critical review of literature
on puma (Felis concolor). Colo. Div. Wildl. Spec.
Rep. 54. 91pp.

ANDERSON, S.H. 1982. Comments on measurement of
habitat. Pages 278-280 in Davis, D.E. ed. CRC Hand-
book of Census Methods for Terrestrial Vertebrates.
CRC Press, Inc. Boca Raton, FL.

ANDREWS, R.D. and E.K. BOGGESS. 1978. Ecology of

coyotes in Iowa. Pages 249-265 in Bekoff, M. ed. Coy-
otes Biology, Behavior and Management. Academic
Press. New York, NY.

BAILEY, T.N. 1974. Social organization in a bobcat popula-
tion. J. Wildl. Manage. 38:435-445.

BANFIELD, A.W.F. 1974. The Mammals of Canada. Univ.
Toronto Press, Toronto. 438pp.

BARNES, C.T. I960. The cougar or mountain lion. The
Ralton Co. Salt Lake City, UT. 176pp.

BARRETT, R.H. 1983. Smoked aluminum track plots for
determining furbearer distribution and relative abun-
dance. California Fish and Game 69:188-190.

BEASON, S.L. and R.A. MOORE. 1977. Bobcat food habit
response to a change in prey abundance. Southwest.
Nat. 21:451-457.

BERG, R.L., L.L. MCDONALD, and M.D. STRICKLAND.

1983 Distribution of mountain lions in Wyoming as
determined by mail questionnaire. Wildl. Soc. Bull.
11:265-268.

BERG, W.K and R.A. CHESNESS. 1978. Ecology of coyotes
in northern Minnesota. Pages 229-247 in Bekoff, M.
ed. Coyotes Biology, Behavior and Management. Aca-
demic Press. New York, NY.

BERNER, A. and L.W. GYSEL. 1967. Raccoon use of large
tree cavities and ground burrows. J. Wildl. Manage.
31:706-714.

BRABANDER, J.J. and J.S. BARCLAY. 1977. A practical
application of satellite imagery to wildlife habitat
evaluation. Proc. Ann. Conf. Southeast Assoc. Fish
Wildl. Agencies 31:300-306.

BRADLEY, W.G. and C.G. HANSEN. 1965. Observations on
the distribution of the ring-tailed cat in southern
Nevada. Southwest. Nat. 10:310-311.

BRAND, C.J. and LB. KEITH. 1979- Lynx demography
during a snowshoe hare decline in Alberta. J. Wildl.
Manage. 40:416-428.

, LB. KEITH, and C.A. FISCHER. 1976. Lynx re-
sponse to changing snowshoe hare densities in cen-
tral Alberta. J. Wildl. Manage. 40:416-428.

BRAY, O.E. and V.G. BARNES. 1967. A literature review on
black bear populations and activities. U.S. Dep. Inter.,
Natl. Park Serv., and Colo. Coop. Wildl. Res. Unit.
34pp.

BROWN, D.E. 1983. The wolf in the Southwest. Univ.
Arizona Press, Tucson. 195pp.

BUCK, S., C. MULLIS, and A. MOSSMAN. 1978. A radio
telemetry study of fishers in northwestern California.
Cal.-Nev. Wildl. Trans. 1979:166-172.

BURNS, J.J. 1964. Comparisons of two populations of mink
from Alaska. Can. J. Zool. 42:1071-1079.

BUSKIRK, S.W. 1984. Seasonal use of resting sites by
marten in south-central Alaska. J. Wildl. Manage.
48:950-953.

CAHALANE, V.H. 1964. A preliminary study of distribution
and numbers of cougar, grizzly and wolf in North
America. NY. Zool. Soc. 12pp.

CANNON, R.W., FL. KNOPF, and L.R PETTINGER. 1982.
Use of landsat data to evaluate lesser prairie chicken
habitats in western Oklahoma. J. Wildl. Manage.
46:915-922.

CARBYN, L.N. 1982. Coyote population fluctuations and
spatial distribution in relation to wolf territories in
Riding Mountain National Park. Can. Field-Nat.
96:176-183-

CASE, R.M. 1978. Interstate highway road-killed animals, a
data source for biologists. Wildl. Soc. Bull. 6:8-13-

CLARK, F.W. 1972. Influence of jackrabbit density on
coyote population change. J. Wildl. Manage. 36:343-
356.

CLARK, T.W. and T.M. CAMPBELL 1983. A small carnivore
survey technique. Great Basin Nat. 43:438-440.

COLE, G.F. 1972. Grizzly bear-elk relationships in Yellow-
stone National Park. J. Wildl. Manage. 36:556-561.

CONNER, M.C., R.F. LABISKY, and DR. PROGULSKE, Jr.
1983- Scent-station indices as measures of population
abundance for bobcats, raccoons, gray foxes, and
opossums. Wildl. Soc. Bull. 11:146-152.

COWAN, I. McT. 1947. The timber wolf in the Rocky
Mountain National Parks of Canada. Can. J. Res.
25:139-174.

and R.H. MACKAY. 1959. Food habits of marten

(Mattes americana) in the Rocky Mountain region of
Canada. Can. Field-Nat. 64:100-104.

CRABB, W.G. 1948. The ecology and management of the
prairie spotted skunk in Iowa. Ecol. Monogr. 18:201-
233.

CRAIGHEAD, J.J, J.S. SUMMER, and G.B. SCAGGS. 1982. A
definitive system for analysis of grizzly bear habitat
and other wilderness resources utilizing IANDSAT
multispectral imagery and computer technology.

492

Carnivores

Wildlife-Wildlands Inst. Monogr. 1, Univ. Montana

Foundation, Missoula.
CUMBIE, P.M. 1975. Mercury levels in Georgia otter, mink

and freshwater fish. Environ. Contam. Toxicol.

14:193-196.
DAVIS, D.E. and R.L. WINSTEAD. 1980. Estimating the

number of wildlife populations. Pages 221-246 in

Giles, R.H. ed. Wildlife Management Techniques Man-
ual. The Wildl. Soc. Washington, DC.
DE VOS, A. 1952. Ecology and management of fisher and

marten in Ontario. Tech. Bull. Ontario Dep. Lands and

Forests.
DOUGLASS, R.J., L.G. FISHER, and M. MAIR. 1983. Habitat

selection and food habits of marten, Martes atneri-

cana, in the Northwest Territories. Can. Field-Nat.

97:71-74.
EGOSCUE, H.J. 1962. Ecology and life history of the kit

fox in Tooele County, Utah. Ecology 43:481-497.
ELLIS, R.J. 1964. Tracking raccoons by radio. J. Wildl.

Manage. 28:363-368.
ERRINGTON, P.L. 1935. Food habits of mid-west foxes. J.

Mammal. 16:192-200.
EWER, R.F. 1973. The Carnivores. Cornell Univ. Press.

Ithaca, NY. 494pp.
FLOYD, B.L. and MR. STROMBERG. 1981. New records of

the swift fox (Vulpes velox) in Wyoming. J. Mammal.

62:650-651.
FRITZELL, E.K 1978. Habitat use by prairie raccoons

during the waterfowl breeding season. J. Wildl. Man-
age. 42:118-127.
FULLER, T.K 1982. Wolves. Pages 225-226 in Davis, D.E.

ed. CRC Handbook of Census Methods for Terrestrial

Vertebrates. CRC Press, Inc. Boca Raton, FL.
and LB. KEITH. 1980. Wolf population dynamics

and prey relationships in northeastern Alberta. J.

Wildl. Manage. 44:583-602.
GAMBLE, R.L. 1981. Distribution in Manitoba of Mustela

Frenata Longicanda, the long-tailed weasel, and the

interrelation of distribution and habitat selection

in Manitoba, Saskatchewan and Alberta. Can. J. Zool.

59:1036-1039.
GASAWAY, W.C., R.O. STEPHENSON, J.L. DAVIS, P.E.K.

SHEPHERD, and O.E. BURRIS. 1983. Interrelationships

of wolves, prey, and man in interior Alaska. Wildl.

Monogr. 84. 50pp.
GASHWILER, J., W.L. ROBINETTE, and O.W. MORRIS.

1961. Breeding habits of bobcats in Utah. J. Mammal.

42:76-84.
GIER, H.T. 1975. Ecology and behavior of the coyote

(Canis latrans). Pages 246-262 in Fox, M.W. ed. The

Wild Canids. Van Nostrand Reinhold Co. New York,

NY.
GILBERT, F.F. and E.G. NANCEKJVELL. 1982. Food habits

of mink (Mustelavison) and otter (Lutra canadensis)

in northeastern Alberta Canada. Can. J. Zool. 60:1282-

1288.
GRABER, DM. 1982. Ecology and management of black

bear in Yosemite National Park. Cooperative National

Park Resources Study Unit. Univ. of California at

Davis. Tech. Rep. 5.
GREEN, J.S. and J.T. FLINDERS. 1981. Diets of sympatric

red foxes and coyotes in southeastern Idaho. Great

Basin Nat. 41:251-254.
GREENWOOD, R.J. 1982. Nocturnal activity and foraging

of prairie raccoons in North Dakota. Am. Midi. Nat.

107:238-243-
GREER, KR. 1955. Yearly food habits of the river otter in

the Thompson I.akes region, northwestern Montana,
as indicated by scat analysis. Am. Midi. Nat. 54:299-
313.

GRENFELL, W.E. and A.J. BRODY. 1983. Seasonal foods of
black bears in Tahoe National Forest, California. Calif.
Fish and Game 69:132-150.

and M. FASENFEST. 1979. Winter food habits of

fishers, Martes pennanti, in northwestern California.
Calif. Fish Game 65:186-189.

GRINNELL, J, J.S. DIXON, andJ.M. LINSDALE. 1937. Fur-
bearing mammals of California, Vol. I. Univ. Calif.
Press, Berkeley. 375pp.

HALEY, D. 1975. Sleek and savage. North America's weasel
family. Pacific Search. Seattle, WA. 298pp.

HALL, E.R. 1951. Weasels. Univ. Kansas Publ. Mus. Nat.
Hist. No. 4.

and KR. KELSON. 1959. The mammals of North

America, Vol. II. The Ronald Press. New York, NY.

HARGIS, CD. and DR. McCULLOUGH. 1984. Winter diet
and habitat selection of marten in Yosemite National
Park. J. Wildl. Manage. 48:140-146.

HARLOW, H.J. 1981. Torpor and other physiological adap-
tations of the badger (Taxidea taxus) to cold envi-
ronments. Physiol. Zool. 54:267-276.

HART, E.B. 1982. The raccoon (Procyon lotor) in Wyo-
ming. Great Basin Nat. 42:599-600.

HATCHER, R.T. and J.H. SHAW. 1981. A comparison of
three indices to furbearer populations. Wildl. Soc.
Bull. 9:153-156.

HAWTHORNE, V.M. 1972. Coyote food habits in Sagehen
Creek basin, northeastern California. Calif. Fish and
Game 58:4-12.

HAYS, R.L, C. SUMMERS, and W. SEITZ. 1981. Estimating
wildlife habitat variables. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-81/47. Washington, DC. 111pp.

HEMKER, T.P. 1982. Population characteristics and move-
ment patterns of cougars in southern Utah. M.S. The-
sis. Utah State Univ., Logan. 59pp.

HENRY, C.J., L.J. BLUS, S.V. GREGORY, and C.J. STAF-
FORD. 1981. PCBs and organochlorine pesticides in
wild mink and river otters from Oregon. Pages 1 763-
1789 in Chapman, J.A. and Pursley, D. eds. World-
wide Furbearer Conf. Proc. 1980. Frostburg, MD.

HENSHAW, R.E. 1975. Reintroduction of wolves into the
wild. Pages 420-444 in Klinghammer, E. ed. The
Behavior and Ecology of Wolves. Symp. Proc. Garland
STPM Press. New York, NY.

, R. LOCKWOOD, R. SHIDELER, and R.O. STEPHEN-
SON. 1975. Experimental release of captive wolves.
Pages 319-345 in Klinghammer, E. ed. The Behavior
and Ecology of Wolves. Symp. Proc. Garland STPM
Press, New York, NY.

and R.O. STEPHENSON. 1974. Homing in the gray

wolf (Canis lupus).). Mammal. 55:234-237.

HERRERO, S. 1972. Aspects of evolution and adaptation in
American black bears (Ursns americanus Pallus) and
brown and grizzly bears ( U. asctos Linne ) of North
America. Pages 221-231 in Herrero, S. ed. Bears —
Their Biology and Management. New Series 23, Inter-
national Union for Conservation of Nature. Morges,
Switzerland.

. 1978. A comparison of some features of the evolu-
tion, ecology and behavior of black and grizzly/brown
bears. Carnivore 1:7-17.

HILTON, H. 1978. Systematics and ecology of the eastern
coyote. Pages 209-228 in Bekoff, M. ed. Coyotes
Biology, Behavior and Management. Academic Press.

Carnivores

493

New York, NY.

HOFFMAN, CO. and J.L. GOTTSCHANG. 1977. Numbers,
distribution, and movements of a raccoon population
in a suburban residential community. J. Mammal.
58:623-636.

HOLLEMANN, D.F. and R.O. STEPHENSON. 1981. Prey
selection and consumption by Alaskan wolves in
winter. J. Wildl. Manage. 45:620-628.

HORNOCKER, M.G. 1969. Winter territoriality in moun-
tain lions. J. Wildl. Manage. 33:457-464.

. 1970. An analysis of mountain lion predation upon

mule deer and elk in the Idaho Primitive Area. Wildl.
Monogr. 21. 39pp.

and H.S. HASH. 1981. Ecology of the wolverine in

northwestern Montana. Can. J. Zool. 59:1286-1301.

HOUSTON, D.B. 1978. Elk as winter-spring food for carni-
vores in northern Yellowstone National Park. J. Appl.
Ecol. 15:653-662.

HUMPHREY, S.R. and T.L. ZINN. 1982. Seasonal habitat
use by river otters and Everglades mink in Florida. J.
Wildl. Manage. 46:375-381.

INGLES, L.G. 1965. Mammals of the Pacific States. Stanford
Univ. Press, Stanford, CA. 506pp.

JOHNSON, M.K and R.M. HANSEN. 1979. Coyote food

habits on the Idaho Natural Engineering Laboratory. J.
Wildl. Manage. 43:951-955.

JONES, J.H. and N.S. SMITH. 1979. Bobcat density and
prey selection in central Arizona. J. Wildl. Manage.
43:666-672.

JONKEL, J. and I. McT. COWAN. 1971. The black bear in
the spruce-fir forest. Wildl. Monogr. 27.

JORDAN, PA., PC. SHELTON, and D.L. ALLEN. 1967.
Numbers, turnover, and social structure of the Isle
Royale wolf population. Am. Zool. 7:233-252.

KEITH, L.B. 1963. Wildlife's ten-year cycle. Univ. Wiscon-
sin Press, Madison. 201pp.

KELLEYHOUSE, D.G. 1980. Habitat utilization by black
bears in northern California. Pages 221-228 in Mar-
tinka, C.J. and KL. McArthur eds. Bears — Their Biol-
ogy and Management. Bear Biol. Assoc. Ser. 4.

KILGORE, D.L., Jr. 1969. An ecological study of the swift
fox (Vulpes velox) in the Oklahoma Panhandle. Am.
Midi. Nat. 81:512-534.

KING, CM. 1983- Mustela erminea. Mammalian Species
195. 8pp.

KNOWLTON, F.F. 1972. Preliminary interpretation of

coyote population mechanics with some management
implications. J. Wildl. Manage. 36:369-382.

KOEHLER, G.M. and M.G. HORNOCKER. 1977. Fire ef
fects on marten habitat in the Selway-Bitterroot Wil-
derness. J. Wildl. Manage. 41:500-505.

, , and H.S. HASH. 1979. Lynx movements

and habitat use in Montana. Can. Field-Nat. 93:441-
442.

KOFORD, C.B. 1977. Status and welfare of the puma (Felis
concolor) in California, 1973-1976. Final Rep. De-
fenders of Wildlife and the National Audubon Society.
Mus. Vert. Zool. Univ. Calif., Berkeley. 57pp.

KOLENOSKY, G.B. 1972. Wolf predation on wintering

deer in east-central Ontario. J. Wildl. Manage. 36:257-
369.

KORSCHGEN, L.J. 1959. Food habits of the red fox in
Missouri. J. Wildl. Manage. 23:168-176.

KUYT, E. 1972. Food habits and ecology of wolves on

barren-ground caribou range in the Northwest Terri-
tories. Can. Wildl. Serv. Rep. Ser. 21. 36pp.

LAMPE, R.P. 1982. Food habits of badgers in east-central

Minnesota. J. Wildl. Manage. 46:790-795.
— and M. SOVADA. 1981. Seasonal variation in home

range of a female badger (Taxidea taxiis). Prairie Nat.
13:55-58.

LARRISON, E.J. and D.R.JOHNSON. 1981. Mammals of
Idaho. Univ. Idaho Press, Moscow. 166pp.

LECOUNT, A.L. 1980. Some aspects of black bear ecology
in the Arizona chaparral. Pages 175-180 in Martinka,
C.J. and KL. McArthur eds. Bears — Their Biology
and Management. Bear Biol. Assoc. Ser. 4.

LEWIS, J.C 1970. Wildlife census methods: A resume. J.
Wildl. Dis. 6:356-364.

LINDZEY, F.G. 1978. Movement patterns of badgers in
northwestern Utah. J. Wildl. Manage. 42:418-422.

and F.F. KNOWLTON. 1975. Determining the rela-
tive abundance of coyotes by scent station lines.
Wildl. Soc. Bull. 3:119-124.

and E.C MESLOW. 1977. Home range and habitat

use by black bears in southwestern Washington. J.
Wildl. Manage. 41:408-415.

LINHART, SB. and W.B. ROBINSON. 1972. Some relative
carnivore densities in areas under sustained coyote
control. J. Mammal. 53:880-884.

LONG, CA. 1973- Taxidea taxus. Mammalian Species 26.
The Amer. Soc. of Mammal. Lawrence, KS. 4pp.

and CA. KILLINGLEY. 1983. The badgers of the

world. Charles C. Thomas, Publisher. Springfield, IL.
404pp.

LOTZE, J.H. and S. ANDERSON. 1979. Procyon lotor.

Mammalian Species 119. The Amer. Soc. of Mammal.
Lawrence, KS. 8pp.

LYNCH, G.M. 1974. Some den sites of Manitoba raccoons.
Can. Field-Nat. 88:494-495.

MARTINKA, C.J. 1976. Ecological role and management of
grizzly bears in Glacier National Park, Montana. Pages
147-156 in Pelton, MR, J.W. Lentfer, and G.E. Folk
eds. Bears — Their Biology and Management. IUCN
Publ. New Ser. 40. Morges, Switzerland.

McCORD, CM. and J.E. CARDOZA. 1982. Bobcat and lynx.
Pages 728-766 in Chapman, J.A. and G.A. Feldhamer
eds. Wild Mammals of North America: Biology, Man-
agement and Economics. John Hopkins Univ. Press.
Baltimore, MD.

MEALY, S.P. 1980. The natural food habits of grizzly bears
in Yellowstone National Park, 1973-74. Pages 281-292
in Martinka, C.J. and KL. McArthur eds. Bears — Their
Biology and Management. Bear Biol. Assoc. Ser. 4.

MECH, L.D. 1966. The wolves of Isle Royale. U.S. Dep. In-
ter., Natl. Park Serv. Fauna Ser. 7. Washington, DC.
210pp.

. 1975. Some considerations in re-establishing

wolves in the wild. Pages 445-457 in Klinghammer, E.
ed. The Behavior and Ecology of Wolves. Symp. Proc.
Garland STPM Press. New York, NY.

. 1977. Productivity, mortality, and population

trends of wolves in northeastern Minnesota. J. Mam-
mal. 58:559-574.

and L.D. FRENZEL. 1971. Ecological studies of the

timber wolf in northeastern Minnesota. U.S. Dep.
Agric, For. Serv. Res. Pap. NC-52. 62pp.

and P.D. KARNS. 1977. Role of the wolf in deer

decline in the Superior National Forest. U.S. Dep.
Agric, For. Serv. Res. Pap. NC-51. 62pp.

MELQUIST, WE. and M.G. HORNOCKER. 1979. Methods
and techniques for studying and censusing river otter
populations. Univ. Idaho For. Wildl. and Range Exp.
Sta. Tech. Rep. 8. 1 7pp.

494

Carnivores

and

-. 1983- Ecology of river otters in west-

central Idaho. Wildl. Monogr. 83- 60pp.
— J.S. WHITMAN, and M.G. HORNOCKER. 1981.

Resource partitioning and coexistence of sympatric
mink and river otter populations. Pages 187-220
in Chapman, J.A. and D. Pursley eds. Worldwide Fur-
bearer Conf., Vol. 1. Frostburg, MD.

MESSICK, J.P. and M.G. HORNOCKER. 1981. Ecology of
the badger. Taxidea taxns in southeastern Idaho.
Wildl. Monogr. 76. 53pp.

MILLER, F.L. and R.H. RUSSELL. 1977. Unreliability of strip
aerial surveys for estimating numbers of wolves on
western Queen Elizabeth Islands, Northwest Territo-
ries. Can. Field-Nat. 91:77-82.

MOORE, RE. and N.S. MARTIN. 1980. A recent record of
the swift fox (Vulpes velox) in Montana. J. Mammal.
61:161.

MORE, G. 1 976. Some winter food habits of lynx (Felix
lynx) in the southern Mackenzie district, Northwest
Territories. Can. Field-Nat. 90:499-500.

MURIE, A. 1944. The wolves of Mount McKinley. U.S. Dep.
Inter., Natl. Park Serv. Fauna Ser. 5. 238pp.

. 1951. Coyote food habits on a southwestern cattle

range. J. Mammal. 32:291-295.

NELLIS, C.H. and LB. KEITH. 1968. Hunting activities and
success of lynx in Alberta. J. Wildl. Manage. 32:718-
722.

and . 1976. Population dynamics of coyotes

in central Alberta, 1964-68. J. Wildl. Manage. 40:389-
399.

, S.P. WETMORE, and LB. KEITH. 1972. Lynx-prey

interactions in central Alberta. J. Wildl. Manage.
36:320-329.

NOVICK, H.J., J.M. SIPEREK and G.R. STEWART. 1981.
Denning characteristics of black bears (Ursus ameri-
canus) in the San Bernardino mountains of southern
California. Calif. Fish and Game 67:52-61.

and G.R. STEWART. 1982. Home range and habitat

preferences of black bears (Ursiis americantis) in
the San Bernardino Mountains of southern California,
U.S.A. Calif. Fish and Game 68:21-35.

NUDDS, ID. 1977. Quantifying the vegetative structure of
wildlife cover. Wildl. Soc. Bull. 5:113-117.

OKONIEWSKI, J.C. and RE. CHAMBERS. 1984. Coyote
vocal response to an electronic siren and human
howling. J. Wildl. Manage. 48:217-221.

PARKER, G.R. 1973. Distribution and densities of wolves
within barren-ground caribou range in northern main-
land Canada. J. Mammal. 54:341-346.

, J.W. MAXWELL, L.D. MORTON, and E.J. SMITH.

1983- The ecology of the lynx (Lynx canadensis) on
Cape Breton Island. Can. J. Zool. 61:770-786.

PEDERSON, J.C. and R.C. TUCKFIELD. 1983. A compara-
tive study of coyote food habits on two Utah deer
herds. Great Basin Nat. 43:432-437.

PETERSON, R.O. 1975. The wolves of Isle Royale: New
developments. Pages 3-18 in Klinghammer, E. ed. The
Behavior and Ecology of Wolves. Symp. Proc. Garland
STPM Press. New York, NY.

1977. Wolf ecology and prey relationships on Isle

Royale. U.S. Dep. Inter., Natl. Park Serv. Sci. Monogr.
Ser. 11. 210pp.

PIEKIELEK, W. and T. BURTON. 1975. A black bear pop-
ulation study in northern California. Calif. Fish and
Game 61:4-25.

PIMLOTT, D.H. 1967. Wolf predation and ungulate popu-
lations. Am. Zool. 7:267-278.

POWELL, R.A. 1982. The fisher, life history, ecology, and
behavior. Univ. Minnesota Press, Minneapolis. 217pp.

PULLIAINEN, E. 1975. Ecology of the wolf in the settled
areas of Finland. Pages 84-92 in Klinghammer, E.
ed. The Behavior and Ecology of Wolves. Symp. Proc.
Garland STPM Press. New York, NY.

RAINE, R.M. 1982. Ranges of juvenile fisher (Martes pen -
nanti) and marten, (Martes americana) in southeast-
ern Manitoba. Can. Field-Nat. 96:431-438.

RAUSCH, R.A. and A.M. PEARSON. 1972. Notes on the

wolverine in Alaska and the Yukon Territory. J. Wildl.
Manage. 36:249-268.

RICHARDS, S.H. and R.L. HINE. 1953- Wisconsin fox pop-
ulations. Tech. Wildl. Bull. 6. Wisconsin Conserv.
Dep., Madison.

ROBINETTE, W.L., J.S. GASHWILER, and O.W. MORRIS.
1959. Food habits of the cougar in Utah and Nevada.
J. Wildl. Manage. 23:261-273.

ROUGHTON, R.D. and M.W. SWEENY. 1982. Refinements
in scent-station methodology for assessing trends in
carnivore populations. J. Wildl. Manage. 46:217-229-

RUTHERFORD, W.H. 1954. A record of the ringtail (Bas-
saricas astutns) in Colorado. J. Mammal. 35:442-443

SARGENT, A.B., W.K PFEIFER, and S.H. ALLEN. 1975. A
spring aerial census of red foxes in North Dakota.
J. Wildl. Manage. 39:30-39-

and D.W. WARNER. 1972. Movements and denning

habits of a badger. J. Mammal. 53:207-210.

SAUNDERS, J.K, Jr. 1963- Food habits of the lynx in New-
foundland. J. Wildl. Manage. 27:384-390.

SCHEMPF, P.F. and M. WHITE. 1977. Status of six furbear-
ers in the mountains of northern California. U.S. Dep.
Agric, For. Serv. Calif. Reg. 51pp.

SCHOFIELD, R.D. I960. A thousand miles of fox trails in
Michigan's ruffed grouse range. J. Wildl. Manage.
24:432-434.

SCHOWALTER, D.B. and JR. GUNSON. 1982. Parameters
of populations and seasonal activity of striped skunks.
Mephitis mephitis, in Alberta and Saskatchewan. Can.
Field-Nat. 96:409-420.

SCOTT, T.G. 1955. Dietary patterns of red and gray foxes.
Ecology 36:366-367.

and L.F. SELKO. 1939. A census of red foxes and

striped skunks in Clay and Boone counties, Iowa.
J. Wildl. Manage. 3:92-98.

SEBER, G.A.F. 1973- The estimation of animal abundance.
Griffin. London. 506pp.

SEIDENSTICKER, J.C, M.G. HORNOCKER, W.V. WILES,
and J.P. MESSICK 1973 Mountain lion social organi-
zation in the Idaho Primitive Area. Wildl. Monogr. 35.
60pp.

SEN, A.R. 1982. A review of some important techniques in
sampling wildlife. Can. Wildl. Serv., Occ. Paper 49-
17pp.

SHAW, H.G. 1979. A mountain lion field guide. Arizona
Game and Fish Dep. Spec. Rep. 9, Phoenix. 27pp.

SHORT, ILL. 1979. Food habits of coyotes in a semidesert
grass-shrub habitat. U.S. Dep. Agric, For. Serv. Res.
Note RM-364. 4pp.

SIMMS, DA. 1979. North American weasels resource utili-
zation and distribution. Can. J. Zool. 57:504-520.

SINGER, F.J. 1978. Seasonal concentrations of grizzly
bears. North Fork of the Flathead River, Montana.
Can. Field-Nat. 92:283-286.

SMITH, J.W. and B.J. VERTS. 1982. Mephitis mephitis.

Mammalian Species 173- The Amer. Soc. of Mammal.
Lawrence, KS. 7pp.

Carnivores

495

SOUTIERE, EC. 1979. Effects of timber harvesting on
marten in Maine. J. Wildl. Manage. 43:850-860.

SPALDING, DJ. and J. LESOWSKJ. 1971. Winter food of
the cougar in south-central British Columbia. J. Wildl.
Manage. 35:378-381.

SPENCER, W.D., R.H. BARRETT, and W.J. ZIELINSKI. 1983.
Marten habitat preferences in the northern Sierra
Nevada. J. Wildl. Manage. 47:1181-1186.

SPERRY, C.C. 1941. Food habits of the coyote. U.S. Dep.
Inter., Fish and Wildl. Serv. Res. Bull. 4. 70pp.

SPRINGER, J. T. and J.S. SMITH. 1981. Summer food habits
of coyotes in central Wyoming. Great Basin Nat.
41:449-456.

STAINS, HJ. 1975. Distribution and taxonomy of the cani-
dae. Pages 3-26 in Vox, M.W. ed. The Wild Canids.
Van Nostrand Reinhold Co. New York, NY.

STORM, G.L. 1972. Daytime retreats and movements of
skunks on farmlands of Illinois. J. Wildl. Manage.
36:31-45.

STREETER, R.G. and C.E. BRAUN. 1968. Occurrence of
pine marten (Martes americana) in Colorado alpine
areas. Southwest. Nat. 13:449-451.

STRICKLAND, M.A., C.W. DOUGLAS, M. NOVAK, and N.P.
HUNZIGER. 1982. Martin. Pages 599-612 in Chap-
man, J. and G. Feldhamer eds. Wild Mammals of
North America: Their Biology, Management and Eco-
nomics. The John Hopkins Univ. Press. Baltimore, MD.

TODD, AW. and LB. KEITH. 1983. Coyote demography
during a snowshoe hare decline in Alberta. J. Wildl.
Manage. 47:394-404.

TOWEILL, D.E. and EC. MESLOW. 1977. Food habits of
cougars in Oregon. J. Wildl. Manage. 41:576-578.

TRAPP, G. and D.L HALLBERG. 1975. Ecology of the gray
fox (Urocyoti cinereoargenteus): a review. Pages
164-178 in Fox, M.W. ed. The Wild Canids: Their Sys-
tematics, Behavioral Ecology and Evolution. Van Nos-
trand Reinhold Co. New York, NY.

TREADWELL, B.D. 1979. A provisional framework for
denning black bear habitat. Pages 319-330 in
LeCount, A. ed. First Western Black Bear Workshop.
Ariz. Game and Fish Dep., Phoenix.

TURKOWSKI, F.J. 1980. Carnivora food habits and habitat
use in ponderosa pine forests. U.S. Dep. Agric, For.
Serv. Res. Pap. RM-215. 9pp.

ULLIMAN, J.J., E.O GARTON, andJ.A. KEAY. 1979. Wild-
life habitat classification using large-scale aerial pho-
tography. Pages 414-422 in Frayer, W.F. ed. Forest
Resource Inventories, Vol. 1. Colo. State Univ., Ft.
Collins.

URBAN, D. 1970. Raccoon populations, movement pat-
terns, and predation of a managed waterfowl marsh. J.
Wildl. Manage. 34:372-382.

VAN BALLENBERGHE, V., AW. ER1CKSON, and D. BY-
MAN. 1975. Ecology of the timber wolf in northeast-
ern Minnesota. Wildl. Monogr. 43- 43pp.

VAN ZYLL DE JONG, C.G. 1966. Food habits of the lynx
in Alberta and the Mackenzie District, N.W.T. Can.
Field-Nat. 80:18-23.

. 1975. The distribution and abundance of the wol-
verine (Gulo gulo) in Canada. Can. Field-Nat. 89:431-
437.

VARNEY, JR. 1973. An evaluation of the use of ERTS-1
satellite imagery for grizzly bear habitat analysis.
NASA Prog. Rep., Mont. Coop. Wildl. Res. Unit. Univ.
Montana, Missoula.

VERTS, B.J. 1967. The biology of the striped skunk. Univ.
Illinois Press, Urbana. 218pp.

WAGNER, FH. and LC. STODDART. 1972. Influence of
coyote predation on black-tailed jackrabbit popula-
tions in Utah. J. Wildl. Manage. 36:329-342.

WECKWERTH, R.P. and V.D. HAWLEY. 1962. Marten food
habits and population fluctuations in Montana. J.
Wildl. Manage. 27(1):5574.

WEISE, T.F., W.L ROBINSON, R.A. HOOK and L.D. MECH.
1975. An experimental translocation of the eastern
timber wolf. Pages 19-42 in Klinghammer, E. ed. The
Behavior and Ecology of Wolves. Symp. Proc. Garland
STPM Press. New York, NY.

WENGER, C.R. and AT CRINGAN. 1978. Siren-elicited
coyote vocalizations: an evaluation of a census tech-
nique. Wildl. Soc. Bull. 6:73-76.

WOOD, J.E. 1954. Food habits of furbearers of the upland
post oak region in Texas. J. Mammal. 35:406-415.

. 1959. Relative estimates of fox population levels. J.

Wildl. Manage. 23:53-63.

YOCOM, C.F. 1973. Wolverine records in the Pacific

coastal states and new records for northern California.
Calif. Fish and Game 59:207-209.

. 1974. Status of marten in northern California,

Oregon, and Washington. Calif. Fish and Game 60:54-
57.

and M.T. McCOLLUM. 1973. Status of the fisher in

northern California, Oregon, and Washington. Calif.
Fish and Game 59:305-309.

YOUNG, S.P. 1958. The bobcat of North America. The
Stackpole Co. Harrisburg, PA. and The Wildlife Man-
agement Inst. Washington, DC.

ZIMEN, E. and L. BOITANI. 1975. Status of the wolf in

Europe and the possibilities of conservation and rein-
troduction. Pages 43-83 in Klinghammer, E. ed. The
Behavior and Ecology of Wolves. Symp. Proc. Garland
STPM Press. New York, NY.

496

Carnivores

24

BATS

Stephen P. Cross

Department of Biology
Southern Oregon State College
Ashland, OR 97520

Editor's Note: Most biologists have little experience
with bats, and therefore tend to exclude them from
inventories or monitoring studies. Bats, however,
are an important resource and are often locally
abundant.

The study of bats is a specialized discipline that re-
quires the use of distinctive techniques. This chapter
provides a summary of bat study methods and sug-
gests appropriate circumstances for their use.

INTRODUCTION

Bats are often a neglected segment of wildlife
inventories; they are difficult to sample, they are mo-
bile and have secretive nocturnal habits, and they
are considered by many people to be a mysterious
and unlikable group of wildlife. However, bats play
important roles in many ecosystems and they cer-
tainly deserve attention in wildlife habitat and popu-
lation inventories.

Roughly 40 species of bats exist in North Amer-
ica north of Mexico (Jones et al. 1982; Hall 1981).
In some habitats the number of bat species exceeds
that of all other mammal species combined, and
some species can occur in very large numbers if
conditions are favorable. Bats have many distinctive
morphological, physiological, and behavioral adapta-
tions that are associated with some unique uses of
habitat. Consequently, methods of inventory and
monitoring of habitat and populations also tend to
be unique. Only methods most suitable for applica-
tion to the bats of temperate North America will be
considered here.

A knowledge of basic behavioral characteristics,
especially daily and seasonal activity patterns, is es-
sential when contemplating any study of bats. Be-
cause bats are extremely mobile, some habitat inven-
tory and monitoring techniques are similar to those
for birds. Unlike most birds, however, bats are active
at night and are not easily detected or distinguished
by direct observation or unaided audition. Further,
the local distribution of bats is usually patchy and
uneven. This is caused by variation in the distribu-
tion of essential habitat resources and seasonally
variable tendencies of bats to congregate at specific
and often unique resting, breeding, feeding, and
watering places. Consequently, although population
measurements and determination of species presence
may be accomplished by censusing at such congrega-
tion sites, determination of absolute density is vir-
tually impossible.

Daily and seasonal activity patterns are impor-
tant aspects of behavior that are directly related to
inventory of populations and analysis of habitat asso-
ciations. During the summer a typical daily activity
pattern for North American bats begins with emer-

Bats

497

Pallid bat.

gence from a daytime shelter (roost) during evening
twilight or after dark. A roost is any site where a bat
lands when it voluntarily ceases flying. By conven-
tion, roosts usually include resting sites where food
may be devoured and digested but exclude sites
where a bat lands to capture prey. Some species or-
dinarily congregate in large numbers whereas others
roost singly.

Bats may begin feeding immediately after leav-
ing the day roost or they may fly several miles be-
fore reaching a foraging site. Most North American
species feed on flying insects. After feeding, they may
fly to a drinking site or, alternatively, a night roost. A
night roost is generally not at the same site as a day
roost. Night roosts are occupied for various lengths
of time depending upon the species and purpose for
which the roost is being used. Some night roosts
may be used simultaneously by several species that
do not normally share a day roost.

The active-inactive pattern may be repeated one
or two times during the night after which the bats
return to their original day roost. This daily pattern

varies depending on the species, geographic location,
season, and environmental conditions. Water require-
ments and drinking activity vary with the tempera-
ture and humidity. During late spring or early sum-
mer the females of many species congregate to form
colonies at maternity roost sites where parturition
and maturation of the young take place. During this
period, water requirements may be elevated due to
lactation and high environmental temperatures, and
the bats may begin their nocturnal activity with a
trip to a drinking site. While females congregate in
maternity colonies, males of some species may form
separate bachelor colonies.

Some night roosts are used primarily in the late
summer and fall, presumably for breeding. Congrega-
tion at such sites is often referred to as swarming.
For some species that are usually found solitary or
sexually segregated, late summer and fall is a time
when individuals come together, thus facilitating
mating and reducing the probability of inbreeding.

Roosts used during the winter for dormancy are
called hibernacula. Such roosts may be at different

498

Bats

elevations or in different geographic regions than
those of summer roosts. Often these are the same
sites visited during swarming earlier in the year.
Those species that do not hibernate may undertake
extensive latitudinal migration to find suitable condi-
tions during different periods of the year. It becomes
evident that several different types of roosts, each
with distinct characteristics, are required throughout
the year for many species.

Although techniques are not standardized, there
is a logical sequence of procedures that may be fol-
lowed when investigating bat populations and their
habitat relationships at a local level. The first step is
to compile a list of species potentially in the area of
concern by reviewing publications on mammal dis-
tribution in North America (e.g., Hall 1981), one
that deals specifically with the distribution of bats
(e.g., Barbour and Davis 1969), or one that deals
with the distribution of mammals or bats in a more
specific region (e.g., Maser et al. 1981). The known
habitat affinities may also be determined by consult-
ing general publications on bats, especially Barbour
and Davis (1969), or on regional natural history. For
many regions in North America, one may consult the
very specific descriptions of wildlife-habitat associa-
tions and distribution that have been compiled in re-
cent years (e.g., Thomas 1979).

The next phase of study depends on the ques-
tions) being asked. One question may be simply
whether a given species or group of different species
is present in an area or habitat. Techniques for veri-
fying the presence of predicted species should be se-
lected with the habits of the species in mind. For
some species an investigation of potential roost sites
is appropriate. For species whose roost sites are inac-
cessible or little known, a technique such as mist-
netting at water holes or foraging areas may be best.
Even more specialized techniques, such as the use of
ultrasonic detectors, may be appropriate for some
species. These techniques may complement one an-
other depending on the objectives of the study (e.g.,
see Paige et al. 1985). The important point is that
many species or groups of species have distinctive
behaviors that must be taken into consideration
when attempting to verify presence.

Questions related to the use of specific habitat
features are sometimes difficult to answer. Determin-
ing whether bats use a given cave as a roost site is
relatively simple, but the investigator must be aware
that just because bats or their sign are not found at
one time of a day or year does not necessarily mean
that they do not use the site at another time. Be-
cause some roost areas, such as cliffs and trees, are
often inaccessible, it is very difficult to establish hab-
itat relationships.

Questions regarding population size also require
consideration of both general and specific behaviors.

Although measurement of absolute density is not a
feasible objective, an estimation of population size
utilizing a specific area or habitat feature is possible
for those species whose members congregate at spe-
cific locations. For many species these congregation
sites are the summer day roosts, especially maternity
colonies, or winter hibernacula. Even though there
may be a great distance between such sites, most
bats seem to exhibit loyalty to them — at least on a
seasonal basis. This fidelity makes it possible to com-
pare population size at a given site during the same
time from year to year. Other species may use a par-
ticular site year around. In any event, knowledge of
such species-specific behavior is imperative when
framing questions about the importance of a habitat
feature or change that might influence population
size.

For species that do not congregate at accessible
daytime roosts or hibernacula, it may be possible to
obtain rough estimates of population size by captur-
ing them in other areas of concentration. Although
drinking sites are often difficult places to estimate
species population sizes, they may be the best places
to assess bat community structure. Whereas many
species segregate for day roosting and dormancy,
they may come together at common drinking sites
during the summer. To a lesser extent this may also
be true for night roosts during the fall. However,
both drinking places and night roosts may have char-
acteristics that are differentially attractive to various
species so one must not assume that samples taken
at those sites are necessarily representative of local
community structure.

When population estimates are compared spa-
tially or temporally, the techniques need to be iden-
tical. In contrast, to determine bat community com-
position in a particular area, as many sampling
techniques as possible should be employed. Relative
density of the entire community may not be easily
measured, but it is possible to measure the portion
of the community using a common local resource,
and the data obtained need only be qualified as such
to be useful for descriptive or comparative purposes.

HABITAT FEATURES CORRELATED
WITH BATS

Many bats are closely tied to specific habitat fea-
tures. The primary activities associated with such
features are drinking, feeding, and roosting. Drinking
is obviously associated with physical features, partic-
ularly the presence of a suitable source of water,
whereas feeding often shows some association with
vegetative features. Because roosting behavior may
be associated with both physical and vegetative fea-
tures, a brief consideration of the general physical
attributes of the various functional categories or
roosts follows.

Bats

499

Day roosts are generally inaccessible to preda-
tors and have relatively stable microclimatic condi-
tions thus providing safe and physiologically efficient
places to spend most of the inactive time. Much less
is known about night roosts but they seem essential
for bat survival. They are often located much closer
to foraging and drinking sites than day roosts and
provide a relatively safe place to rest, eat and digest
food, and groom. The physical structure of night
roosts is often distinctly different from that of day
roosts. For example, many bats that roost in crevices
during the day select caves or other relatively open
sites as night roosts.

The largest congregations of bats occur in ma-
ternity roosts. Even some bats that are solitary dur-
ing most of the year may congregate in small colo-
nies for birthing purposes. This indicates either that
sites with the proper conditions for birthing are
scarce or that the presence of large numbers of bats
is necessary to create the proper roost conditions.
Maternity roosts are warmer than other roosts — a
condition which increases the growth rate of young
during gestation and afterbirth (McNab 1982).

Maternity colony of the Yuma Myotis in a man-made struc-
ture.

Conditions in bat hibernacula that are crucial
for survival include relatively stable temperatures
that are low enough to foster the reduced metabolic
activity necessary to conserve energy, but not so low
as to be lethal. For a comprehensive treatment of
current knowledge of bat roosting ecology, see Kunz
(1982).

Physical Features

Compared with other wildlife, bats exhibit a rel-
atively strong association with prominent physical
features of their habitat. Most of these physical fea-
tures fall into the category that Maser et al. ( 1979)
refer to as "unique" habitats. These sites make up a

very small percentage of the total land base but are
often disproportionately important as habitat. Distri-
bution and abundance of bats are often directly re-
lated to the availability of these habitat features,
which may be natural or man-made structures.

There are two principal activities associated di-
rectly with the use of physical habitat features. The
first, roosting, occurs in a variety of natural and man-
made structures. The second, drinking, usually
requires an open water source. Specific structural
features are described under these two functional
headings.

Roosting sites. Physical features that serve as roost-
ing sites may be categorized in a number of ways.
The following list is not intended to be comprehen-
sive but does include the structures used most often
for roosts. Bats appear to choose these structures on
the basis of their specific and relatively constant mi-
croclimatic conditions and lack of disturbance.

Caves and Cave-like Structures. Caves are uti-
lized by bats more than by any other group of wild-
life. Caves provide shelter from adverse weather, a
relatively stable environment, darkness, and protec-
tion from predators. Bats may roost in caves singly
or in large congregations. As many as 25 million Bra-
zilian free-tailed bats (Tadarida brasiliensis) have
been found in single caves in the southwestern
United States in the summer (Cockrum 1969; Davis
et al. 1962). Some species use caves or cave-like
structures for all roosting activities whereas others
use them only for one type of roosting behavior. For
example, Townsend's big-eared bat {Plecotus town-
sendii) and the gray myotis {Myotis grisescens)
seem restricted to cave-like roosts, whereas the long-
eared myotis (M. evotis) and long-legged myotis (M
volans) ordinarily use caves only as night roosts
(Barbour and Davis 1969).

Many caves are unsuitable as bat roosts and
some are used regularly for only a portion of the
year. The important characteristics of a cave are tem-
perature, humidity, light, size, and interior configura-
tion and surface texture. Caves that offer the most
optimal range of all of these features are those most
likely to be used by the greatest number of individu-
als and species. Caves with large spacious chambers
are more likely to house large aggregations. Cold and
damp caves are more suitable as hibernacula or as
summer night roosts. Warm caves are more suitable
for maternity roosts or for species that are active
year-round. Total darkness does not seem to be an
absolute requirement for most cave dwelling bats, al-
though they may select the darkest portion of a par-
ticular cave. Caves used as day roosts are often more
likely to be occupied simultaneously by fewer spe-
cies than those used as night roosts.

500

Bats

A group of Townsend's big-eared bats roosting on the ceiling of a lava tube at Lava Beds National Monument.

Many types of man-made cave-like structures
have access to the surface. Tunnels and mines are
the most obvious. Although not usually as structur-
ally diverse as natural caves, they may be used by
bats for all types of roosting. Generally they are eas-
ier for humans to find and access than natural caves
because of habitat modifications related to excava-
tion. Not only are the locations more visible but
they are often recorded on legal or topographic
maps. Old mine shafts and tunnels are often very
dangerous — great care should be exercised when in-
vestigating such abandoned excavations.

Cliffs, Crevices, and Talus Slopes. Cliffs pro-
vide some roosting structures — especially crevices —
favored by many bat species. Rock crevices in cliffs
may have more suitable temperature and moisture
conditions for bats than rock crevices elsewhere be-
cause of their vertical exposure and large mass asso-
ciated with large heat-retaining capacity. Further,
cliffs provide a safer refuge from predators than
roost structures more accessible to flat terrain.

The largest and smallest species of bats in the
U.S. are cliff-crevice dwellers. The greater mastiff bat
(Eumops perotis), roosting in small groups, occupies
large crevices with downward facing openings at
least twice as wide as its body (Vaughan 1959). The
western pipestrelle (Pipistrellus hesperus) usually
roosts singly in small, up to 2.5 cm (1 in.) wide, ver-
tical rock crevices (Cross 1965). A few other species
have been found in intermediate situations but, as
Kunz ( 1982) pointed out, little is known about the
ecology of crevice-dwelling bats because of the diffi-
culty of finding them in these relatively inaccessible
places.

In general, those cliffs with greatest fracturing
are more likely to provide a diversity of rock crev-
ices from which bats may choose roosting sites. Ori-
entation of cliff faces with respect to the sun may be
important for determining the temperature condi-
tions of the associated crevices or other cavities. Oc-
cupied cliff crevice roosts offer some thermal buffer-
ing and a variety of thermal choices (O'Shea and
Vaughan 1977; Hayward and Cross 1979).

Bats

501

Talus is the accumulation of broken rocks at the
base of cliffs or other steep slopes. Talus creates a
variety of crevices, cavities, and even small caves, de-
pending on the size of the breakdown debris. Bats
roosting in such habitat are more susceptible to nat-
ural predation than in caves or cliffs.

Man-Made Structures. Many species of North
American bats readily use certain man-made struc-
tures as roost sites. These structures appear to sup-
ply suitable substitutes for "natural" roost sites. Some
species, such as the big brown bat (Eptesicus fus-
cus), the little brown bat (Myotis lucifugus), and
the Yuma myotis (M. yumanensis), show a great de-
pendency upon these roosts, especially as maternity
sites (Barbour and Davis 1969). There is evidence to
indicate that both ranges and population sizes of
some species have increased as a result of using
man-made structures for roosts (see Kunz 1982). A
great variety of man-made structures are used by
bats. In addition to the cave-like structures men-
tioned above, the other structures may be grouped
into three categories: buildings, bridges, and miscel-
laneous structures.

Buildings offer an endless variety of roosting
sites that environmentally resemble natural crevices,
caves, or cavities. For example, spaces under roofing
or siding provide crevice-like structures. Attics, base-
ments, and abandoned buildings provide cave-like
conditions. Intermediate-sized spaces may resemble
natural rock or tree cavities. Many semi-open struc-
tures such as carports, porches, or woodsheds serve
as temporary night roosts. Some structures have
been built specifically to attract roosting bats (see
Greenhall and Paradiso 1968; Greenhall 1982).

Bridges also provide a variety of roosting sites,
sometimes in remote areas otherwise having rela-
tively featureless terrain (Davis and Cockrum 1963).
Some concrete bridges have cave-like chambers in
the ends. Several types of crevices are available, es-
pecially in older style bridges. Modern bridges tend
to have fewer protected roosting sites but some are
still useful as night roosts and, in some instances,
could be modified to serve as day roosts.

Other potentially useful man-made structures in-
clude dams, windmill support structures, storm sew-
ers, and any structures that provide cavities, crev-
ices, or cave-like areas with suitable temperature,
humidity, light, and protection.

Drinking Sites. The availability of open water is
crucial to the survival of most bats, and therefore
greatly influences their distribution and abundance.
Most bats drink while flying, which means that the
surface area of the open water must be sufficiently
large to accommodate this behavior. Rapid flyers,

Highway bridge.

Crevices formed between wooden planks of bridge.

Bats roosting in one of the crevices.

502

Bats

I

such as members of the family Molossidae (free-
tailed bats), need a very large surface area from
which to drink whereas some of the highly agile
flyers, such as several small Myotis, can drink from a
pool only a few centimeters in diameter while still in
flight. Because free water is an essential requirement
for most species, seasonal fluctuations in the amount
and distribution of open water greatly influence the
local distribution of bats, especially in arid regions
where water is scarce.

Man has inadvertently made many otherwise un-
inhabitable regions useful to bats by providing water
in the form of seasonal storage reservoirs, ponds and
watering devices for livestock, heliponds, some types
of wildlife water catchment devices (guzzlers), and
swimming pools. Any habitat modification that re-
sults in the availability of open water during dry pe-
riods may result in greater use of the areas by bats if
other requirements are available.

teris mexicana) feeds on nectar. The Mexican long-
nosed bat (Leptonyc teris nivalis) and the other
long-nosed bat (L. yerbabuenae or sanborni) feed
on nectar but also utilize pollen and fruit.

Insectivorous species may be more numerous
near areas such as marshes and agricultural lands
where the vegetation supports high insect popula-
tions. Often in such places, chemical insecticides are
used to control the insects and there is evidence
that these accumulate secondarily in the tissues of
some bats and may be harmful to their populations
(Geluso et al. 1976; Henny et al. 1982). Some forest
dwelling bats respond to tree density on the basis of
foraging style. Those that are strong flyers and catch
their insect prey in flight need open forests whereas
those that are very agile or glean their food from the
vegetation can utilize more densely vegetated areas.

Discussion

Vegetation Features

Little information is available regarding bat-vege-
tation associations. It appears that most North Ameri-
can bats respond to major structural differences in
vegetation rather than to vegetation species differ-
ences. Roosting and feeding are two prominent activ-
ities that are often associated with vegetative fea-
tures. Because bats are very mobile, these two
activities may occur in widely separated and vegeta-
tively distinct areas.

Roosting Sites. Some North American bats roost on
vegetation but relatively little is known about spe-
cific requirements. Most known vegetative roosts are
daytime retreats, but some evidence exists that bats
may also use trees and shrubs as night roosts (Kunz
1982). Vegetative day roosts are most prominently
associated with living trees or snags and logs.

A few North American species regularly roost in
the foliage of trees. These include the hoary bat
(Lasiurus cinereus), red bat (L borealis), northern
yellow bat (L. intermedins), seminole bat (L. semi-
nolus), and probably the silver-haired bat {Lasionyc-
teris noctivagans). Some species, such as the big
brown bat and the pallid bat {Antrozous pallidus),
may use tree cavities or crevices. Species such as
Keen's myotis (M. keenii) and long-eared myotis
have been found in tree crevices or under exfoliat-
ing bark, primarily in standing trees (Barbour and
Davis 1969). Many other species are also thought to
use these sites because of their close association
with forested habitats.

Because bats are closely associated with distinc-
tive habitat features, their occupancy of a particular
area may be predicted to some extent on the basis

Feeding Relationships. Three species with limited
distribution in the southwestern U.S. use plants di-
rectly for food. The long-tongued bat (Cboeronyc-

Conifer snag with hole and exfoliating bark — creating
roosting sites for bats.

Bats

503

of the presence or absence of these features. The
best places to look for bats or signs of their presence
are in and around essential habitat features. Any geo-
graphic area of interest should first be surveyed for
potential roost and drinking sites. In addition to the
presence or absence of essential habitat features,
their proximity to other features is also important.
For example, the absence of open water within a few
miles of potential roosting or feeding sites will limit
some species. The available surface of open water
will limit others. The types of roost sites available
will also limit some species. Cave and crevice dwell-
ers are often restricted to just those specific roost
structures.

Apparent or potential roost structures should be
investigated first. If bats are present, it is possible to
obtain samples for species identification and esti-
mates of population size. However, because bats are
only associated with some roosts seasonally, absence
from a potential roost site may not mean it is not
used. If bat guano or other sign, such as urine stains
or remains of chewed insects are found, the roost
should be visited at other times for sampling. In
some instances, bats may use a roost structure with-
out leaving obvious evidence of their visits. This is
true at some night roosts where several hundred
bats may use a site without leaving much sign. If
such a situation seems possible based on other types
of evidence of bats presence in the area, the investi-
gator must visit the roost at different times to ascer-
tain usage.

Potential water sources should be investigated.
If water is limited, bats may be highly concentrated
at available sources. The presence of bats in an area
may sometimes be determined easily by simple ob-
servation at an isolated water source during the eve-
ning activity period. Observations should be sched-
uled on partially moon-lit nights or may be aided by
a low intensity light, such as a gas lantern, to en-
hance observation after dark.

In addition to drinking places, bats may also be
seen in their foraging areas, especially those species
that begin their activities during the twilight period.
Some species can be identified by their characteristic
flight but this often requires considerable experi-
ence. Even if species are not identified, the relative
number of bats associated with different habitat fea-
tures may be estimated.

POPULATION MEASUREMENT

Measuring populations of nocturnal flying mam-
mals presents some unusual problems. It is virtually
impossible or impractical to achieve accurate mea-
surements of absolute density. However, because
many North American bat species tend to congregate
at preferred roosting sites or at isolated water

sources, it is often feasible to determine presence
and broad habitat associations. In addition, the popu-
lation features of size, relative density, and age struc-
ture of bats using the congregation sites may often
be obtained.

Presence

Presence may often be determined by simple
observation. Bats are sometimes readily observed
while feeding, drinking, or roosting. Counts may be
made but species identification usually involves cap-
turing some of the animals. There are three practical
capture methods: hand capture, mist netting, and
trapping. These capture techniques all involve simi-
lar handling methods associated with species identifi-
cation, sexing, aging, and marking. Handling tech-
niques are described in association with the first
sampling method. For more detailed and complete
accounts, see "Bats and Bat Banding" by Greenhall
and Paradiso (1968) or "Behavioral and Ecological
Methods for the Study of Bats," T.H. Kunz, editor (in
press).

Hand Capture. Opportunities to capture bats di-
rectly with the hand are rare. Such instances usually
come about when bats are in a torpid state, such as
when hibernating or when a bat is sick or injured
and partially immobilized. Bats should be handled
only if absolutely necessary and gloves should be
worn. Arousal from torpor may cause serious deple-
tion of stored energy that might lead to death at a
later time. Handling or disturbance of some species
can force them to abandon a preferred roost for one
that is less favorable for survival.

Hand-capture techniques also include those in-
volving some implement manipulated by hand. It is
common to encounter bats that are out of normal
reach or grasp, often in crevices or flying in a con-
fined space such as a cave or building. There are two
implements that are particularly valuable aids for
capturing such bats. Long-nose scissor forceps (bot-
tle forceps) are useful for reaching into cracks, crev-
ices, and corners where a hand will not fit or cannot
be extended. The tips of the forceps should be pad-
ded by covering them with rubber tubing, plastic
rubber, or some substance that will reduce chances
of injury. Another useful capture tool is a wide-
mouth insect net, preferably with an adjustable-
length handle. The net can be used to capture bats
in flight, especially in confined spaces, or may be
placed over a small roost opening or over bats that
are roosting in the open but are out of reach. For
capturing bats in flight, sweeping from behind may
be more effective and have less chance of causing in-
jury than a head-on sweep. Both of these equipment
items may be purchased from biological supply com-
panies (see Dowler and Genoways 1976) and are
relatively inexpensive. No special training is required
for their use. Efficient use of both of these capture

504

Bats

implements will usually be aided by a hand-held
flashlight or, preferably, a head lamp.

Handling Techniques. Whether captured by
hand or by some other method, it is often necessary
to handle individual bats. Handling may be necessary
to identify species, determine sex and age, or mark
the bats for population estimates. Handling bats
should be done with care to avoid injuring the ani-
mals or being bitten. Like most wild mammals, bats
will attempt to defend themselves by biting. They
are capable of carrying rabies. Although the inci-
dence of rabies is not unusually high (estimated at
less than 1% in natural populations), some precau-
tions should be taken such as wearing gloves and,
where there is unusually high risk of bites, acquiring
protection from rabies by pre-exposure immuniza-
tion. Information regarding preexposure immuniza-
tion may be obtained from a physician or local pub-
lic health service agency. In most instances a pair of
pliable leather gloves will provide adequate protec-
tion against bites and allow enough feeling for sensi-
tive manipulation of the bat.

Care should be taken not to apply too much
pressure to a bat during handling because their wing
(finger) bones are easily broken and some species
are easily suffocated or may go into shock. Placing
the thumb or forefinger under the chin of the bat
while cradling the body in the palm of the hand al-
lows immobilization without excessive pressure (Fig-
ure 1 ). Alternatively, a bat may be held with its two

wings open by grasping each wing with the thumb
and forefinger at the distal end of the bat's forearm
near the clawed thumb. Using this method requires
that measurements and marking be done by a second
worker. Although the bat's head (and teeth) are far-
ther from the handler's fingers, many bats will strug-
gle violently in this position and there is often a
greater chance of either being bitten or injuring the
bat. Considering these problems and the difficulty in
grabbing a bat in this position to start with, it seems
more sensible to use the finger-under-the-chin

A long-eared Myotis handled by grasping wings near the
wrist.

Figure 1. An effective and relatively safe way to hold a bat without being bitten
or injuring the bat. The worker's forefinger or thumb is placed under the
bat's lower jaw while the remaining fingers restrain the body and wings. This
leaves the other hand free for manipulation, marking, or measuring.

Bats

505

A big brown bat held with the wings extended. When holding by the wings, take care to avoid grasping the delicate finger
(wing) bones.

Grasping this Townsend "s big-eared bat with thumb and forefinger of one hand allows for manipulation with the other
hand.

506

Bats

method of immobilization. The most difficult part of
this method is the initial seizure of the animal, which
is most easily accomplished by pinning it against
something solid.

It is often necessary to retain some individual
bats temporarily. For small numbers a cloth sack is
adequate as a holding container. For larger numbers
additional sacks or some type of cage-like container
may be necessary. Greenhall and Paradiso ( 1968) de-
scribed a number of such holding cages. If a cage is
deemed necessary, it should be constructed to fit the
specific conditions under which it is likely to be
used.

The first task of inspection is species identifica-
tion. This is relatively simple for some species but
difficult for others. A good place to start is with a
key to the bat species of North America such as that
found in Barbour and Davis (1969). There have been
some taxonomic revisions since that work was pub-
lished, so it would be wise to check a more recent
publication, perhaps one that is focused on the spe-
cific region of work. For some species there is no
foolproof method of identification in the field, such
as distinguishing the Yuma myotis from the little
brown myotis in areas where they occur together.
Some other species of Myotis are particularly diffi-
cult to identify without experience. Time spent with
a regional key and collection, such as one might find
at a local college or university, may be helpful.

It is often necessary to determine sex, and
sometimes age, of individuals so that the type of ag-
gregation (breeding, maternity, or bachelor) may be
ascertained. The sex of bats is generally easy to de-
termine. The males have an obvious penis and the
clitoris of the females is not enlarged. Testes of
males and mammary nipples of females are seasona-
bly obvious. Aging is difficult after the juvenile pe-
riod. Juveniles less than 4-5 months old can be rec-
ognized by the cartilaginous areas in the finger
joints. The joints are lighter in color and relatively
smooth compared with the darker, knobby appear-
ance of the adult joints. Tooth wear can only give a
general impression of age; canines and premolars be-
come obviously round in very old individuals.

After inspection, bats may be released by care-
fully launching them into flight such as one might
pitch a horseshoe, or by quickly releasing the grip
and allowing them to take off on their own. The
launching method is preferred for fast flying species
that need some initial speed to gain flight and may-
be used for all species unless holding time is pro-
longed. Some species become progressively more le-
thargic the longer they are held in captivity, espe-
cially under cold conditions, and they should not be
launched or dropped. The best release method in
this situation is to simply place the bat as high as
possible above the ground on a solid substrate such

as a rock, wall, or tree. As soon as the bat warms up
it will fly off on its own. Bats released from grasp in
this fashion will often attempt to bite if the restrain-
ing hold is gradually reduced. Those released by
launching or rapid relaxation of the grip usually
make no attempt to bite.

Marking Techniques. Bats are often marked to
aid in determining movements or to identify them as
previously captured individuals for population esti-
mates. A variety of marking techniques are available,
but two — banding and punch-marking — are most
useful for studying populations related to habitat in-
ventories. For details of other techniques, see Bar-
clay (in press).

Banding involves placing a small plastic band
around the distal end of the forearm where the lead-
ing edge of the wing membrane is very narrow. For
the North American species being considered here,
use of lipped or flanged plastic bands applied loosely
so they slide freely along the forearm will reduce in-
juries to the bats (Stebbings 1978).

Banding with aluminum bands should be
avoided because identification numbers may be lost
as a result of chewing by the bats. The use of plastic
bands with large embossed numbers or color codes
will solve this problem (Bonaccorso et al. 1976). For
local studies of abundance and movements, the col-
ored plastic bands are easily coded to represent date
and location of banding. Bands may be purchased
from A.C. Hughes, 1 High St., Hampton Hill, Middle-
sex, Eng. TW121WA (colored plastic); Gey Band and
Tag Co., Box 363, Norristown, PA 19404; National
Band and Tag Co., 721 York St., Newport, KY 41071;
and Ball Chain Mfg. Co., Inc., 741 South Fulton Ave.,
Mt. Vernon, NY 10550.

Aluminum band on the forearm of a long-legged Myotis.

Bats

507

A second method of marking bats is punch-
marking (Bonaccorso and Smythe 1972). The tech-
nique consists of punching small holes in the form of
numbers through the outstretched wing membrane.
This is accomplished with a tattoo instrument nor-
mally used to mark domestic livestock. These instru-
ments may be purchased from veterinary supply
companies or some livestock supply stores, and they
come in at least two sizes, for large and small ani-
mals. With care, the large animal tattoo marker can
be used for all species of North American bats. It has
the advantage of larger perforations (both size of
numbers and size of holes) that are more legible and
longer lasting than those made by the small animal
instrument. Although initially touted as a suitable
substitute for banding, it was quickly learned that
punch-marking had the drawback of being short-
lived (Bonaccorso et al. 1976). Punch marks remain
legible for only about 5 months. Nevertheless, the
advantages are numerous and the technique is suita-
ble for short-term population inventories related to
habitat monitoring.

Several publications are available that describe
the technique in considerable detail. See Bleitz
( 1984) and Keyes and Grue ( 1982) for general pro-
cedure and review of history and techniques. See
Greenhall and Paradiso (1968) and Kunz (in press)
for application of the techniques to bats. A general
description of the technique will be given here with
emphasis on those areas that are important or
unique with respect to bats. This information should
allow an investigator to determine if the technique is
suitable for a given bat sampling situation. The
proper use of mist nets requires experience and is

Punch-marking a long-eared Myotis with a large animal
tattoo device.

A Yuma Myotis with punch-marked wing.

greatly facilitated by field training. If instruction is
not available from someone working with bats, the
best alternative is to work with a knowledgeable
bird netter. Names and addresses of such individuals
may be obtained from a regional bird banding associ-
ation (see Bleitz 1984). Attempts to "start from
scratch" may result in injury or stress to the bats,
damage to the nets, and frustration to the investiga-
tor.

Each individual bat may be marked with a dis-
tinct number but this is usually unnecessary. A sim-
pler and adequate approach for population census
and local movement studies is to use a distinctive
numerical sequence (letters may be included) for
each marking episode at each location. If necessary,
additional punch marks may be made in the same
wing or opposite wing on subsequent encounters.

Mist-Netting. The mist net was developed by the
Japanese to capture birds and is still used extensively
for that purpose. Because mist-netting is the most
commonly used capture method in bird research, it
has received a great deal of attention from ornitholo-
gists (see review by Keyes and Grue 1982). The
basic technique and many of the recent innovations
developed by bird netters are readily applicable to
bats.

Basic Technique. The basic technique consists
of placing a large rectangular net, having several tiers
made of fine hairnet-like material, in a position
where it will be in the pathway of flying bats. For
best results, these nets are placed where bats are
concentrated such as near their unique habitat fea-
tures or drinking sites. Bats attempting to enter or
leave a roost site, drink water from an open water
source, or forage in a relatively confined area are
vulnerable to capture with mist nets. Bats are cap-
tured when they strike the netting and fall into ham-
mock-like pouches at the lower side of each shelf of
the net (Figure 2). Entangled bats are removed from
the net by hand.

Equipment. Mist nets are available in a variety
of colors, sizes, and construction material. A list of
suppliers may be found in Dowler and Genoways

508

Bats

The wing of a long-eared Myotis with perforations resulting from the punch-marking process. The tattoo device used for
making the marks is shown below the wing.

Figure 2. Diagrammatic view of a mist net setup. The net has five cross lines forming four tiers and associated
pockets where the bats are usually entrapped.

Bats

509

(1976). The capture location, situation, and size of
bats to be captured should be considered when de-
termining the most suitable net characteristics. Black
nets are suitable for bats because most capturing is
done at night and the color does not contrast with
the background.

Net sizes vary in length, height, number of tiers,
and mesh. All sizes for nets are given in the fully-
stretched position. The most common sizes used for
bat netting are about 5m(18ft), 9m(30ft), and
13 m (42 ft). Nets longer than 13 m (42 ft) are diffi-
cult to operate without some sort of center support
or heavy guying. Short-length nets are useful for
roost entrances, small ponds, or foraging pathways in
heavy vegetation. The 9-m (30-ft) net is very useful
for netting over small ponds or in foraging areas
with larger open spaces. Most nets used for captur-
ing bats are about 2 m (7 ft) high and have four tiers
(also referred to as shelves or pockets). The working
height becomes less than 2 m (6.6 ft) when the
cross lines (trammels) are compressed to form the
pockets. Nets with two or five shelves are also com-
mercially available. Other widths (heights) may be
obtained by removing shelves of the standard nets or
by stacking nets to cover a greater vertical distance.

Mesh sizes (designated by the diagonal of each
square) useful for netting North American bats range
from 3-8 cm (1.5 in.) for small species to 6.4 cm
(2.5 in.) for large species. Because many regions sup-
port a combination of large and small species, it is
safest to use a relatively small mesh size.

Nylon and terylene are common construction
materials; terylene is stronger but less pliable.
Strength is determined by the "ply" (number of
strands per thread) and "denier" (a weight measure-
ment). Nets for capturing bats are usually 2 ply and
range from 50 to 110 denier (d), are black color, ny-
lon, 70 d, 2 ply, 3-8-cm (1.5-in.) mesh, 4 tiers, 2 m
(7 ft) high, and 9 m (30 ft) long.

Many types of poles may be used to hold the
ends of a mist net in place. Electrical conduit, which
may be cut into portable sections that can be cou-
pled in the field, is very commonly used for this pur-
pose. Keyes and Grue ( 1982) described many of the
available alternatives. Because deployment condi-
tions usually vary, sometimes greatly between sites,
it is advisable to have versatile equipment.

Nylon cord is an essential item for guying the
tops of any pole setup. Other useful items include a
headlamp, so both hands are free for removing bats
from the net; gloves; cloth holding sacks that are eas-
ily carried under the belt; and insect repellent. When
netting over water, hip boots or chest waders are
often necessary to facilitate removing bats from the
net. Another alternative in such situations is to use a
boat or raft. Some items that appear to be luxuries,

such as a comfortable chair and warm clothing, often
become necessities as a night of netting wears on.

Operation. Nets should be set up well before
dark but the cross lines should be bunched until bats
are seen flying or when fully dark. This prevents cap-
turing birds that are active during the daylight or
twilight periods. Position of the nets is often critical
and influences capture success. If possible, the nets
should be set so they are perpendicular to the direc-
tion that the bats are most likely to fly. This is rela-
tively easy to determine at roost sites but often in-
volves experience with the particular situation at
drinking and foraging sites (Figure 3). If possible,
several nets should be set up in such situations so as
to find the most efficient positions. Generally, when
nets are set over somewhat rectangular bodies of
water they should be positioned perpendicular to
the long axis of the pond. Other features such as ob-
structions to flight should also be considered and the
nets placed where there is the least interference
with the bats' flight patterns. In open areas where
there are no criteria for orientation of a single net,
several nets may be set at different angles. Common
configurations are V, L, and triangles. Vertical stack-
ing of nets to achieve greater height is a variation
particularly useful in foraging areas.

Nets should be checked regularly — at least
every 15 minutes. Bats captured in a net should be
removed as quickly as possible to avoid confounding
entanglements and holes in the net resulting from
the bats chewing or the netter having to break the
strands to free the bats. A captured bat should be re-
moved from the side of the net from which it en-
tered a pocket. A good sequence for untangling is
head, one wing, other wing, and body.

When removing a bat from a net, the bat may
be grasped in such a way as to immobilize the
mouth, as described previously, or by using a cloth
baffle on the opposite side of the net for support and
to provide something the bat can get its teeth into.
The cloth baffle is also useful as a backing to get an
initial hold on the bat. Inducing a bat to release its
bite on the baffle or other object (such as a finger)
may be achieved by mild pulling pressure, by touch-
ing the back of the bat's head, by blowing on the an-
imal, or by a combination of these strategies. After
removal from the net, the bats may be held in a
cloth sack or holding cage, or marked and released
immediately. Bats flying nearby may be attracted to
the calls of captives held near a net or trap site.

The duration and timing of netting depends
upon the objectives of the investigator. It is rela-
tively easy to determine when the evening emer-
gence flight of bats from a day roost is finished. In
contrast, the period of greatest bat activity at night
roosts may be difficult to predict and may occur dur-
ing the middle of the night. Netting at foraging sites

510

Bats

***- y*-i

wmrn

Figure 3. Some possible placements for mist nets to capture bats.

Bats

511

1/7 //
hi hit

y^i ff/h

\ :;J f

u ■/ /.

V V

\ \, ' ' \

in places where there are extremely high concentra-
tions of bats. For example, at the entrance of some
cave roosting sites, the captured bats cannot be un-
tangled quickly enough to avoid bat injury or dam-
aged nets. Windy conditions greatly diminish netting
success because of reduced bat activity or decreased
net effectiveness.

Trapping. An alternative to mist-netting is trapping,
first developed by Constantine (1958). He used a so-
called "harp trap" which is designed to capture and
hold bats. Like a mist net, a trap is placed where it is
likely to intercept passing bats. Unlike the mist net,
it does not require constant attention and can be
used where the density of flying bats is very high
and where space is restricted.

Basic Technique. The basic technique consists
of placing a metal frame, strung with vertical strands
of wire or nylon, directly in the path of flying bats
(Figure 4). The strands of the trap are closely spaced
and kept taut so that when a bat strikes them it will
lose control of its flight and slide to the bottom of
the frame. A sack-like container is suspended from
the bottom of the frame to serve as a retaining struc-
ture or as a funnel to direct the bats into a holding
container. Bats captured in such a device may be
handled immediately or left until a more convenient
time for the investigator.

A bat entrapped in the pocket of a mist net. Captured bats
should be removed as soon as possible to avoid confound-
ing entanglements. Bats will also chew the netting which
results in holes (pictured).

or at drinking sites is usually most productive within
2 to 3 hours after sunset. However, it appears that
different species have different peak periods for feed-
ing and drinking so it is desirable to sample for as
long as possible for species presence and to sample
for a constant period when attempting to estimate
population size for comparative purposes.

Upon completion of a sampling episode the nets
should be dried, if necessary, and all debris removed.
Each net should be folded in such a manner that it
can be easily unfolded for the next use (e.g., see
Greenhall and Paradiso 1968). Proper care at the
end of one session, including tying lines so they stay
in order and making notations regarding the condi-
tion of the net, will greatly facilitate setup at the
next session.

Although mist-netting is a versatile and fairly ef-
ficient method of capturing bats, it has some limita-
tions. Netting is most efficient under very specific
circumstances, e.g., in areas where bats concentrate
to forage or drink and where there is sufficient space
to deploy the net. Mist nets are often inappropriate

Equipment. Bat traps are not readily available
commercially but are fairly easy to build. Informa-
tion presented here will give a rough idea of design;
details of construction are given in the references
cited below.

The original design (Constantine 1958; see also
Greenhall and Paradiso 1968), commonly referred to
as the Constantine trap, consists of a single rectangu-
lar aluminum frame with vertically arranged taut
steel wires spaced 1 in. (2.5 cm) apart. Bats stopped
in flight slide down the wire and are directed
through a large funnel-shaped plastic catchment into
an escape-proof cage. Constantine later modified his
original design by varying the size and collapsibility
of the trap to conform to specific spatial needs and
logistical constraints.

Tuttle (1974) further modified the harp trap
and provided detailed construction plans. Tuttle's
trap uses two adjacent frames, 157.5 cm (62 in.)
square, spaced about 8 cm (3 in.) apart. The tautness
of the vertical strings is adjustable as is the distance
between frames. These adjustments allow for differ-
ences in conditions, and size and flight characteris-
tics of different species of bats. The strings may con-
sist of 0.20-mm (0.008-in.) stainless spring-steel wire
or 6- to 20-lb (3- to 9-kg) test monofilament nylon
line. Ideally, bats hitting the strings should not
bounce off but slide down the first bank of strings or
pass into the space between frames and then slide

512

Bats

TENSION ADJUSTMENTS

FRAMES

NYLON OR
WIRE STRANDS

Figure 4. Harp-type bat trap with major part labeled
(after Tuttle 1974).

down. The canvas holding bag is lined with plastic
that is attached at the top but hangs free at the bot-
tom. The bats slide in easily but instead of crawling
up the outer slick plastic surface, they crawl up the
cloth sides, under the plastic to a blind ending. Here,
unless large numbers are present, they usually re-
main quiet until harvested by the trapper. The width
of the opening of the holding bag is adjustable to ac-
commodate different sized species. The trap also has
adjustable legs to allow for changes in height.

The Tuttle harp trap is designed to be portable
but field assembly requires about 45 minutes. If sam-
ple sites are relatively accessible and the frames can
be transported without dismantling, assembly time
may be reduced. In fact, permanent non-collapsible
frames may be practical in most situations where the
sample sites are accessible and have limited physical
variation.

Tidemann and Woodside ( 1978) gave design
and construction details for a light-weight collapsible
version of the harp trap. By using telescoping sec-
tions of aluminum tubing, the trap becomes a rela-
tively portable package 95 cm (37.4 in.) long and 15
cm (5.9 in.) in diameter, weighing 7 kg ( 15 lb).

They maintain that the trap may be assembled in
about 5 minutes. If potential sample sites are inac-
cessible or if transport of a fully assembled trap is
not possible, this construction option should be
considered.

Operation. Typically, bat traps have much less
capture surface area than mist nets. Consequently,
they are particularly appropriate where bats are
flying in rather confined spaces such as cave en-
trances, small heavily-canopied streams, small iso-
lated pools of water, building openings, and along
narrow pathways in thickly-vegetated areas. In many
situations, the open area around the trap may be
blocked off to direct the bats into the trap. When
necessary, several traps may be used simultaneously
because the bats are retained without harm until it is
convenient for the investigator to deal with them. In
most instances it is necessary to check a trap at least
every hour to ensure that the holding bag is not
overloaded and to release bats as soon as possible so
they can resume their normal activity.

Other Techniques. Some other techniques are oc-
casionally useful for determining presence of bats.
Generally these techniques are used under more spe-
cialized circumstances than those described above
and, therefore, will not be treated in the same detail.

Bat trap set up in cave opening. Bats slide down vertical
strands into holding bag.

Bats

513

Bat trap set up over open area of small stream.

Detection of Echolocation Calls. Flying bats
produce high frequency sounds for communication,
orientation, and prey capture (see Fenton 1982).
Most such calls are above the range of unaided hu-
man detection and are referred to as "ultrasonic." In
recent years, equipment has become available that
allows detection of these ultrasonic calls in the field.
Echolocation calls, especially those used for pursuit
and capture of prey, are often recognizable as spe-
cies-specific (Fenton 1982).

The technique has several features that are use-
ful for wildlife management applications. Species
presence in a general area or specific habitats may
often be determined without actually handling the
animals or interfering with their normal activity. Un-
like mist-netting and trapping, this technique is not
restricted to sites with very specific capture condi-
tions. It may be especially useful in areas where con-
gregation sites are unknown, inaccessible, or not
conducive to sampling by conventional methods.
The technique may also be useful in conjunction
with more conventional methods. For example, an
ultrasonic detector may be a valuable tool in deter-
mining optimal habitats and time intervals for suc-
cessful deployment of mist nets (Kunz and Brock
1975). In other instances, a detector may be used to
monitor degree of activity or to aid in counting the
number of bats at a specific place after some of the
bats have been captured to determine or verify spe-
cies identity.

The technique does have some limitations. Some
bat species are not easily detected and some are not
easily distinguished from others. It cannot be used to
determine population density because calls of spe-
cific individuals cannot be identified. However, it can
sometimes be used to count the number of individu-
als passing by a restricted area in one direction, such
as a roost opening.

Ultrasonic detection equipment varies in com-
plexity, cost, and availability (Simmons et al. 1979;
Fenton in press). The cost of some sophisticated
units is beyond the reach of local management units.
In addition, equipment operation and data interpreta-
tion often require extensive training and experience.
If information that is only obtainable with this
method is deemed essential for a land management
decision, it may be desirable to contract specialists
for its collection rather than to attempt to acquire
the equipment and training at the local level. How-
ever, some inexpensive, easy-to-build, and relatively
easy-to-use systems are becoming available (e.g., see
Paige et al. 1985) and may provide some valuable in-
formation useful for management purposes.

Observation. The presence of bats in general
and a few species in particular can be determined by
direct observation. This is best done in the evening
twilight period. Bats may also be seen during morn-
ing twilight although they are not usually as numer-
ous as in the evening. It is often possible to view
bats around attractive habitat features such as roost-
ing, drinking, and foraging sites, with the aid of low-
intensity artificial red light. Enhancement of vision is
also possible through the use of a night- vision
scope, sometimes aided by an infrared light source,
but the value of this instrument is somewhat limited
because of its relatively small field of view. Such an
instrument is best used where areas of bat flight are
restricted, to determine population size of known
species with a minimum of disturbance (e.g., see
Bagley and Jacobs 1985). For further details con-
cerning observational techniques, see Barclay (in
press).

Advertising. One method of finding concentra-
tions of bats, especially roosting sites in areas fre-
quented by humans, is to advertise. Most local news-
papers will run stories about bats that include a
request for information concerning their wherea-
bouts. Some radio stations, especially public service
types, will do the same. Finally, posters may be
printed and distributed in the area of concern. Most
responses obtained from these methods are from
people wanting to know how to get rid of bats in
man-made structures. This provides the opportunity
for some conservation education in addition to locat-
ing local populations of bats.

Shooting. Because many bats emerge from
their daytime roosts during the twilight period, it is
often possible to collect them by shooting. This may
be done with a standard shotgun using shells loaded
with very small shot (No. 9 or higher). A small-gauge
light-weight shotgun with light loads is usually most
efficient for hitting erratically flying bats in dim light.
A pistol or small-bore rifle loaded with shot shells
may be used in the same manner, but the shot pat-
tern produced is smaller and less uniform than that
of a shotgun.

514

Bats

Killing an animal forecloses the option of subse-
quent release, and shooting should only be used if
absolutely necessary. If roost sites are unknown or
inaccessible and other techniques are not applicable
in an area where bats are known to occur, this op-
tion might be justified to establish presence. If possi-
ble, all specimens collected in this fashion should be
prepared as study skins or preserved in alcohol and
deposited in a regional museum for future reference.
Such specimens may be necessary for species identi-
fication or may serve as vouchers to verify the pres-
ence of species in an area.

The drawbacks of the technique are obvious.
Whereas only one or two specimens may be neces-
sary for species identification or to serve as vouch-
ers, many more may be killed needlessly while trying
to collect specific species. Shooting at twilight is
dangerous and illegal in some areas without special
permission. The technique should only be used as a
last resort when other more selective and less de-
structive methods are not applicable.

Relative Density

Relative density values of bats may be computed
from data obtained by some of the techniques de-
scribed above. It is important to realize that all such
density values are biased to some degree because of
selectivity for certain species. The degree of tech-
nique selectivity and the intended use of the relative
density values should be considered when collecting
and analyzing data. There are two primary, but not
exclusive, uses of relative density values. First, rela-
tive densities of temporally or spatially separated
populations may be compared to detect possible dif-
ferences in community composition. Second, relative
density values may be viewed as valid indicators of
bat community structure. Considering the built-in
biases and possible uses, some specific sampling
situations can be described to illustrate possible
applications.

Capture at day roosts is often highly species-se-
lective and therefore of little value for calculating
relative density that reflects overall bat community
structure. However, day roosts that are used by
more than one species may yield relative density val-
ues that are useful for evaluating changes over time
at a particular site. Site-to-site comparisons are usu-
ally not possible because different physical features
make it impossible to sample in an identical manner.
Some night roosts such as caves or tunnels may be
used by several species at once and therefore yield
data that may be used to calculate relative densities.
Again, these values are best used for temporal com-
parisons related to habitat or seasonal changes at a
given site. They are not good indicators of total bat
community structure because not all bats in an area
are likely to use the same type of night roost. They

may, however, indicate the community of species
that is utilizing a particular habitat feature.

Sampling at foraging sites is also selective and,
therefore, not a good measure of total community
structure. Sampling at communal drinking sites offers
perhaps the best opportunity to collect relative den-
sity data that reflect community structure. This is
particularly true where water sources are scarce and
the site is large enough to accommodate all species
potentially in an area. Data obtained at such sites
may also be used to calculate species diversity values
that may be useful for comparative and descriptive
purposes.

To determine species richness for a habitat or
area, one should use as many forms of capture and
detection as are available.

Absolute Density

It is very difficult, if not impossible, to deter-
mine the absolute density of bat populations. There
are two basic problems: ( 1 ) samples are taken at
points of concentration rather that at random, and
(2) it is extremely difficult to determine the total
area used by individuals in the sample. A few at-
tempts have been made to estimate density of bats
(see Gaisler 1979) but there are no standard tech-
niques similar to those used for other small mam-
mals or birds.

Whereas measurement of absolute density is not
a feasible objective, it is often possible to measure
population size at the places where bats congregate.
The values thus derived may be compared, at least
temporally, to look for trends (seasonal or annual)
or the effects of environmental alterations. Several
methods have been used to estimate the size of se-
lected bat concentrations (see Laval in press). Often,
with small colonies in roosting sites, it is possible to
simply count the bats present. This is especially ap-
plicable to hibernacula where the bats are extremely
immobile and not easily disturbed if care is taken by
the investigator. In large aggregations, where it is im-
possible to count every individual, the number of
bats in several sample areas of known size may be
counted and the estimate made by extrapolation
from the total area covered. Photographic analysis
may be used in the same manner and may also be
used to estimate population size by taking pictures
of bats emerging from day roosts (Altenbach et al.
1979; Warden 1980). The number of flying bats can
be counted in each sample picture leading to the es-
timate of bats per unit of time. If the total time of
the emergence flight is also known then population
size may be estimated.

Population estimates may also be made from
capture-recapture data. However, it appears that

Bats

515

most capture methods cause some alienation to a
particular site and chances of recapture are dimin-
ished. Nevertheless, if such estimates are obtained in
a consistent manner from place to place and time to
time, they can provide a quantitative value that is
useful for comparison. Such values are usually over-
estimates of the real population because animals are
more likely to become capture-shy rather than cap-
ture-prone. With consideration of these weaknesses,
a simple Lincoln/Peterson estimator, or some modifi-
cation, such as that of Gaisler ( 1979), will yield ade-
quate estimates.

DISCUSSION

The problems of monitoring bat populations and
their habitats are challenging. To address these prob-
lems, one must be aware of the peculiarities of bat
behavior and the limitations of the specialized tech-
niques associated with their study.

Two characteristics of many bat species make
their populations particularly susceptible to sharp
population declines. First, their tendency to congre-

gate often makes them especially vulnerable to den-
sity-independent mortality factors. These factors in-
clude both natural events and human-caused death
and disturbance. A few thoughtless or misguided
people can cause severe population reductions
either by direct killing, vandalizing roosts, or disturb-
ing the bats during critical periods. Bats disturbed
during hibernation may use essential energy reserves
during the arousal process and those disturbed at
maternity sites may lose contact with their young.

Second, bats have a very low reproductive rate;
the females of most species only produce one young
per year. This low reproductive potential makes re-
covery from sudden increases in mortality relatively
slow. Reduction of populations of North American
bats during recent years has been documented for
several species (Harvey 1976).

Little doubt exists that bats fill important niches
and the impact of their insectivorous habits in tem-
perate ecosystems should not be underestimated. It
is important to acknowledge this vital role and at-
tempt to inventory and monitor bat habitat and pop-
ulations to facilitate their effective management and
conservation.

516

Bats

LITERATURE CITED

ALTENBACH, J.S., KN. GELUSO, and D.E. WILSON. 1979.
Population size of Tadarida brasiliensis at Carlsbad
Caverns in 1973- Pages 341-348 in Genoways, H.H.
and Baker, R.J. eds. Biological Investigation in the
Guadalupe Mountains National Park, Texas. Natl. Park
Serv. Proc. and Trans. Ser. 4.

BAGLEY, F. and J. JACOBS. 1985. Census technique for
endangered big-eared bats proving successful. Endang.
Species Tech. Bull. 3:5-7.

BARBOUR, R.W. and W.H. DAVIS. 1969. Bats of America.
University Press of Kentucky, Lexington. 286pp.

BARCLAY, R.M.R. (In press). Marking and observational
techniques, in Kunz, T.H. ed. Behavioral and Ecologi-
cal Methods for the Study of Bats. Smithsonian Institu-
tion Press. Washington, DC.

BLEITZ, D. 1984. Mist nets and their use. Bleitz Wildlife
Foundation, 5334 Hollywood Blvd., Hollywood, CA.
18pp.

BONACCORSO, F.J. and N. SMYTHE. 1972. Punch-mark
ing: an alternative to banding. J. Mammal. 53:389-390.

, , and S.R. HUMPHREY. 1976. Improved

techniques for marking bats. J. Mammal. 57:181-182.

COCKRUM, EL. 1969. Migration in the guano bat, Tadar-
ida brasiliensis. Pages 303-336 in Jones, J.K, Jr. ed.
Contributions in Mammalogy. Misc. Publ. 51. Univ.
Kansas, Lawrence

CONSTANTINE, D.G. 1958. An automatic bat-collecting
device. J. Wildl. Manage. 22:17-22.

CROSS, S.P. 1965. Roosting habits of Pipistrellus hesperus.
J. Mammal. 46:270-279.

DAVIS, R.B., C.F. HERREID II, and H.L. SHORT. 1962. Mex-
ican free-tailed bats in Texas. Ecol. Monogr. 32:311-
346.

DAVIS, R.P. and EL. COCKRUM. 1963. Bridges utilized as
day roosts by bats. J. Mammal. 44:428-430.

DOWLER, R.C and H.H. GENOWAYS. 1976. Museology:
supplies and suppliers for vertebrate collections.
Texas Tech Press. Lubbock. 83pp

FENTON, MB. 1982. Echolocation, insect hearing, and
feeding ecology of insectivorous bats. Pages 261-28 in
Kunz, T.H. ed. Ecology of Bats. Plenum Publ. Corp.
New York, NY.

. (In press). Detecting, recording, and analyzing the

vocalizations of bats, in Kunz, T.H. ed. Behavioral and
Ecological Methods for the Study of Bats. Smithsonian
Institution Press. Washington, DC.

GAISLER,J. 1979. Ecology of bats. Pages 281-342 in Stod-
dard, D.M. ed. Ecology of Small Mammals. Chapman
and Hall, London.

GELUSO, K.N, J.S. ALTENBACH, and D.E. WILSON. 1976.
Bat mortality: pesticide poisoning and migratory
stress. Science 194:184-186.

GREENHALL, A.M. 1982. House bat management. U.S.
Dep. Inter., Fish and Wildl. Serv. Resour. Publ. 143-
53pp.

and J.L. PARADISO. 1968. Bats and bat banding.

U.S. Dep. Inter., Bur. Sport Fish, and Wildl. 72. 48pp.

HALL, E.R. 1981. The mammals of North America. 2nd Ed.
John Wiley & Sons. New York, NY. 1 181pp.

HARVEY, M.J. 1976. Endangered Chiroptera of the south-
eastern United States. Proc. Annu. Conf. Southeast. As-
soc. Game Fish Comm. 29:429-433-

HAYWARD, B.J. and S.P. CROSS. 1979. The natural history of
Pipistrellus hesperus ( Chiroptera: Vespertilionidae ). Of-
fice Res. West. New Mexico Univ., Silver City. 3: 1-36.

HENNY, C.J., C. MASER, JO. WHITAKER, and T.E. KAISER.
1982. Organochlorine residues in bats after forest
spraying with DDT. Northw. Sci. 56:329-337.

JONES, J.K Jr., DC. CARTER, and H.H. GENPWAUS. 1982.
Revised checklist of North American mammals north
of Mexico. Occ. Pap. 62. Mus. Texas Tech. Univ., Lub-
bock.

KEYES, BE. and C.E. GRUE. 1982. Capturing birds with
mist nets: a review. N. Am. Bird Bander. 6:1-14.

KUNZ, T.H. 1982. Roosting ecology. Pages 1-46 in Kunz,
T.H. ed. Ecology of Bats. Plenum Publ. Corp. New
York, NY.

, ed. (In press). Behavioral and ecological methods

for the study of bats. Smithsonian Institution Press.
Washington, DC.

and C.E. BROCK 1975. A comparison of mist nets

and ultrasonic detectors for monitoring flight activity
of bats. J. Mammal. 56:907-911.

LAVAL, R.K (In press). Census techniques, in Kunz, T.H.,
ed. Behavioral and Ecological Methods for the Study
of Bats. Smithsonian Institution Press. Washington, DC.

MASER, C.J.E. RODIEK and J.W. THOMAS. 1979. Cliffs,
talus, and caves. Pages 96-103 in Thomas, J.W. tech.
ed. Wildlife Habitats in Managed Forests. U.S. Dep.
Agric. For. Serv. Agric. Handbook 553-

, BR. MATE, J.F. FRANKLIN, and C.T. DYRNESS.

1981. Natural history of Oregon coast mammals. U.S.
Dep. Agric, For. Serv. Gen. Tech. Rep. PNW-133-
496pp.

MENAB, B.K 1982. Evolutionary alternatives in the physi-
ological ecology of bats. Pages 151-200 in Kunz, T.H.
ed. Ecology of Bats. Plenum Publ. Corp. New York,
NY.

O'SHEA, T.J. and T.A. VAUGHAN. 1977. Nocturnal and sea-
sonal activities of the pallid bat, Antrozous pallidas. J.
Mammal. 58:269-284.

PAIGE, K.N., L.A. MINK and V.R. McDANIEL. 1985. A

broadband ultrasonic field detector for monitoring bat
cries. J. Wildl. Manage. 49:11-13.

SIMMONS, J.A., MB. FENTON, W.R. FERGUSON, M. JUT-
TING, and J. PALIN. 1979. Apparatus for research on
animal ultrasonic signals. Life Sci. Misc. Publ. Royal
Ontario Mus. 31pp.

STEBBINGS, R.E. 1978. Marking bats. Pages 81-94 in
Stonehouse, R. ed. Animal Marking. Univ. Park Press,
Baltimore, MD.

THOMAS, J.W. tech ed. 1979. Wildlife habitats in managed
forests: the Blue Mountains of Oregon and Washing-
ton. U.S. Dep. Agric, For. Serv. Agric. Handbook 553-
512pp.

TIDEMANN, C.R. and DP. WOODSIDE. 1978. A collapsi-
ble bat-trap and comparison of results obtained with
the trap and with mist nets. Australian Wildl. Res.
5:355-362.

TUTTLE, M.D. 1974. An improved trap for bats. J. Mam-
mal. 55:475-477.

VAUGHAN, T.A. 1959. Functional morphology of three
bats: Eumops, Myotis, Macrotus. Publ. Mus. Nat. Hist.
Univ. Kansas, Lawrence. 12:1-153

WARDEN, T. 1980. Notes on the determination of bat
populations using photographic measurements. Nat.
Speol. Soc Bulletin. 42:70-71.

Bats

517

25

Ungulates

Raymond J. Boyd and Allen Y. Cooperrider

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

Peter C. Lent

U.S. Bureau of Land Management
920 Valley Road
Reno, NV 89512

James A. Bailey

Colorado State University
Fort Collins, CO 80523

Editor's Note: Ungulates occupy a great diversity of
habitats in North America and have life-history
strategies uniquely adapted to these habitats. Yet,
many inventory and monitoring techniques are
similar for each of the species. Ungulates are ex-
tremely important because of their recreational
hunting and viewing values; they also can compete
with livestock for forage and cause damage to var-
ious agricultural crops.

Many common ungulates in North America, such
as elk and antelope, were once quite rare and extir-
pated from many parts of their historical range.
This suggests that given adequate public support,
cooperation, and funding ungulates can be main-
tained or increased in natural habitats with ade-
quate food, cover, and water. Intensive efforts in
studying inventorying and monitoring such habi-
tat factors have been largely responsible for the
successful increase in North American ungulate
numbers. These efforts have been detailed in much
literature over the past 50 years. This chapter, there-
fore, only introduces and provides an overview of
this knowledge base. A biologist working with un-
gulates will need to review the sources listed in the
chapter for more detailed information.

"All animals are created equal, but some are created
more equal than others."

INTRODUCTION

-George Orwell, Animal Farm

Wild ungulates are among the most prized wild-
life groups, and management of their habitat is an
important task of wildlife biologists. Most areas of
North America contain habitat or potential habitat
for one or more ungulate species. These animals are
highly valued for aesthetic reasons, and public hunt-
ing of wild ungulates is an important cultural heri-
tage. Recreational hunting is an important source of
revenue for local economies and for manufacturers
of hunting-related equipment. License sales for these
animals provide a primary source of funding for state
wildlife agencies.

On the other hand, wild ungulates can cause
obvious, short-term habitat alterations through graz-
ing and browsing. They can damage agricultural
crops, directly compete for forage with livestock, as
well as degrade their own habitat.

Inventory and monitoring of habitat for wild
ungulates is thus one of the most important tasks of
wildlife biologists; it is also difficult, challenging, and
interesting work. Although biologists have studied
and attempted to manage wild ungulate habitat more
intensively than most other species habitats, our
understanding and ability to manage it is still
primitive.

In particular, biologists often have difficulty
predicting how ungulate populations will respond to
habitat changes. Because ungulates are adaptable, a

Ungulates

519

Haystack damage by elk.

species such as white-tailed deer (Odocoileus virgi-
nianus) may live in habitats ranging from tropical
Florida swamps to hot Arizona deserts, or northern
coniferous forests of British Columbia. Furthermore,
since the Pleistocene epoch, the number of ungulate
species in North America has been quite limited,
consisting of "generalists" (as opposed to some of
the highly specialized forms found in Africa). In
adapting to different habitats, ungulate populations
acquire distinct patterns of forage selection, habitat
use, and population dynamics. Therefore, generaliz-
ing about the species from study of an individual
population is not only risky, but may increase or per-
petuate misinformation.

Ungulates are highly intelligent animals. For
example, choice of migration routes for species such
as bighorn sheep results not from random selection
among appropriate alternative routes, but rather
selection learned from travel with other individuals
during previous migrations. Such learned behaviors
greatly affect how the animals use habitat and re-
spond to habitat changes. Such use and response to
management cannot always be determined by look-
ing at the habitat.

Because ungulates are relatively long-lived, re-
sponses to habitat changes may not be apparent for
many years. Such time lags make assessing the effects
of management difficult unless careful, well-planned
records are kept for many years.

Finally, habitat suitability often is not the sole
factor limiting ungulate populations. Predation, dis-
ease, or hunting may limit a population. Each of
these factors can, of course, be related to habitat
conditions or interact with habitat factors. However,
there are well-documented cases where changes in
predation, disease, or hunting, in the absence of any
change in habitat, have caused significant changes
in ungulate populations.

Because ungulates have been studied so exten-
sively, this chapter only provides a broad overview.
It is intended for biologists having limited experi-
ence with ungulates and for those wanting guidance
on finding additional information in the literature.
In the following sections we describe general habitat
requirements of ungulates and methods for measur-
ing population variables. We then describe specific
requirements and techniques for individual North
American species. These accounts are necessarily
brief, but should assist the reader in finding more de-
tailed information in the literature.

HABITAT REQUIREMENTS

Habitat requirements of ungulates are analyzed
in terms of their needs for food, water, cover, and
reproduction as well as the way they must be ar-
ranged in space and time.

Food

Forage supply often limits ungulate populations,
either by itself or in conjunction with some other
factor such as cover. Food is almost always a qualita-
tive rather than a quantitative problem with ungu-
lates. By this we mean that the quality of available
food rather than the total amount of food limits a
population. Ungulates have died of malnutrition with
their stomachs full of food. Thus, the habitat biolo-
gist must consider not only the amount of food avail-
able, but its nutritional content. Furthermore, forage
availability and nutritional quality of available plants
vary seasonally due to phenology of plants, utiliza-
tion by other wild and domestic animals, and
weather conditions that may make plants unavailable
(e.g., deep snow) or alter phenology or nutritional
quality (Figures 1 and 2). Thus, analyzing the ade-
quacy of the forage supply for an ungulate is com-
plex. Biologists have tried to simplify the problem by
identifying measurable limiting factors among the
vast number of forage-related factors. This approach
has been quite successful in local areas, but unsuc-
cessful when knowledge of limiting factors from one
region has been blindly used as the basis for evaluat-
ing forage supply in another. As background for un-
derstanding methods of analyzing forage supply for
ungulates, we describe the annual forage cycle, pat-
terns of forage preference and food habits, potential
limiting factors, and mineral licks.

Annual Forage Cycle. Since ungulates may eat
many plant species in different regions and different
seasons, biologists need to classify forage species
according to similar nutritional and/or phenological
characteristics. Wildlife biologists and range man-
agers have found it convenient to classify forage as
grasses, forbs, and browse (Table 1 ). This classifica-
tion is based primarily on structure and phenology
of the plants, but also correlates well with seasonal

520

Ungulates

840-
5 800-

a.
O
g 600-

o

z

o

z

< 400-

co

CO

<

d 200-
<
>
<

/

TOTAL ANNUAL PRODUCTION

NUTRITION QUALITY (K cal DE/g)

O O (S3 O

/ N< NUTRITIONAL VALUES

VSE

1

1 FORBS

2 GRASS
& BRO\

/ N. NECESSARY FOR
/ X. MAINTENANCE

Mm

Ji

v M J JASONDJ F

M A

\ M J J ASODXJ FMA

Figure 1. Typical annual cycle of available standing
crop of forage (from Cooperrider and Bailey
1984).

Table 1. Classification of ungulate forages.

Figure 2. Typical annual cycle of average nutri-
tional value of forage on temperate ranges
(modified from Cooperrider and Bailey 1984).

Class

Description

Taxonomy

Notes

Grass and grass-like
plants

True grass and grass-
like sedges and rushes

Family Gramineae and
grass-like Cyperaceae
and Juncaceae

This category often is
abbreviated to grass,
but still includes sedges
and rushes.

Forbs

All herbaceous plants
other than grass and
grass-like plants

Most monocots other
than three families listed
above; herbaceous
dicots

Since classification of
forbs and browse is not
based on taxonomy,
classification of some
semi-woody species is
somewhat arbitrary.

Browse

All woody plants

Woody dicots

nutritional qualities, making it a basis for describing
such an annual cycle. The nutritional quality of a
given plant species varies with parts of a plant, indi-
vidual plants, ecotypes, subspecies, regions, and
physical characteristics of the site.

A general picture of the forage cycle can be
provided by considering a typical pattern of produc-
tion, availability, and forage quality of grasses, forbs,
and browse in temperate and arctic areas. Forbs
are mostly available for a short time in the spring
and summer (Figure 1); during this time they are
generally highly nutritious as measured by digestible
energy content or similar measures (Figure 2).
Grasses, on the other hand, are available year-round
but are most nutritious during the spring and early
summer when they are green. They also often begin
growing earlier than any other forages (Figures 1
and 3). Dry dormant grass, on the other hand, has
limited nutritional value. Browse, by contrast, is
available year-round with relatively similar nutri-
tional value among seasons (Figure 3) Significant
variations may occur because of species differences,

annual versus perennial grasses, local weather pat-
terns, deciduous versus evergreen browse species,
etc. However, understanding this basic model along
with local variations should help in understanding
food habits and forage preferences of a specific ungu-
late herd and thus in identifying potential limiting
factors related to forage.

Forage Preference and Food Habits. All North
American ungulates are generalist herbivores, choos-
ing opportunistically among hundreds of local plant
species. A typical ungulate will eat over 100 plant
species during the course of a year. However, due to
availability and preference, less than a dozen plants
usually constitute over 75% of the animal's seasonal
diet. These are the species that should be the initial
concern of management.

Forage cycles can be used to help interpret food
habits of the various ungulate species. Beginning in
spring, ungulates will turn to plants that are actively
growing. For instance, the food habits of pronghorn

Ungulates

521

3.0-]

90-

vs

___--■"

_,

^ FORBS

\

S*~

O)

80-

LU

/ ** "" "

[" \

h-

\

/

Q

/ •

V N

UJ

o

*: 2.0-

>
i-

/ / ,'-

T N ~~ -- -.

Q

Z
111

o
<

DC

70-
60-
50-

\ BROWSE

y

/
/

/

/

/ -';

7'" i

1 i

i

X

BROWSE

<

i

GRASS

o

"■--

——■*

O

/

LL

o

/

LL

40-

_l

/

o

< 1.0-

1

H

z

/

Z

30-

o

/

O

V FORBS

cr

//

LU

20-

O

//

Q.

/ GRASS

z

r /

10-

0

s /

■<„

> i i

— i i 1^ i

A S 0 N

D J F

A

A M J J

■ i i

ii ii ii

A M

J J A

S O N D J F

M

A

Figure 3. Typical annual cycle of nutritional qual-
ity of grass, forbs, and browse.

shown in Figure 4 can be, to a large extent, ex-
plained by the pattern of forage production availabil-
ity and nutritional value shown in Figures 1 and 3.
Typically, the actively growing plant species are first
grasses, then forbs. As the season progresses and
forbs become less nutritious or available, ungulates
turn more to browse or grass, depending on the
species and its habitat and forage preference. During
winter, most ungulates depend on dry grass or
browse to maintain themselves. Ungulates reduce
their forage intake during this period and typically
lose weight over winter.

Although North American ungulates are oppor-
tunistic, significant differences in forage preferences
are presumably a reflection of their ability to gather
and digest different forage classes. Most notable of
these differences is the distinction between grass
eaters and browse eaters. All North American ungu-
lates eat grass or browse when it is green, succulent,
and nutritious. However, when grass or browse is
dry or dormant, some ungulates prefer one over the
other. In general, grass has poorer nutritional quality
at this time, but is usually available in greater con-
centrations, favoring animals that can gather greater
quantities in a short period of time. Browse, on the
other hand, is usually more nutritious, but an animal
needs to be more selective and therefore take more
time to gather the same quantity. In general, the
smaller ungulates (pronghorn antelope [Antilocapra
americana], mule [Odocoileus hemionns] and
white-tailed deer [O. virginianus\) tend toward
browse at this time, while the larger species (bison
[Bison bison] and muskox [Ovibos moschatus])
prefer grasses.

Habitat preferences and evolutionary history
also play a role. The North American bovids (big-
horn [Ovis canadensis] and Dall's sheep [O. dalli],

Figure 4. Food habits of pronghorn antelope on
Trickle Mountain, Colorado (from Bailey and
Cooperrider 1982).

muskox, bison, and mountain goat [Oreamnos amer-
icana]) are generally animals of open country (grass-
lands) and are capable of surviving on dry grass in
winter. On the other hand, North American cervids
(mule and white-tailed deer, elk [Cervus elaphus],
moose [Alces alces]) tend to be forest or forest-edge
animals or have evolved from forest-dwelling ances-
tors; all can survive on dormant browse. Caribou
(Rangifer tarandus), an apparent exception, survive
on lichens during winter over much of their range.
Caribou probably evolved as forest animals, eating
arboreal lichens but some learned to move out into
the tundra to eat lichens on the ground. Some North
American ungulates seem capable of surviving either
way, most notably elk; although less well-studied,
mountain goats appear to have the same capability.

Food habits are foods that an animal actually
eats, based on forage availability and preference.
Measurement or observation of food habits can be an
important tool toward understanding limiting factors
related to forage quality and quantity. If observed
food habits differ significantly from typical or ex-
pected food habits, it may suggest a problem or lim-
iting factor. For example, if deer are eating
significant quantities of dry grass during winter (an
atypical pattern), it could suggest that their winter
range lacks suitable browse species. This could be
confirmed by observing or measuring plant species
composition on that range. Similarly, if antelope are
eating small quantities of forbs in spring, it could
suggest that the range is lacking good quality forbs.

Food habits must always be interpreted in terms
of local conditions and the ungulate species under
consideration. For example, the food habit pattern
for mule deer shown in Figure 5 would appear, at
first glance, to be atypical of the species as a whole.
However, it can easily be explained by recognizing

522

Ungulates

that these deer live in California in a Mediterranean
climate where winters are mild and all precipitation
comes in fall, winter, and spring. Therefore, grass
is green and nutritious during winter but dry and un-
palatable during summer.

Analysis of Limiting Forage Factors. Several
methods have been developed to analyze limiting
factors related to forage supply. In order of increas-
ing complexity, these are the —

( 1 ) Foraging area approach,

(2) Key species approach,

(3) Carrying-capacity approach, and

(4) Nutritional approach.

Items 2 and 3 have been used in one form or
another for over 50 years in North America. The
foraging area approach, however, has only been for-
malized in recent years in response to land manage-
ment needs. Similarly, the nutritional approach has
only recently been developed and formalized, al-
though earlier attempts were made to incorporate
nutritional considerations into analysis methods.

Shrubland Diet

GRASS FORBS

SHRUBS

26%

19%

55%

Figure 5. Food habits of mule deer in California
chaparral (from Taber and Dasmann 1958).

Foraging Area Approach. The foraging area
approach is a landscape-based analysis system. With
this approach, the biologist identifies the aspects
of topography, vegetation structure, and other physi-
cal features required to provide suitable foraging
areas for ungulate species. For example, suitable win-
ter foraging areas for bighorn sheep in northern Col-
orado might be considered to be south-facing, grass-
covered slopes below 2,400 m ( 7,920 ft ) elevation,
within 400 m (1,320 ft) of suitable rocky escape
cover. The biologist then analyzes the range to de-
termine if it contains an adequate acreage or distri-
bution of foraging areas. This analysis does not result
in a quantitative estimation of carrying capacity.
The foraging area approach was developed for elk
and mule deer in the Blue Mountains of eastern Ore-
gon by Black et al. ( 1976) and Thomas et al. ( 1979).
The technique has the advantage of requiring only
limited data collection, with data frequently obtained
from existing sources such as aerial photographs,
topographic maps, etc. The technique is limited in
that forage quality is not considered in terms of
plant species composition, range condition, total
forage weight, browse species condition, or nutri-
tional value. However, it is a good way to begin an
analysis of ungulate range and may serve as the basis
for more detailed analyses if these are required later.

Key Species Approach. The key species ap-
proach is based on the assumption that a biologist
can identify the critical (limiting) range of the ani-
mal and also the critical or key forage species on
that range. It was developed primarily for mule and
white-tailed deer in northern areas of the U.S. and
Canada. In winter, these deer typically concentrate
on traditional winter ranges and subsist on one or
more species of palatable browse. The browse spe-
cies on these ranges are assumed to be the factors
limiting these herds. By concentrating on key
browse species on key areas, the biologist can sim-
plify a complex situation. The condition, trend, and
production of the key species can be measured and
monitored; if total production is estimated, it can
lead to a carrying-capacity determination.

The utility of this technique depends entirely on
two factors: ( 1 ) the degree to which ungulates de-
pend on key species and range and ( 2 ) the biolo-
gists' ability to correctly identify these key species
and ranges. Recent research has demonstrated the
importance of all seasonal ranges in determining the
health of deer and other ungulate herds (Mautz
1978). For many years, biologists concentrated on
winter ranges because this was when deer and other
ungulates lost weight; furthermore, losses from star-
vation typically were observed during late winter
and early spring while deer were still on winter
ranges. However, recent field studies and nutritional
research have demonstrated that the condition of
ungulates before winter may be as important or
more important than the condition of the winter

Ungulates

523

food supply in determining overwinter survival.
Therefore, biologists have begun to pay more atten-
tion to spring, summer, and fall range conditions.

The key species approach has provided and will
continue to provide a useful way of analyzing ungu-
late forage supplies. However, blind adherence to
this approach has caused much misinformation and
inefficient use of limited habitat management funds.
In extreme cases, dogmatic belief in key forages as
limiting factors has caused biologists to overlook
limiting factors such as predation, hunting, or dis-
ease. Biologists should use the key species approach
with caution.

Carrying-Capacity Approach. When discuss-
ing carrying capacity, biologists tend to bog down in
semantics; for the purpose of this discussion, carry-
ing capacity is the number of animals that can be
supported on a range with a given amount of forage
production without adversely affecting the long-term
productivity of the forage supply. For a more de-
tailed discussion of the concept, its use and misuse,
see Bailey (1984:280-288) or Caughley ( 1 979 ).

To use the carrying-capacity approach, a biolo-
gist must measure or estimate the total amount of
forage available on a year-round or seasonal range,
determine the amount of forage that can be used
without damaging future productivity, and then cal-
culate the number of animals this amount of usable
forage can sustain. Although seemingly simple, the
approach is difficult at each stage. Measuring total
forage is expensive and complex in itself. Further-
more, determining the amount of forage that can be
used without damage is more of an art than a sci-
ence (Skiles et al. 1980). Finally, calculating the
number of animals that can be sustained requires
making assumptions or estimates on the quality or
condition of the animals, sex and age ratios, etc.
Furthermore, it may require information on intake
rates or food habits that are not available or can only
be obtained at great expense.

To alleviate these problems, biologists tend to
make simplifying assumptions and use "rules of
thumb" in making the calculations. For example, the
50% or "graze half/leave half" rule developed by
range managers is commonly used to determine the
percentage of total forage that can be consumed.
Similarly, many agencies have developed tables of
forage intake rates for both domestic and wild
ungulates.

Calculating carrying capacity is thus a very
crude measure of the capability of the habitat to
support ungulates on a sustained basis. Nonetheless,
it has been useful and will continue to be used. It
is especially useful where more than one wild or
domestic ungulate is found on the same range. In
particular, when livestock are grazing areas used by

Winter-killed mule deer under overbrowsed juniper.

wild ungulates, land managers must determine the
amount of forage to be allocated for livestock and
the amount to be reserved for wild animals. Forage
allocation requires a carrying-capacity approach, and
several systems have been developed for formalizing
these calculations ( MacPherson et al. 1982; Cooper-
rider and Bailey 1984; Nelson 1984).

Calculating carrying capacity, therefore, requires
a great deal of information which may be costly to
obtain, and the final measure may only be a very
crude determination of the ability of the habitat to
provide forage for the animals. Furthermore, most
carrying-capacity approaches do not consider infor-
mation on the nutritional requirements of animals
nor the nutritional content of plants.

Nutritional Approach. The nutritional ap-
proach attempts to consider both the quality and
quantity of forage in a carrying-capacity model.
When considering the nutritional value of forage,
many factors can be measured. For many years, wild-
life biologists used proximate analyses (Table 1 ) to
evaluate forages. These analyses were developed
by livestock nutritionists and proved to have some
value in evaluating wildlife forages. Crude protein
was found to be especially useful since it was easy
and cheap to measure, and levels of crude protein in
most plants seemed to correlate with their palatabil-
ity to wild ungulates. There were some notable ex-
ceptions, however, including arid land browse
species plants such as big sagebrush {Artemisia tri-
dentata) and conifers such as Douglas fir (Pseudot-
suga menziesii). These types of plants typically
contained high levels of crude protein but were fre-
quently far less palatable or digestible than other
plants with similar levels of crude protein. Under-
standing wildlife nutrition was greatly enhanced by
the discovery of essential oils in plants, which inhib-

524

Ungulates

ited ruminant digestion and made the plants unpalat-
able (Nagy et al. 1964; Longhurst et al. 1968).
Although some plants can have relatively high levels
of certain key nutrients, these nutrients are unavail-
able to the animal because of the presence of other
non-nutritive substances.

This discovery and other developments stimu-
lated the search for new measures of nutritional
quality. Two such measures are commonly used
now: digestible protein (DP) and digestible energy
(DE). Both effectively measure the overall quality of
a forage for ungulates, although they can be expen-
sive to obtain. Furthermore, these measures tend
to be highly correlated with each other and with
certain other measurements such as concentration of
certain vitamins. Since the basic requirements for
maintenance, reproduction, lactation, etc. have been
developed for animals in terms of digestible energy,
animal requirements can be related directly to
intake.

Nutritional approaches to carrying capacity re-
quire biologists to limit analyses to a very few possi-
ble nutrients or nutritional measurement. Everything
from trace minerals to vitamins has been suggested
as factors limiting ungulate populations, and such
substances may be important in local situations.
However, the biologist is well-advised to begin a
nutritional investigation or analysis by measuring the
more obvious and basic parameters such as crude
protein, digestible protein, or digestible energy.

Systems for integrating nutritional measurements
such as DP and DE into carrying-capacity determina-
tions have been developed and described by Mautz
(1978), Robbins (1973), Wallmo et al. (1977),
Hobbs et al. (1982), Cooperrider and Bailey (1984),
and others.

Nutritional approaches to determining carrying
capacity require much sophisticated information
on nutritional requirements of animals and nutrient
content of forages. In addition, much data are re-
quired for traditional carrying-capacity models. Se-
lecting appropriate measurement s) of nutritional
quality is critical both in terms of costs and predic-
tive power; nutritional approaches should only be
used when ungulate populations and management is-
sues warrant the high cost in money and personnel.
Not surprisingly, we do not know of any situations
where a nutritional approach is being used in an
operational mode by land- or wildlife-management
agencies.

Mineral Licks. On some ranges, ungulates are at-
tracted seasonally to mineral licks. While these areas
may certainly affect animal distribution, the effect
on herd productivity has not yet been demonstrated.
Mountain sheep, mountain goats, and moose may
be an exception as discussed later in this chapter, al-
though the effects have only been hypothesized.

Biologists once provided salt on ungulate ranges as
livestock operators do on cattle ranges. This prac-
tice, however, does not seem to affect animal pro-
ductivity, survival, or abundance.

Water

Free Water. Although some desert species such as
bighorn sheep can survive for extended periods
without free water or with free water in plants, all
North American ungulates need to have water
sources in their habitat for survival. The biologist
must determine the spacing or density of water
sources: what is minimal for a given species and
what is optimal. In most northern areas of the conti-
nent, water is abundant and relatively well-distrib-
uted. As a result, lack of water as a limiting factor
has rarely been suggested or studied. In the south-
western desert areas, however, lack of water has
been identified as an important factor limiting bigh-
orn sheep (Turner and Weaver 1980), pronghorn
antelope (Yoakum 1980), and mule deer (Wood et
al. 1970). For desert bighorn, minimal spacing of
water sources has been estimated at 10 km (6 mi.)
and optimal spacing as 2 km (1.2 mi.) (Turner and
Weaver 1980). Optimal spacing of water has been
estimated at 4 to 5 km (2.5 to 3 mi.) for mule deer
(Wood et al. 1970) and 5 to 7 km (3 to 4 mi.) for
pronghorn (Yoakum 1978). Without further study, a
biologist may conclude that distribution of free-
water sources (streams, lakes, or perennial springs)
every 2 to 5 km (1 to 3 mi.) is an optimal free-water
supply for ungulates in North America.

Water Quality. Water quality requirements for wild
ungulates have seldom been investigated. However,
biologists have documented that waters high in total
dissolved solids (TDS) or with a high pH will not
be used by some species. Sundstrom ( 1 968 ) noted
that pronghorn antelope avoid waters with TDS in
excess of 5,000 mg/1 or pH higher than 9.2. In gen-
eral, biologists need not worry about water quality
for ungulates unless there is evidence that some
waters are being avoided. In these cases, water qual-
ity measurements should be considered, particularly
if water quality can potentially be improved through
management actions.

Snow. The presence or absence of snow packs is
such an obvious determinant of habitat suitability for
ungulates in the northern areas of this continent
that it is frequently omitted from elementary discus-
sions on the subject. However, more than one nov-
ice biologist has seriously misjudged habitat
suitability by examining a range during summer
without understanding conditions of the range in
midwinter. More seriously, some agencies have for-
malized procedures that calculate earning capacities
based on forage production without considering
snow depths on the range. By using such systems,

Ungulates

525

forage may be considered available to wild ungulates
even though it is under 10 feet of snow during the
time of year they need it.

On the basis of much observation and measure-
ments, snow depth has two important effects on
ungulates: it covers up and eventually makes ground
forage unavailable or unlocatable, and it reduces
mobility. However, these effects are quite species-
specific and, even then, conventional wisdom or
"rules of thumb" may be misleading.

Most ungulates, even those that are not particu-
larly adapted to living in snowy areas, will paw
through a light snow cover of up to 10 cm (3 in. )
and forage without difficulty. Snow covers over 10
cm (3 in.) may inhibit foraging on the ground for
some species, although others such as caribou can
locate by smell and paw for food through depths of
approximately 15 to 18 cm (6 to 7 in.) (Bergerud
1978). Elk have been reported to shift from herba-
ceous forage to browse when depths exceeded 61
cm (24 in.) (Skovlin 1982).

Effects of snow on mobility are partly caused by
depth but also by snow density and crusting. As a
general rule, most ungulates can travel through snow
that reaches up to their bellies without great diffi-
culty. Some species, such as caribou, have wide
hooves that help them walk on snow rather than
through it. Kelsall (1969) suggested that snow
depths from 36 to 43 cm (14 to 17 in.) would re-
strict white-tailed deer mobility. Similarly, snow
depths greater than 76 cm (30 in.) seriously curtail
movement of elk (Skovlin 1982). Sweeney and Stein-
hoff (1976) suggested that depths over 71 cm (28
in.) usually prohibited use by elk.

Many ungulates move to areas with lighter snow
cover during winter. This may require long migra-
tions, movement to lower altitudes, or more limited
shifts in use of particular cover types or areas with
special topographic features. In mountainous areas,
south-facing slopes generally have lower snow
depths, and many ungulates such as mule deer and
elk will concentrate there. However, some ungulates
such as mountain goats use windswept ridges at
high elevations.

Snow cover can easily be measured. However,
obtaining enough samples throughout a range during
winter would be exceedingly difficult and of ques-
tionable value unless conditions were uniform
throughout the range from year to year. Mountain-
ous areas that are likely to remain relatively snow-
free can be predicted to a limited extent from
knowledge of elevation, slope, and aspect. Such areas
can be delineated simply and easily through direct
observation, or through direct observation combined
with use of low-level aerial photographs taken at
appropriate times during winter.

Elk tracks and pawing craters in snow.

Cover

Cover, for purposes of this discussion, is defined
as any structural feature of the environment that is
used for protection from the environment (e.g., ther-
mal cover) or from predators (e.g., security cover).
As Bailey (1984:1 10) pointed out, cover has con-
noted vegetation cover in the past. However, for an
ungulate that relies on sight to detect predators and
on open country for escape, absence of vegetation
may serve the same function as presence of vegeta-
tion does for a hiding species. For example, bighorn
sheep rely on sight to detect predators and avoid
areas of dense vegetation; escape cover for bighorn
sheep consists of steep, open, rugged areas (Bailey
1984:112).

For many years, cover for wildlife species was
described in terms of the habitat they favored, with-
out distinguishing the many biological functions such
cover was providing. Biologists have begun to iden-
tify the purposes that a given type of cover serves.
Black et al. (1976) distinguished between thermal
and hiding cover; the former is defined as overstories
that give protection from the weather and sun and
the latter, as vegetation used for escape and protec-
tion from predators and humans. This is a useful
distinction, but will be used here to include all
structural features, rather than just vegetation, and
will be termed shelter or security cover. In some
cases, both functions may be served by the same
physical structure, but frequently this is not the case.
Shelter and security, however, are the two most
commonly recognized year-round cover needs. As
with many other wildlife species, cover requirements
for reproduction (birthing and early rearing) of un-
gulates are quite specific and will be discussed sepa-
rately. Some species also have special cover
requirements and these are mentioned.

526

Ungulates

# • w 1

" A

(^ ' *

? ■#*

^^^i

ik^feF*

i« m

'&&>*•&

T%^

■' / .

--

vs \. JH|4i

JBP'^s

'•

■&> - w» '. m ^ h v ^ *

^ ■kill

t w ^? win?

MftiCi (£>y J

Escape cover for bighorn sheep.

Shelter. Ungulates use shelter to escape the ex-
treme conditions of weather, particularly extreme
heat or cold or rain, hail, or snow. Most forest or
forest-edge species use tree or brush overstories for
this purpose. During winter, conifer or other ever-
green species are obviously most effective. Open
country animals such as bighorn sheep, pronghorn
antelope, and bison may appear at first to survive
without using shelter cover. However, many field bi-
ologists have noted that these species will use minor
topographic features as shelter from weather. On
cold, windy, winter days, pronghorn antelope will
frequently be found in swales where wind speed and
chill factor are substantially less. Similarly, bighorn
sheep will often take advantage of overhanging cliffs
to provide thermal cover in winter and shade in
summer; desert bighorn have even been observed
using caves during the summer. To our knowledge,
no system for quantitatively identifying such topo-
graphic cover features has been devised or pub-
lished. There are, however, many guidelines on what
constitutes suitable shelter or thermal cover in terms
of vegetation by animal species. For example, Wit-
mer et al. ( 1985) define thermal cover for elk as
a "stand of evergreen coniferous trees over 40 ft tall
with crown closure over 70% with an optimum
size of 30 to 50 acres."

Security. Security cover is used for protection
from predators or humans. North American ungu-
lates have several distinct strategies for avoiding
predation. Hiders such as the white-tailed deer rely
on suitable vegetation cover to conceal them from
predators; they can also usually run effectively
through dense vegetation to escape predators. Many
quantitative definitions are available for what consti-
tutes hiding cover for an individual ungulate species.
For example, Thomas et al. (1979) has defined hid-
ing cover for elk as "vegetation capable of hiding an

eik from view at 200 ft or less with an optimum
size of 30 to 50 acres."

Watchers such as pronghorn antelope live in
open country and rely on observation to protect
them from predators. Some watchers can effectively
outrun their predators in open country, which may
be level terrain in the case of the pronghorns or
rocky slopes in the case of bighorn and Dall's sheep.
Other watchers rely on group defense (e.g., musk-
oxen) or individual defense (e.g., moose) as protec-
tion. Watchers tend to congregate into groups
ranging from small family units to large herds (over
100 animals). For these ungulates, security cover
consists of suitable open habitat. Pronghorn ante-
lope, for example, prefer habitat with a mean height
vegetation of 37.5 cm (15 in.); areas with vegetation
over 60 cm (24 in.) are less preferred and those
over 75 cm (30 in.) are rarely used (Yoakum 1980).
This preference can be attributed to security cover
since suitable forage is often more abundant in areas
with taller vegetation.

Many ungulates use more than one strategy. For
example, mule deer tend to hide; yet on many
ranges they will congregate into large herds in open
areas, such as for feeding in late winter and early
spring, and use a watcher strategy. Often this feeding
takes place near rocky terrain or more thickly vege-
tated areas where deer can more effectively outrun
predators.

Ungulates may also use different strategies in
different parts of their zoogeographic range or even
from one range to another. Elk, for example, tend
to use coniferous forests for hiding cover in much of
the West; however, in recent years they have natu-
rally colonized shrub-steppe regions of Washington,
which are devoid of trees, and also sagebrush

Ungulates

527

steppes of Wyoming, where they use tall shrubs for
security cover. Similarly, animals may use different
strategies against different predators. Bighorn sheep,
for example, will run from wolves, but they may
turn and defend themselves from coyotes.

Because of the different strategies and the diver-
sity of security cover that ungulates may use, a biolo-
gist must be cautious in assessing what is suitable
security cover. First-hand knowledge of the range
being analyzed and of the behavior of the animals
that use it are necessary.

Special Cover. Some ungulates use special
areas for cover from other forces. For example, cari-
bou may seek windy areas in summer where they
are more protected from biting and stinging insects.

Relationship between Cover Preference and
Requirements. Preference for a cover type does
not necessarily indicate that the type is required for
survival or that its absence will limit the herd (Allen
1977; Peek et al. 1982). For example, bighorn sheep
have evolved a predator-evasion strategy and use
rocky terrain as escape cover from wolves. Bighorn
sheep continue to use such habitats even in areas
where wolves have been extirpated for over 50
years. Do they need it or merely prefer it? Such
questions are difficult to answer and raise questions
of value (e.g., should we be managing for "natural"
animals; when does a wild animal become domesti-
cated?). These questions are, for the most part, be-
yond the scope of this publication. However, the
subject of cover deserves some cautionary notes.

If an animal has evolved a strong preference for
using a certain type of security or shelter cover and
this cover is available within its range, then that
preference will effectively determine which portion
of the range the animal will use. For example, big-
horn sheep often deplete forage on small areas of
suitable open, rugged terrain to the detriment of the
herd, even though vast expanses of coniferous forest
with suitable forage surround such areas. Under
these conditions, the question of preference or re-
quirement is moot. Biologists should be cautious in
assuming they understand better than the animal
what it needs. If an animal has evolved or learned a
strong preference for a given type of cover, it is
probably adaptive, even though the reasons may not
be obvious to the biologist.

Birthing Areas. Many ungulate species use special
birthing areas. The lambing areas of wild sheep
(Geist 1971a) and calving grounds of caribou (Lent
1974a) are well-known examples. Birthing habitat
requirements of species such as pronghorn ( Fichter
1974) have been of management concern. Such spe-
cial areas arc frequently identified in the develop-
ment of land-use plans, habitat management plans,
and environmental impact assessments. For those

species in which clear birthing habitat descriptions
are available, the wildlife biologist should be alert to
identify areas meeting these specifications. Docu-
menting early infant mortality rates in ungulates is
extremely difficult, however, and data to describe
the demographic consequences of loss or alteration
of birthing areas are not available. Most examples are
anecdotal and relate to direct human disturbance
(Kurt 1968; Lent 1974a).

Special habitat requirements for cover may be
anticipated, especially for hider species (Lent 1974a)
that depend on predator defense to isolate the partu-
rient female from others of the species and subse-
quently isolate the newborn for long periods of the
day. Such antipredator strategy may be significant for
up to 2 months after birth, like in mule deer (Geist
1971a, 1981).

Migration and Movements

Most North American ungulate populations
show predictable patterns of movements over the
course of a year, periodically returning to some orig-
inal locality. Nomadic movements are those in which
such a fixed, recurrent locality seems to be absent.
The more general term "movements" is used in the
literature, particularly in reference to migrations
of relatively short distance or when the type of
movement is not clear.

Even white-tailed deer, generally considered to
be relatively sedentary, may make seasonal migra-
tions of 20 to 100 km (12 to 60 mi.). They occupy
more restricted home ranges (X = 44 ha [110 a.])
in winter, larger ones in summer ( 1 00 to 500 ha
[250 to 1,250 a.]) (Nelson 1979; Nelson and Mech
1981). This illustrates the phenomenon of seasonal
home range prevalent in many temperate and boreal
populations. That is, there are relatively localized
seasonal use areas where movements are over short
distances, connected by corridors of longer distances
traversed relatively quickly. Moose in Wyoming rep-
resent another well-documented example (Houston
1974).

The biologist must be particularly alert for situa-
tions where resident (relatively sedentary) and mi-
gratory populations of the same species may overlap
during certain seasons. This has been reported for
white-tailed deer (Nelson and Mech 1981 ), caribou
(NPR-A Task Force 1978, 1979; Cameron and Whit-
ten 1979), and moose (LeResche 1974). Such over-
laps, which often require marking or telemetry
programs to identify and understand, have obvious
implications for management. Whether one needs to
estimate seasonal carrying capacities, regulate har-
vest by establishing zones or seasons, or assess im-
pacts of proposed developments, one needs to be
able to discriminate between such resident and mi-
gratory populations.

528

Ungulates

POPULATION MEASUREMENT TECHNIQUES

Measuring ungulate population attributes is one
of the most important, yet most difficult and expen-
sive, aspects of ungulate management. Mule deer
have been studied, measured, and censused for over
50 years in the West, yet Connolly ( 1981:226)
wrote —

"No reliable estimate of mule . . . deer numbers
exists for any state or province. The only avail-
able estimates are speculative and often they
are no more than guesses by the best-qualified
persons."

Determining numbers or density of ungulates
and trends for these measurements are extremely
important not only as a measure of the effectiveness
of habitat management actions, but also as a basis for
setting harvest quotas or goals. Biologists have stud-
ied this problem extensively, and a vast body of liter-
ature has accumulated. Some state agencies are
making substantial progress in improving methodol-
ogy (Gill et al. 1983). There are many disagreements
on the subject, however, and much work needs to
be done.

Because so much work has been done on this
subject, this section only serves as a brief overview
to literature on the subject. Biologists will need to
consult original references for detailed information
on methodology and extended discussions of advan-
tages and limitations of specific techniques.

Neck-banded cow elk in summer range herd

Because state wildlife agencies are responsible
for managing resident ungulate species, they are
usually more active in developing and implementing
techniques for gathering population data on ungu-
lates. Biologists in land-managing agencies, who re-
quire population data, should consult with state
agencies to determine whether such data are avail-
able and can be obtained from them or through a
cooperative effort. Ungulate surveys are expensive;
publicly funded wildlife programs cannot afford
duplications.

Presence/ Absence

Evidence of presence or absence of an ungulate
species in a given area can be determined by di-
rectly observing animals or tracks, droppings, or
other signs. Tracks of most North American ungu-
lates are relatively distinct (Murie 1954). Droppings
may be more difficult to distinguish; droppings of
mule and white-tailed deer, pronghorn antelope,
bighorn sheep, and mountain goat overlap in size,
shape, and appearance. However, a biologist may be
able to rule out all except one of these species
through knowledge of habitat requirements and geo-
graphic distribution.

Although data on presence/absence are rela-
tively easy to collect, they are of limited value for
ungulates. Management agencies and the public are
more concerned with numbers, density, or relative
density. Presence information is most useful in deter-
mining geographic distribution or use of a range.
The geographic distribution of North American un-
gulates is well-known on a continental or regional
basis; however, local patterns of distribution are
often poorly understood. For example, only in the
past 35 years has the distribution of bighorn sheep
in North America been described in enough detail to
indicate whether a given mountain range supported
sheep populations. Furthermore, much of this infor-
mation has only been gathered in the past 1 5 years;
some states still do not have this information. Finally,
detailed information on habitat use, seasonal ranges,
lambing areas, etc. is lacking for most of these herds,
and much of the information obtained in the past
needs to be updated. Some ungulates such as moose
are expanding their range; other species are being
locally extirpated and many herds are abandoning or
no longer using parts of their range. Knowledge of
seasonal ranges including migration corridors and
birthing areas is essential for good habitat manage-
ment. This is basic and very important information
and often can be obtained with equipment no more
sophisticated than a field notebook, topographic
maps, field glasses, and a stout pair of hiking boots.

Relative Density or Relative Abundance

Relative density or relative abundance are com-
monly used to measure the effectiveness of habitat

Ungulates

529

management actions. Collecting this data provides an
index of population abundance, such as deer ob-
served per square mile or antelope counted per mile
of highway. Many such indexes have been devel-
oped; some require directly observing animals,
whereas others use indirect signs, such as tracks and
fecal pellets. Indexes can then be compared from
year to year or from area to area. Population indexes
do not usually require the biologist to make the
same sort of assumptions as required with popula-
tion estimation methods. Furthermore, they are eas-
ier and cheaper to obtain and, for measuring
population trends or comparing areas, they may be
adequate.

The major problem with indexes is ensuring
observations or counts are replicated as nearly as
possible. Factors that may need to be standardized
include routes, flying or driving conditions, weather,
experience of personnel, training, protocol for such
things as counting an animal in or out of an area,
time of day, etc. The value of indexes for measuring
year-to-year changes in abundance significantly de-
creases and becomes useless if the techniques or
protocol for collecting the information is periodi-
cally changed.

One technique, the pellet group count, needs
further explanation. Biologists have attempted with
varying success to use this technique for quite differ-
ent purposes for over 50 years (Neff 1968). As a
method for determining presence, it is sound. As a
method for population density estimation, it has
significant problems (Neff 1968). As a method for
determining relative density, it is useful for compar-
ing years or large areas, but may be misleading if
used to compare habitat use between or among
small adjacent areas. The problem is that defecation
does not correlate well with habitat use. Elk, for
example, often defecate while traveling (40% of all
defecations), yet this activity consumes only about
5% of their day (Collins and Urness 1979). Using
such a technique within the home range of an elk
herd would erroneously suggest that areas used for
travel were heavily used. On the other hand, if the
technique is used to compare average relative densi-
ties on areas large enough to contain the daily home
ranges of animals, then such factors are of negligible
importance.

Density-
Estimating density or total numbers in a herd is
difficult and expensive, yet if done carefully, it can
provide very useful information. The many tech-
niques available fall into four general categories: total
counts (census), sample counts, ratio methods, and
indirect methods.

Total Counts. Total counts are most successful
with diurnal ungulates on open country, such as

530 Ungulates

pronghorn antelope. Observers on foot often try to
totally count small isolated herds that are concen-
trated, such as on a winter range or at waterholes in
summer. Such counts usually require good planning
and coordination to ensure that all animals are ob-
served and are not counted more than once.

Various methods have been devised to correct
or adjust for these problems, but none are entirely
satisfactory. For this reason, information from many
such efforts are used as indexes rather than total
counts.

Total counts are often attempted from helicop-
ters or fixed-wing aircraft, although more often some
form of sampling is used. The practical problems
are similar, but sampling schemes make the calcula-
tions more complex.

Sample Counts. Sample counts are of two general
types, strip samples and plot samples. Strip samples
have been used effectively to count ungulates from
airplanes, motor vehicles, or on foot (Riney 1982:78-
84). Many different types of strip samples are avail-
able; Robinette et al. (1974) tested many of these
and Burnham et al. (1980) summarized assumptions
and statistical analyses.

Plot counts have been used less frequently, es-
pecially from airplanes. However, a helicopter quad-
rat system (Kufeld et al. 1980) is being used in
Colorado and initial results and tests appear promis-
ing (Gill et al. 1983).

Ratio Methods. Ratio methods, including mark-
recapture methods, are based on knowing the num-
ber of marked animals in a population or removed
from a population at some point. These methods
generally require animals to be marked or tagged —
expensive and time-consuming with ungulates. Many
variations on the basic capture/recapture models
have been developed. Refer to Riney (1982:42-74)
for practical aspects of capturing and marking ani-
mals, to Bailey (1984:311-315) and Gilbert
( 1979:302-304) for a general discussion of ratio
methods, and to White et al. (1982) for a discussion
of detailed statistical treatment.

Indirect Methods. Indirect methods correlate ani-
mal density with some measurable and observable
sign left by the animals. With ungulates, the indirect
method used most frequently is the pellet group
density measurement (Freddy and Bowden 1983a,b;
Neff 1968).

General Cautions. Gill et al. (1983) suggested that
biologists obtain reliable estimates of ungulate popu-
lation density to track changes over time. This is
certainly laudable and biologists with agencies re-
sponsible for population management would be well-
advised to pursue it. However, because such esti-

mates are expensive and difficult to obtain and the
results often debatable, many biologists and man-
agers prefer to use indexes. Furthermore, although
many techniques have serious disadvantages when
used to count and estimate populations, most of
these disadvantages disappear when they are used as
indexes (Riney 1982:85). The biologists charged
with habitat management or working for a habitat
management agency should consider whether a pop-
ulation or condition index will be adequate for their
purposes.

Animal Condition

Condition indexes have been developed and
used experimentally for many years to measure un-
gulate health, but few have been used operationally
by management agencies. Generally, condition meas-
urements or observations are collected on individual
animals and then averaged to derive an index of the
condition of the herd or population. Animal condi-
tion indexes can be considered a function of the
habitat and the number of animals. Thus, unlike most
other population measurements, they reflect ecologi-
cal density or the number of animals relative to the
quantity and quality of habitat available (Bailey
1984:318). Indeed, many measures of population
structure, such as doe to fawn ratios (a measure of
productivity) or behavioral indexes, can be consid-
ered special types of population condition indexes
(Bailey 1984:318-324; Hanks 1981).

Population Structure and Productivity

With most North American ungulates, sex and a
limited number of age classes (young of the year,
yearling, adult) can be distinguished in the field. In
addition, more detailed age information can be ob-
tained from dead or harvested animals using tooth
eruption characteristics, dental cementum annuli, or
horn growth rings. Biologists have collected much
data on sex and age structures of ungulate popula-
tions and expended much effort in developing and
refining methods to analyze such data (Caughley
1977; Riney 1982; Fowler and Smith 1981). This
type of information is now commonly used in popu-
lation models that can simulate the dynamics of herd
numbers and structure over time (Conley 1978).
The models are now used routinely by several west-
ern state game agencies as a source of information to
be used with other sources in making recommenda-
tions about harvest goals and quotas for big game
species (Pojar and Strickland 1979).

Some very good work has been done with popu-
lation structure measurements. However, manage-
ment biologists must keep in mind the original
reasons for collecting population structure data and
avoid getting bogged down in the complexities of
population modeling. For management purposes, the
simplest and directly observable measures, such as

cow to calf ratios, have often proven to be the most
useful.

Behavioral Indicators

Interest in the use of behavioral attributes as
indexes of population and habitat condition is in-
creasing. Hanks (1981:61) reported that several stud-
ies of population condition correlate declines in
physiological condition or demographic vigor with
certain behavioral attributes of the population. He
recommended considering these attributes as a pos-
sible early warning system of adverse population
trends: ( 1 ) rate and quality of social interaction, (2)
population density, and ( 3 ) feeding strategy. As an
example, Horejsi (1976) concluded that duration of
suckling and total time spent suckling could be re-
lated to population and habitat quality in bighorn
sheep.

MAJOR SPECIES

The following descriptions treat each major
ungulate species or species group separately because
of the importance of these animals. Because many
of these species are similar in many ways, the sec-
tions emphasize aspects of the biology and manage-
ment that are unique to the species.

Many books and monographs are available on
North American ungulates. Some of the more recent,
major publications about population or habitat meas-

Jaw board used as aging reference at big game check sta-
tions.

Ungulates

531

Table 2. Major recent references on North American ungulates that cover habitat and population

i

measurements.

Species

Periodicals

Books/Monographs
Bibliographies

General
Individual Species

Schmidt and Gilbert
(1978)

Collared peccary

Sowls (1984)

Elk

Proceedings of Western States Elk Workshop

Thomas and Toweill
(1982)

Mule deer

Wallmo (1981)
Kerr (1979)

White-tailed deer

Halls (1984)

Moose

Alces (Proceedings of North American Moose Confer-
ence and Workshop

Caribou

Proceedings of International Reindeer and Caribou Sym-
posiums

Pronghorn

Proceedings of Pronghorn Antelope Workshop

Yoakum (1980)
Autenrieth (1978)

Bison2

Mountain goat

Proceedings of Biennial Symposium of the Northern
Sheep and Goat Council

Samuel and Mac-
Gregor (1977)

Muskox

Proceedings of International Muskox Symposiums

Klein et al. (1984)

Bighorn sheep
and

Proceedings of Desert Bighorn Council

Monson and Sumner

(1980)
Trefethen (1975)

Dall's Sheep

Proceedings of Biennial Symposiums of Northern Sheep
and Goat Council

Krausman et al. (1984)

'Only references since 1975 are cited.
2No recent general references available.

urements are listed in Table 2; others are cited in
specific species subsections.

Collared Peccary

Of all large game animals in North America, the
collared peccary (Tayassu tajacu) has been the last
to get the attention of wildlife biologists. In many
places it was considered vermin or a pest to live-
stock operators. In the past 30 years, however, the
peccary has become a protected big game species in
Arizona, New Mexico, and Texas. Careful censuses
and regulated harvests now ensure its survival if
habitats can be preserved. The most serious limiting
factor to the future survival of this animal is habitat
loss. Areas of unspoiled habitats need to be set aside
to ensure their future existence (Sowls 1984). Preda-
tion levels are thought to be low, although hard

i^*-"p.

^^t?^^:. *

Collared peccary.

532

Ungulates

data are not available on the effects of predation on
population levels (Sowls 1984). Although the col-
lared peccary could have both ecto- and endoparas-
ites, they do not seem to cause significant population
losses (Sowls 1984).

The species is not long-lived (mean length of
life is about 4 years), but the high reproductive rate
and its ability to withstand severe food and water
shortages for short periods enables the species to
maintain good, young animal populations.

Habitat Requirements and Analysis. In the South-
west, collared peccaries occupy a variety of habitats.
Habitat types occupied by peccaries in Arizona have
been described by Knipe ( 1957), Neal (1959), and
Eddy ( 1961 ). Arizona populations are found primar-
ily in the southeastern quarter of the state and are
generally contiguous with populations in New Mex-
ico (Bissonette 1982). The Arizona populations in-
habit areas of low desert-scrub vegetation where the
principal plants are saguaro (Carnegia gigantea),
mesquite (Prosopis sp. ), paloverde (Cercidium sp. ),
and many species of cactus. The oak-woodland, chap-
arral, and lower areas of ponderosa pine {Pinus port-
derosa) forests are occupied during the summer,
but can be too cold for peccaries in the winter. Pec-
caries can inhabit desert environments because they
make full use of microclimates, and during very hot
summer periods, they move only at night.

Peccaries in New Mexico occupy two types of
vegetation cover. One is the oak-juniper association
between elevations of 1,500 and 1,940 m (5,000 and
6,400 ft) just below the ponderosa pine belt. The
other cover type is mesquite, beginning at elevations
about 1,303 m (4,300 ft) and extending to about
1,879 m (6,200 ft) on some exposures (Donaldson
1967). Other protective cover includes caves and
mine shafts, natural crevices, rock overhangs, and un-
dercut arroyo banks.

Habitats occupied by peccaries in Texas vary
from dry deserts in the west to lush chaparral and
deciduous brushlands in the east. Peccaries occupy
about 260,000 km2 (100,000 mi.2) in this state in
areas having rainfall varying from 40 cm (16 in.) per
year in the west to 75 cm (27 to 30 in.) per year
in the east (Jennings and Harris 1953; Low 1970;
Bissonette 1976; Ellisor and Harwell 1979).

Food and water are supplied by succulent food
(Sowls 1978). The prickly pear cactus is the pre-
ferred food of the peccary in Arizona (Knipe 1957),
but tubers, bulbs, and rhizomes are also important.
In Texas, good peccary habitat consists of heavy
brush with an abundance of prickly pear (Ellisor and
Harwell 1979). In New Mexico, major food items
include oak (Quercus sp.) mast, juniper (Juniperus
deppeana and/ monosperma), pinyon (Pinus edu-
lis), mesquite beans, leaves and fruits of Engelmann

prickly pear (Opuntia engelmanni), and mescal
(Lophophora williamsii) or Palmer agave (Agave
palmer i) (Donaldson 1967).

Peccaries do not depend entirely on free water,
but use it if it is available. Peccaries can usually ob-
tain ample amounts of water from prickly pear, the
basal ends of mescal, and underground bulbs (Don-
aldson 1967). Thus they can withstand long dry
periods without ill effects.

Since prickly pear cactus seems to be the most
common of all peccary foods, habitat measurements
should include determination of abundance of
prickly pear. Because the greatest threat to peccary
populations is habitat loss (Sowls 1978), the vegeta-
tion trend of peccary habitat should be monitored,
particularly that of prickly pear.

Population Measurements. Compared with more
common big game animals, peccaries present new
problems to researchers trying to gather and inter-
pret biological data. The principal problem is that
peccaries have, a year-round breeding season. This,
coupled with the female's ability to become preg-
nant within a few days after giving birth, gives the
species a very high breeding potential. Because of
these factors, these animals do not fit as easily into
age classes or cohorts like elk, deer, and other big
game animals (Sowls 1984). Life tables have been
constructed from ages of hunter-killed animals, but
problems, primarily unequal birth and survival rates,
make use of these analytical techniques suspect
(Low 1970).

In Texas, aerial surveys, road censuses by
ground vehicles, and track counts have been tried
for inventorying these animals. These census meth-
ods, however, are not feasible in brushy country.
In Arizona, extensive testing of peccary census meth-
ods involved track counts, counts at waterholes,
helicopter counts, surveys by biologists on foot, and
questioning hunters about numbers of animals seen.
Results indicated that helicopter counts were too
expensive and variable, and track counts were too
variable. The Arizona Game and Fish Department
found that deer hunter counts of peccaries, obtained
right after a hunt, were more accurate than sightings
reported later in questionnaires. They also reported
that data from deer hunters compared well with data
obtained by biologists surveying on foot (Sowls
1978). Thus, the most promising method seems to
be questioning hunters in peccary range immediately
after hunting seasons for other species, particularly
deer. Combined with foot or horseback observations
by biologists, this method gives good indications of
population number and trend. Roadblocks and estab-
lished hunter checking stations also provide impor-
tant data on peccary populations, and data on
productivity can be obtained by estimating ages of
harvested animals.

Ungulates

533

Discussion. The most important management con-
sideration is maintaining as much suitable peccary
habitat as possible. The collared peccary has high
reproductive potential and is remarkably adaptable.
For these reasons, healthy and productive popula-
tions of this unique animal can be effectively man-
aged throughout its range. Although habitat factors
limiting peccary populations are not well-known,
evidence suggests that density of prickly pear is very
important and should be a high priority for inven-
tory and monitoring.

Elk

At one time, the American elk (Cervus elaphus)
was the most widely distributed member of the deer
family in North America. It was found from the At-
lantic to Pacific coasts and from Mexico to northern
Alberta, except the southern coastal plains and Great
Basin (Hall and Kelson 1959). However, as people
moved westward, elk disappeared from settled re-
gions until they virtually vanished from most of their
historic ranges. Remnant herds were scattered in
the Rocky Mountain region and parts of the Pacific
Northwest and Canada. In the past 50 years, elk have
expanded their range and been reintroduced so that
they are fairly widespread throughout the mountain-
ous areas of the West.

Habitat Requir ements and Analysis. Much of the
work in recent years on elk habitat has concentrated
on ( 1 ) identifying areas used for foraging, cover,
and other needs, and the necessary spatial relation-
ships of these areas using the Thomas et al. ( 1979)
landscape approach, and (2) determining the effects
of logging, roads, fire, livestock grazing, and other
human disturbance on habitat suitability for elk.

Thomas et al. (1979) defined optimum elk habi-
tat for the Blue Mountains of Oregon and Washing-
ton in terms of a proper mixture of foraging and
cover areas. They further subdivided cover into hid-
ing cover and thermal cover. This landscape ap-
proach has been modified for use in many other
areas of the West. For example, Witmer et al. ( 1985)
have used this approach for the elk in the coastal
forests of Oregon and Washington, and Wisdom et al.
(1986) have formalized this into a habitat evaluation
model.

The greatest limitation of such a landscape ap-
proach is that the quantity, quality, or availability of
forage in "foraging areas" is not considered. Elk con-
sume more forage per animal than most of the other
common North American ungulates, and their forage
preferences are similar to those of domestic cattle
(Cooperrider 1982). Therefore, where elk are reach-
ing carrying capacity or where they occupy common
range with cattle, forage quantity and quality can be
a problem.

For a current and more thorough review of elk
habitat requirements and evaluations, see Skovlin
(1982).

Population Measurements. Enumerating various
components (age structure, sex and age ratios, and
total population) of an elk population is difficult, but
important. These data tell the biologist or manager
much about the dynamics and general health of the
herd, and with adequate sample sizes, they provide
the information needed for proper elk population
management.

Age Structure. Two methods are used to de-
termine age structure. The first involves examining
tooth wear and replacement in elk killed by hunters.
Aging by this method is based on the sequence of
tooth eruption and wear. Several good aging guide-
lines are available, but techniques developed by
Quimby and Gaab (1957) are used by most wildlife
researchers and managers. One problem associated
with this method, however, is obtaining a large
enough sample to enable the manager to make an
accurate estimate of the herd's age structure.

To increase samples, hunters are invited to re-
move the lower jaws from their kills and bring them
to check stations. However, ages obtained by this
method do not necessarily represent a herd's age
composition. If a wildlife management agency is
willing to spend the time and money, rather large
sample sizes of jaws can be gathered by setting up
several check stations around one herd area. This
method is effective for special research areas or ex-
tremely important herds, but is not satisfactory for
statewide age-structure estimations.

The second method is the dental cementum-
aging procedure, which requires only an incisor

Cross-section of an incisor from an unknown age elk killed
in October; age estimated at six years by cementum annuli
count.

534

Ungulates

tooth from a harvested elk (Keiss 1969). By using
this method, age samples from all elk herds in a state
can be greatly increased. The tooth is sectioned with
a special sectioning saw, put through various staining
procedures, and then examined through a micro-
scope for cementum annuli which can be counted in
a manner similar to counting rings in a tree stump.
By sending every hunter, or a statistically reliable
sample of hunters, a small postage-paid envelope in
which to collect and mail an incisor, sample sizes
can be increased substantially. A hunter can more
easily remove and carry an incisor than a whole jaw.

Sex and Age Ratios. Ground counts and aerial
counts are two general methods used to estimate sex
and age ratios. Ground counts are useful in areas
where elk congregate in large numbers, such as feed-
ing grounds or winter ranges. Ground sex and age
classifications can be considerably biased, however,
because larger bulls tend to winter away from the
cow/calf/young bull groups. Ground sex and age
classifications are generally only possible from De-
cember through February, when elk are congregated
on their winter ranges. This means that some mortal-
ity has occurred after the hunting season so that
ratios are not exactly the posthunt ratios needed for
some population formulas. In addition, large samples
of prehunt sex and age ratios can rarely be obtained
by ground-count methods. While the total number of

elk on winter ranges can be satisfactorily estimated
from a fixed-wing aircraft, sex and age ratio counts
cannot. Low and slow-flying helicopters are needed
to accurately determine these ratios.

Prehunt sex and age ratio counts should be
made in mid-September at the peak of the rut when
the largest percentage of bulls can be observed. Elk
can be classified as spike bulls; young bulls (typical
antler conformation, but with light beams and usu-
ally four or five points); mature bulls (typical antler
conformation, but with heavy beams); cows; and
calves. Calves can be distinguished from yearling
cows primarily from the face or nose conformation,
but color, relative thickness of the neck, and com-
parative size are other criteria (Boyd 1970, 1978).

Biologists can facilitate classifications by using
tape recorders to tally various categories and later
extract the data. This method allows them to watch
the elk continually while recording the sex or age of
each animal.

Sex and age ratio counts can be made in early
December, before the larger bulls have moved to
high and isolated wintering areas away from the
main herd, to estimate changes caused by fall hunt-
ing. Techniques and criteria of posthunt counts are
not different from prehunt counts.

Calves are distinguished from yearling cows primarily on the basis of nose or face conformation (a calf has a short, blunt
nose). The neck on a calf appears to be shorter and heavier than that of a mature cow. A calf has a fuzzy, rounded patch
on its forehead and in general has a "puppylike" appearance.

Ungulates

535

Total Population. To properly manage an elk
population, the wildlife manager must estimate the
total elk population on a given range. Several census-
ing methods can be used to make these estimates.

Elk on winter ranges have been aerially counted
since about 1946. These sample counts, for the most
part, are accomplished by using fixed-wing aircraft
but do not encompass all elk in a survey area. In
Colorado, the percentages of deer or elk observed
from the air were compared with the total popula-
tion present for several years. Only about 37% of the
elk present on a sagebrush-oakbrush-serviceberry
(Artemesia-Quercus-Amelanchier sp. ) wintering area
actually were seen by aerial observers (Boyd 1958).
However, winter counts are used by many wildlife
management agencies to evaluate the population
trends and to set annual hunting seasons and
regulations.

Indirect census methods have been used suc-
cessfully by several management agencies by con-
trasting prehunt and posthunt sex ratio counts and
harvest estimates (Boyd 1970). Among the many
references concerning these methods, one excellent
review applicable to elk is by Hanson (1963), which
provides a discussion on calculating abundance of
vertebrates from sex-age ratios.

While population figures resulting from these
methods are extremely valuable, they do not tell the
entire story of elk herd dynamics. Missing data in-
clude estimates of nonhunting mortality and formu-
las to calculate the consequences of past and future
management on elk populations.

A wildlife manager also needs to set manage-
ment goals and priorities for managing wildlife popu-
lations. Obtaining reliable information on which to
base these objectives, however, is difficult. Without

Mature bull elk.

tools or methods to assist in gathering and analyzing
data, management often ignored many elk population
factors when setting harvest limits, regulations, and
management priorities. Now, with the development
of the computer, these factors can be considered
(Gross 1970; Pojar 1981). All factors that contribute
to the life cycle of an elk population (i.e., sex-age
ratios, cow-calf ratios, harvest by sex and age,
wounding loss) are needed to conduct a simulation
(Boyd et al. 1975).

One new approach to censusing elk involves the
use of mark-recapture techniques to estimate num-
bers of elk (Bear and Green 1980). Individual elk in
a wintering area are randomly trapped and marked
with colored ear tags. An attempt is made to capture
about 10% of the estimated herd. The recapture
effort is accomplished by surveying the entire winter
range with a helicopter as soon after trapping as
possible. All ear-tagged individuals are counted to
determine the ratio of marked to unmarked animals.
Three or four recapture flights are made each year.
This system is quite precise (SE ± 5% ). Accuracy is
unknown, but the estimates agree very well with life-
table analyses, other aerial counts, and estimates
from biologists familiar with the area (Gill et al.
1983).

Discussion. The major concerns of managing
elk habitat focus on the effects of competing land
uses, particularly logging, grazing, and other human
disturbance. Logging especially is a concern because
it physically alters the habitat through cutting trees,
piling slash, and other forestry-related activities. In
addition to the immediate disturbance from such
activity, habitat is also subsequently disturbed from
roads built as part of the logging job. Not only do
elk tend to avoid well-used roads, such roads also
provide increased hunter access. The effects of log-
ging, roads, and other human disturbances are de-
scribed by Lyon et al. (1985), Lyon and Ward
(1982), and Heib(1976).

Grazing is another concern because elk compete
with livestock for forage. The competition is most
severe with elk and cattle because they occupy
many common ranges, their diets are similar, and
both species consume large amounts of forage
(Cooperrider 1982).

Mule and White-tailed Deer

The mule deer {Odocoileus hemionus) ranges
over most of temperate North America between the
Pacific Coast and the 100th Meridian. The white-
tailed deer (O. virginianus) is the most popular and
widespread big game animal in North America, found
in almost all the contiguous 48 states. It thrives in a
wide range of climatic and habitat conditions and
adapts quickly to changes.

536

Ungulates

Mule and white-tailed deer are often victims of
automobile accidents, diseases, parasites, predation,
starvation, wounds, fence entanglements, poaching
and, of course, legal hunting. A primary limiting fac-
tor for these species would be habitat degradation
caused by urban development, deteriorating forest
habitat, and agricultural development. Of all these
limiting factors, the primary management concern is
habitat loss.

Habitat Requirements and Analysis. Mule deer
and white-tailed deer have been intensively studied
for many years. However, because they occupy such
diverse habitats, generalizations about habitat re-
quirements are extremely difficult to make. A biolo-
gist should, therefore, consult the literature on a
particular region for specific habitat requirements.
Wallmo (1981) and Halls (1982) provide extensive
reviews of habitat evaluation and management for
mule and white-tailed deer, respectively.

Because mule deer are adaptable, they may be
found in all major climatic and vegetational zones of
the western U.S. except the arctic, tropics, and ex-
treme deserts. Generally, mule deer inhabit semiarid,
open forest, brush, and shrublands found in conjunc-
tion with steep, broken, or otherwise rough terrain.
However, extensive populations are also found in
prairie habitats along the eastern limits of their dis-
tribution and in semidesert shrub habitats of the
Southwest (Mackie et al. 1982). Black-tailed deer,
which consist of two mule deer subspecies, are
found in temperate, coniferous forests along the Pa-
cific Coast from central California, north to south-
eastern Alaska. Populations can also be found in
adjacent woodland-chaparral areas of the coastal
ranges of California.

In general, the white-tailed deer does best in
subclimax habitat (Hesselton and Hesselton 1982),
particularly cut over forests. This is particularly true

in the eastern U.S. However, west of the 100th Me-
ridian, white-tailed deer are often associated with
riparian areas along major river drainages. Many of
these areas consist of forested stands of cottonwoods
(Populus sp. ), willows (Salix sp. ), and other tree
species. Although white-tailed deer have not been
studied as extensively in the West, the evidence
suggests that these habitats are very important if not
essential for many white-tailed deer herds.

The same caution applies to mule deer. Whereas
early research and writing emphasized the impor-
tance of early successional habitat, recent research
suggests that, in some areas, climax forest is the
most important of the available or potential habitat
types for maintaining high year-round carrying ca-
pacity (Wallmo and Schoen 1981 ). This is particu-
larly true of black-tailed deer at the northern areas of
their geographic range.

Both mule and white-tailed deer require a diver-
sity of plant species as food at any one time and
during an entire year. Diverse vegetation is also re-
quired for hiding, escape, and thermal cover. Be-
cause of this, interspersion of several individual
forage and habitat types may be more important than
the occurrence of individual types.

Population Measurements.

Direct Counts. Four basic methods are used to
count mule and white-tailed deer: aerial counts,
drives or "complete" surveys, flushing counts, and
automobile counts. For direct counts, aerial census is
probably the only method to be considered. These
counts are usually conducted during the winter
when there is a good counting background,
deciduous trees have shed their leaves, and deer are
concentrated on limited winter ranges.

Note differences in antler conformation between the white-tailed deer buck (left photo) with the points coming off the
main beam and the mule deer buck (right photo) with dichotomous branching.

Ungulates

537

Aerial counts from fixed-wing aircraft are gener-
ally limited to total counts of entire herds or total
counts on trend areas, flown annually. As stated for
elk, sex and age ratio data cannot usually be ob-
tained with the use of fixed-wing aircraft. The use of
aerial censusing has been revolutionized by the heli-
copter, however. Sex, age, and area data can now
be obtained that were not possible from fixed-wing
aircraft. Helicopter counts are more expensive, but
better quality data can be obtained.

Present-day aerial census has moved from total
counts to sample-based estimates of population size
(Gill et al. 1983). A helicopter-quadrat system of
censusing deer has been developed in Colorado
(Bartmann 1974; Kufeld et al. 1980) that consists of
1.6-km (1-mi. ) quadrats, randomly located in win-
ter range areas and permanently marked so the same
areas can be flown each year. The accuracy, preci-
sion, frequency, and resolution of the quadrat census
method are currently being studied (Gill et al.
1983).

Indirect Counts. The primary indirect census
method is the pellet counts. Pellet counts are best
used in wintering areas. This technique is based on
assumptions of the number of pellet groups depos-
ited every 24 hours by an individual deer. A winter
range is censused for pellet groups after the deer
leave an area. All pellet groups within a series of sta-
tistically designed plots are then counted. The num-
ber of pellet groups per hectare (acre) is multiplied
by the number of hectares (acres) per habitat type,
resulting in an estimate of the total number of pellet
groups per habitat type. Another factor that must
be closely estimated is the number of days that deer
have been on the area sampled. Knowing this, the
number of pellet groups per hectare (acre) for each
habitat type and the daily defecation rate, results in
an estimate of the number of deer using the area
(Neff 1968; Ryel 1971; Anderson et al. 1972).

A review of research on methods to estimate
total populations of mule and white-tailed deer over
large areas revealed that the helicopter quadrat and
pellet group count methods mentioned above were
conceptually feasible for implementation on large
areas of the western U.S. No single system can apply
to all areas of the western U.S. Topography, weather,
and vegetation composition are primary items that
dictate the system to be used at different localities.

Helicopter aerial surveys appear to be the
method of choice wherever possible, because they
require less time, are more accurate than fixed-wing
aircraft counts, and provide direct estimates of popu-
lation size. Aerial censuses cannot be used, however,
in areas of turbulent winds or where combinations
of topography and mottled snow background pre-
clude accurate, consistent counts.

Research in Colorado revealed that pellet group
counts on permanently marked plots and on tempo-
rary plots yielded comparable estimates of popula-
tion size. Seemingly, temporary plots are more
practical because they are easier to implement and
are less costly (Freddy and Bowden 1983a).

Discussion. Because habitat and population meas-
urements are so similar for mule and white-tailed
deer, the species were combined in this section to
save repeating much of the information. Some of the
problems encountered while inventorying and moni-
toring deer include a lack of accurate and precise
census methods, extreme difficulty in determining
carrying capacity on seasonal ranges, and the fact
that some populations inhabit areas of the western
U.S. where winters are so mild that winter range
concentrations are unknown and year-round ranges
are normal.

Factors limiting deer populations in the short
term are still not well understood. For example,
available evidence indicates that mule deer declined
substantially over most of the species' range from
the early 1960s to around 1976, then began to in-
crease with no generally accepted explanation for
either the decline or recovery. Connolly (1981:243)
wrote —

"It reveals how little control biologists and
managers have over the deer they purport to
manage. Just as they were powerless to halt the
decline, the biologists and managers now are
unable to show in any scientifically acceptable
way that improved management put the herds
on the road to recovery."

In spite of the problems, a wealth of information
on techniques can be used as is or modified slightly
to fit site-specific situations. The brief writeups and
literature cited in this section should give guidance
on where to obtain additional information for either
the biologist with no or limited experience or the
experienced biologist who has need for more de-
tailed information.

Moose

The moose (Alces alces) is a valued trophy and
game animal throughout most of its range. In many
areas (Isle Royale, Denali National Park, Kenai Moose
Range), it is also important as a large, impressive,
and relatively photogenic focus for tourists and
photographers.

Female reproductive performance in moose
populations varies, apparently reflecting levels of nu-
trition which, in turn, can be related to habitat char-
acteristics. Yearling pregnancy rates are especially
variable, ranging from 0 to 62% according to various
studies. Adult pregnancy rates are somewhat less

538

Ungulates

variable, but the twinning rate among adult females
in a population may be an especially good indicator
of habitat quality (Franzmann 1978).

Such reproductive performance indicators
should relate to quite specific habitat areas, because
moose generally have small home ranges (3 to 6 km'
[1.8 to 3.6 mi. ]). In some cases, however, these are
seasonal home ranges, separated by migration corri-
dors of up to 50 km (30 mi.) (LeResche 1974;
Franzmann 1978).

Moose are widely distributed in North America,
extending from coastal areas bordering the Arctic
Ocean to the mixed deciduous/coniferous forests of
Minnesota and Maine and as far south as Utah in
the western montane forests. Dramatic expansions in
the Rocky Mountains have occurred in this century,
as have apparent expansions and increases in tundra
areas of western and northern Alaska. Climatic
changes and a large increase in the proportion of
fire-seral communities are two reasons most fre-
quently cited for these distributional dynamics (Kel-
sall and Telfer 1974).

Average adult body weights for females are 4 14
kg (920 lb) in Alberta and 445 kg (980 lb) in Alaska;
males attain weights of up to 585 kg (1,300 lb)
(Franzmann 1978). From their size, moose are
clearly an ungulate species requiring large amounts
of forage. Adult energy requirements vary, but aver-
age around 14,000 kcal per day (Belovsky 1978).
Although moose have large bite sizes, they have rela-
tively low bite rates. In a study in Alberta, Renecker
and Hudson (1986) estimated that consumption
rates ranged from 23 g/min (0.8 oz/min) in July to
11 g/min (0.4 oz/min) in January.

In Alaska, calves grow from birth weights of
1 1.2-135 kg (25-30 lb) to weights of 180 kg (400

Mature bull moose.

lb ) in 5 months. Their size suggests that forage abun-
dance and quality are critical factors for moose. In-
deed, Belovsky (1978) presents convincing evidence
that moose almost always forage to maintain energy
input. They have little room for error if they are
to achieve positive energy balances.

Habitat Requirements and Analysis. Because of
the high forage demand of moose, habitat evaluation
for the species has concentrated on identifying forag-
ing areas and measuring forage quantity and quality.

Browse comprises 75 to 80% of the moose diets
in winter but becomes less important by varying
degrees in summer. Moose can break off large stems
and branches and exploit the canopy up to 3 m (10
ft) above ground level. Throughout the range of
the species, willows (Salix sp. ) are used out of pro-
portion to their abundance. Other highly palatable
and apparently preferred plants include birch (Be-
tula sp.), aspen (Populus tremuloides), and cotton-
wood (Populus sp. ). When snow is not too deep,
low shrubs (such as Vaccinium sp.) or sedges
(Carex sp. and Equisetum sp. ) may also be used.

Brassard et al. ( 1974) examined several varia-
bles potentially useful in measuring relative habitat
value and winter carrying capacity in Quebec. They
found a direct relationship between degree of occu-
pation and quantity of available deciduous stems.
Later work (Polequin et al. 1977) has shown that in
late winter, cover and reduced exposure from wind
may be as important as forage availability in the se-
lection of sites.

In most parts of their range, moose make signifi-
cant use of aquatic plants in summer. Much of this
use is concentrated at a time when twigs and leaves
of browse species have peak nutritional value. In
terms of energy needs, there is no obvious advantage
for moose switching to aquatics. However, Jordan et
al. (1973) and Fraser et al. ( 1984) concluded that
sodium content is being selected since winter
browse may have been insufficient.

Similarly, natural mineral licks are also impor-
tant sources of sodium for moose (Tankersley and
Gasaway 1983). Although nothing directly links
aquatic forage and mineral licks to moose reproduc-
tive dynamics, such links can be hypothesized, at
least physiologically. Thus, habitat managers should
identify and maintain such features.

For many populations, riparian and floodplain
communities are also preferred habitats, apparently
of importance far beyond their relatively restricted
occurrence. This is certainly true of Shiras moose (A
a shirasi) in the Rocky Mountains (Peek 1974) and
of populations that extend into tundra areas (Mould
1977, 1979).

Ungulates

539

Browse quality has been determined by many
measurement techniques (Oldemeyer 1977a). On
the Kenai Pensinsula, Alaska, deteriorating range
quality has been circumstantially linked with declin-
ing moose numbers (Oldemeyer et al. 1977a). Ani-
mal physiological factors, such as hair and blood
chemistry, have also been used in measuring habitat
quality (LeResche et al. 1974; Franzmann et al.
1975). LaPerriere et al. ( 1980) used Landsat images
and cluster analysis techniques to map moose habitat
classes and features in interior Alaska. Through these
techniques, large areas could be mapped with high
accuracy and little field checking.

A model has been developed to evaluate moose
habitat suitability and quality in Alaska. The model
allows quality to be estimated quantitatively, based
primarily on data from studies in south-central and
interior Alaska. A handbook is available that would
help a biologist apply the model (Konkel 1980, as
cited in Mule 1982). However, Mule (1982) com-
pared the habitat quality ratings generated by the
model with those derived by a team of experts, and
he concluded that the model-based ratings were
unacceptably inaccurate. A revision of the model is
described but not fully tested (Mule 1982).

Population Measurements. Techniques for accu-
rately censusing moose need to be perfected. In
spite of their large size, their habitats and frequent
use of forested cover make them difficult to see from
aircraft, even helicopters. Even though at some times
of the year some populations move into more open
terrain, moose are not sufficiently consistent in such
behavior. Furthermore, seasonal aggregations, when
they do occur, do not involve large segments of
the population.

Helicopter surveys, before bulls drop their ant-
lers in the fall, are useful in obtaining adult sex ra-
tios. Yearling bulls may also be reliably identified
from the air, but yearling females are difficult to
reliably identify (LeResche and Rausch 1974). After
males have shed their antlers, adult females may
be correctly distinguished from the ground by their
white vulvar markings (Lent 1974b). More detailed
information on the age structure of populations can
only be obtained by sectioning teeth and reading
cementum annuli from hunter kills or other dead an-
imals (Sergeant and Pimlott 1959).

Timmermann ( 1974) reviewed other tried tech-
niques in assessing relative population levels of
moose, ranging from track-and-fecal dropping counts
to infrared scanning devices. All these techniques
had factors that severely limited their usefulness.

Discussion. Because moose generally prefer early
successional vegetation, carrying capacity can be
improved by manipulating selected habitats. These
manipulations may include mechanical rehabilitation
(Oldemeyer 1977b), prescribed burns, and special
logging practices (Teller 1974). But Peek (1974)
advised caution and careful planning of such manipu-
lations, particularly in areas where more than one
large herbivore species may be present.

Perhaps the most critical management need for
moose is the development of truly accurate and use-
ful census techniques. Until this happens, biologists
will have difficulty assessing the value of various
management practices, including habitat
manipulations.

Caribou

Despite these problems, most census techniques
still involve aerial surveys, using either transects or
intensive complete counts of small blocks of habitat.
Transects have been discontinued in some areas
because of their extremely low accuracy (Timmer-
mann 1974); in other areas, they are used only as
indexes of relative abundance. In the most intensive
tests of aerial counts to date, conducted over enclo-
sures with known numbers of animals, LeResche
and Rausch (1974) found that only about 68% of
the adults were observed under ideal conditions,
with inexperienced observers seeing only 43%. Both
precision and accuracy were low. Recent experience
with conditions similar to the test area, along with
planning and attention to details, was found to be
essential in obtaining valid trend indicators. Similarly,
Gasaway et al. ( 1985) described significant and
widely variable biases in summer aerial surveys of
moose where even the direction of population trend,
except veiy large changes, could not be reliably
detected.

Populations of caribou {Rangifer tarandus)
roam a variety of northern boreal habitats in North
America, from the barren polar deserts of the Cana-
dian Arctic Archipelago southward, occupying much
of Alaska and extending into woodland and alpine
areas of southern Canada. A small remnant popula-
tion moves into the contiguous 48 states in the area
of the Idaho Panhandle during part of the year. The
species is typically divided into two groups of sub-
species: the tundra (cylindricornis) group and the
woodland (compressicornis) group (Banneld 1962).
Ecologically, however, at least four categories should
be considered:

( 1 ) Small groups that remain in tundra habitats
throughout the year;

(2) Large migratory populations, some portions
of which typically use taiga habitats in the
winter months;

540

I 'ngulates

if

Caribou.

(3) Mountain caribou that make marked altitudi-
nal migrations seasonally;

(4) Woodland caribou that generally make only
short annual movements.

Generally the woodland populations are small, show-
ing one of the latter two strategies. There are excep-
tions however. In Labrador, populations of woodland
caribou have reached nearly 200,000 animals and
migrate over 200,000 km2 (80,000 mi.2) of tundra
and taiga (Juniper 1980; Parker 1981).

Reindeer, domesticated members of the same
species, have been introduced into North America,
primarily from Siberia. Herds exist in Alaska and
in the Mackenzie River Delta of Canada. Management
as domesticated stock continues in some areas to
varying degrees. In other areas, reindeer are now
feral, and some have mixed and interbred with na-
tive subspecies.

Habitat Requirements and Analysis. Surprisingly,
few biologists agree on the habitat requirements of
caribou, despite an extraordinary amount of research
and management concern over the past decades.
Much controversy in recent years has focused on the
importance and need of lichens in the winter diet
of caribou. Populations of the high arctic polar de-
sert and some introduced populations make little use

of lichens and yet they thrive. Klein (1982) argued
that these exceptions in no way excluded or refuted
the possibility that the species' evolution may be
closely linked to a lichen-based niche, unexploited
by other herbivores. Lichens, as winter forage, are
low in protein but high in digestible carbohydrates.

Late winter-spring fat deposition may occur
on high-quality lichen ranges. But clearly, caribou,
like other northern ungulates, adapt to long periods
of low-quality forage in winter and indeed do not
normally maintain body weight in winter even on a
relatively high-protein diet (McEwan and Whitehead
1970, 1975). However, actual examples of starvation
or extreme malnutrition among adult caribou in
winter or spring are rare, although it has occurred in
some arctic island populations. Bergerud (1978:98)
points out that —

"It has not been demonstrated that an absolute
shortage of food has caused caribou declines
.... I do not know of any studies where it
could be shown that birth and death rates were
altered due to an absolute shortage of food."

However, the massive declines and die-offs that have
occurred in insular populations demonstrate that,
under certain circumstances, forage can be abso-
lutely limiting (Klein 1968).

Ungulates

541

Thing and Clausen (1980) attributed a high
mortality among calves in Greenland to overuse and
depletion of lichens on winter ranges. Leader-Wil-
liams (1980) also attributed alterations in reproduc-
tive performance to habitat overutilization.
Bergerud's analysis ( 1978), emphasizing hunting and
predation as the key mortality factors in caribou
population dynamics, tended to ignore the possible
role of habitat quality and nutrition in mediating
mortality factors.

Most specialists have stressed the importance of
identifying and protecting the calving grounds of
migratory caribou populations. Nevertheless, there
has again been no agreement as to what critical habi-
tat requirements are linked with such calving
grounds. Isolation from human centers (Lent 1966)
and phenological characteristics ( Kuropat and Bryant
1980; Lent 1980) have been suggested as important.
Kelsall (1968) believed that most Canadian calving
grounds were extremely harsh with poor forage, and
stressed the low density of predators as a factor.
Reimers et al. (1983) showed that females on early
greened calving grounds had less weight loss during
lactation than those still on winter forage at calving
time.

As noted above, specific habitat requirements of
caribou are controversial issues, particularly in win-
ter ranges and calving areas. In addition, many now
believe that the quality of winter forage and, within
broad limits, quantity of winter forage are of rela-
tively little consequence in terms of caribou demog-
raphy. In contrast, managers of domestic reindeer
continue to be concerned about winter range condi-
tion techniques for evaluating these ranges (Eriksson
1980). Workers with woodland caribou have also
developed techniques for assessing arboreal lichen
abundance (Van Daele and Johnson 1983). The rela-
tive availability of forage, as influenced by patterns of
snow cover and ice accumulation, have also been of
great interest in terms of wild and domestic popula-
tions. These techniques are described in Chapter
8, Tundra.

A broad concensus has developed that Rangifer
habitat quality is presently best measured by examin-
ing population attributes. Parker ( 1981 ) measured
both physical and reproductive characteristics in
females, consistent with an expanding population.
Reimers et al. ( 1983) noted several attributes they
believe highly correlate with spring-summer range
quality.

cover is essential for accurate winter counts. Berge-
rud provided many other recommendations on cen-
sus design and execution. Siniff and Skoog (1964)
reported on the use of stratified random sampling
with aerial transects.

In the early 1960s, specialized aerial census
techniques for estimating the size of migratory cari-
bou populations were developed in Alaska. These
techniques take advantage of specific behavioral attri-
butes of migratory caribou: their use of calving
grounds at relatively fixed locations, where almost all
adult females converge for a brief calving season,
and the typical concentration of these females and
other individuals into extremely dense aggregations
during an even briefer post-calving period.

Skoog ( 1962) pioneered the calving-ground
count method and Lent ( 1966) first described the
post-calving aerial photography method. Because
adult males are largely segregated from cows and
widely dispersed at these times, both methods re-
quire later samples to obtain adult sex ratios needed
to determine total population estimates. Thus, the
current technique has commonly been referred to as
the "aerial photo-direct count extrapolation"
method. Such ratios are best obtained from samples
at several locations during the rutting season.

Modern refinements and analyses of these basic
methods have been described in several Pittman-
Robertson (Federal Aid) reports of the Alaska De-
partment of Fish and Game (Pegau and Hemming
1972). These reports are discussed in detail by
Doerr(1979).

Canadian workers have continued to make more
use of winter and spring censuses (Parker 1975).
Aside from using such censuses for total population
estimates, they are universally important in deter-
mining relative use of winter ranges; estimating rela-
tive mortality rates (Davis and Valkenberg 1979);
and assessing vulnerability to hunting pressure, much
of which occurs on winter ranges.

Calef and Heard ( 1980) used aerial transects to
sample and derive population estimates for three
caribou populations. They used a 25% correction
factor for animals overlooked, citing empirical evi-
dence for a similar correction factor of 29% . They
were able to compare the transect estimate with an
estimated calving ground count for the same popula-
tion. The two estimates were 17,225 and 15,884,
respectively, for the Lorillard herd.

Population Measurements. The first aerial cen-
suses of caribou relied primarily on winter and
spring transect data. Bergerud (1963) presented one
of the best early treatments of such aerial censuses,
their problems and uses. He noted, as did Lent
( 1966), that a low but continuous, unbroken snow

Discussion. Many census techniques have been
developed for obtaining good numeric and composi-
tion data for caribou populations, particularly for
those inhabiting tundra and open woodland environ-
ments. The ability to meaningfully relate habitat
quality and trend to population dynamics of Rangifer

542

Ungulates

continues to be marred by a general lack of agree-
ment. Sampling of reproductive and physiological
attributes of individual animals holds the most prom-
ise for monitoring populations and, indirectly, the
status of their seasonal ranges.

Pronghorn

Pronghorn antelope (Antilocapra americana)
symbolize the wide open rangelands of the West.
The range of pronghorn during the early 1 800s in-
cluded most of the Great Plains, the high sagebrush
steppes and grass valleys of the Great Basin states,
and parts of south-central Canada and northern Mex-
ico (Einarsen 1948). By the 1920s, however, the for-
mer densely populated ranges had been drastically
reduced. Ranges occupied by pronghorn today are
virtually identical to historical ranges except areas
along the Mississippi River. Reintroductions since the
1920s are largely responsible for increased popula-
tions of this swift ungulate.

Limiting factors for pronghorn herds include
predation; severe winters with deep snow; fences
that preclude seasonal movements; and neonatal
mortality, possibly caused by a lack of adequate nu-
trition for pregnant does during the last 3 months of
pregnancy. Disease does not appear to be a signifi-
cant mortality factor in pronghorn (Yoakum
1978:107). Management concerns about pronghorn
are related to limiting factors and include coyote and
bobcat predation on newborn fawns; road kills on
high-speed roads; wire fences constructed in ways
that preclude pronghorn movements, especially in
winter; and decreasing availability of diversified vege-
tative ranges.

Habitat Requirements. Pronghorn typically oc-
cupy habitats characterized by low, rolling, wide-
open, expansive terrain. Sites with sparse stands of
conifers may be used at certain times if understory
vegetation allows distant visibility and rapid mobility
(Yoakum 1978).

Habitat requirements for pronghorn are de-
scribed in detail in Autenrieth ( 1978), Yoakum
(1980), Kindschy et al. (1982), and Allen et al.
(1984).

Free-standing water, available year-round, ap-
pears necessary for producing and maintaining high
pronghorn densities. This water should be available
every 1.6 to 8.0 km (1 to 5 mi.). Studies by Sund-
strom (1968) in Wyoming indicated that 95% of the
observed pronghorn were within a 4.8 to 6.4 km (3
to 4 mi. ) radius of water. Highest densities of prong-
horn occur in habitats averaging 25 to 38 cm (10
to 1 5 in. ) of precipitation per year. Where prong-
horn have been transplanted to areas of higher pre-
cipitation, production and survival rates declined.

•*4Bi

Mature pronghorn buck.

Pronghorn also occur in areas of lower precipitation,
but population densities are less (Yoakum 1978).

Most pronghorn ranges receive some snow;
however, when snow accumulations exceed 25 to
30 cm (10 to 12 in.), pronghorns have trouble find-
ing forage. Prolonged winters with deep snow are
especially harmful when combined with factors such
as —

( 1 ) Low quantities or qualities of forage,

(2) Excessive wind, and

(3) Man-made obstacles that impede or restrict
free movement to areas with less snow cover.

Temperatures do not seem to be a significant limit-
ing factor, as pronghorns can adapt to hot deserts or
alpine plateau conditions.

Quality and quantity of vegetation appear to be
the major factors affecting pronghorn population
densities. In the sagebrush-grassland areas of the
Great Basin states, the following vegetative charac-
teristics are found on ranges preferred by pronghorn
(from Kindschy et al. 1982):

• ground cover averaging 50% living vegetation;

• general range composition of 40 to 60% grass,
10 to 30% forbs, and 5 to 20% browse;

• a variety of species, within a vegetative commu-
nity, including 5 to 10 species of grasses, 20 to
40 species of forbs, and 5 to 10 species of
shrubs;

• succulent plants, available in spring and wet
summers;

• open, rolling rangelands having a variety of veg-
etative types rather than monotypic vegetative
communities;

Ungulates

543

• low vegetation structure averaging 38 to 6 lcm
(15 to 24 in.) in height.

Most of these preferred habitat qualities would also
apply to pronghorn ranges east of the Continental
Divide.

West of the 100th Meridian, pronghorn typically
inhabit ranges characterized by low rolling, expan-
sive terrain and are almost never observed for more
than a few minutes at a time where their view was
restricted by terrain or other natural features (Allen
et al. 1984). Kindschy et al. (1982) indicate that
areas with slopes less than 5% were best for prong-
horn in this region.

Microhabitats provided by topographic relief in
this region apparently increase habitat quality during
winter by providing lower wind velocities, less
snow, and less dense snow. During the fall and win-
ter, pronghorn spent more time in basins less than
1.6 km (1 mi.) in diameter than at any other time of
the year (Prenzlow et al. 1968).

Pronghorn typically forage heavily on browse
and forbs and only eat grass when it is green and
succulent. Thus, pronghorn food habits do not gen-
erally overlap those of horses and cattle. However,
pronghorn food habits are similar to those of domes-
tic sheep, and pronghorn may compete with sheep
for available forage.

Wheat ( Triticum aestivum ) is a major portion
of the diet of pronghorn living near winter wheat
fields in Colorado (Hoover et al. 1959). Heavy use of
wheat by pronghorn has also been reported from
Kansas, Montana, and Alberta (Allen et al. 1984).
Winter wheat near or interspersed with rangeland is
assumed to improve the winter food value for prong-
horn if shrubs are present at densities of 75% or
less (Allen et al. 1984).

Fences on pronghorn ranges west of the 100th
Meridian may restrict movements and cause direct
injury or mortality (Allen et al. 1984). Fences may
cause significant adverse impacts when constructed
across migration routes or where they interfere with
daily movements to and from water.

Population Measurements. Most pronghorn are
counted from the air, although some ground counts
are used on sparsely inhabited ranges. Ground sur-
veys are time-consuming and expensive, but they
provide accurate data on fawn survival.

Aerial counts, usually by fixed-wing aircraft,
cover larger areas in less time. Aerial surveys are
conducted during the summer and fall in some
states, and during the winter in others. Winter
counts are usually flown to obtain total numbers, as
fawns are hard to distinguish from yearlings in win-

Twin pronghorn fawns.

ter, and most bucks have shed their horn sheaths by
the time winter aerial counts are made.

If sex and age ratio data are needed, counts are
made in late summer or early fall. Flight patterns
should be either sample strips (usually 0.8 km
[0.5 mi.] on each side of the aircraft and 1.6 km
[1 mi.] apart) or in randomly located (1.6 km"
[1 mi."]) quadrats. Most flying for pronghorn census
should be approximately 46 to 91 m (150 to 300 ft.)
aboveground with speeds of about 112 to 128 km
(70 to 80 mi.) per hour. Usually 30% of the area is
sampled by a strip census while 10% is sampled by
quadrats. The quadrat sample method usually takes
twice as long as the strip census (T.M. Pojar, pers.
commun. ).

Discussion. The pronghorn is unique in that it is
easy to observe during all seasons in all its preferred
habitats; therefore, censuses are relatively easy. Cen-
sus techniques vary, however, from Great Plains
grassland habitats to Great Basin sagebrush habitats.
Inventory and monitoring of pronghorn habitats is
also relatively easy since their habitat requirements
are relatively similar throughout their range, have
been well-described, and can be measured easily.

Bison

The American bison (Bison bison ) today is
being raised in predominantly domestic situations.
With few exceptions, free-rar ging herds of bison are
no longer found. Only in the vicinity of Wood Buf-
falo National Park in Canada and in Yellowstone
National Park in Wyoming can free-ranging bison
populations be found that are neither cropped nor
hunted. Free-ranging herds in Utah and Alaska are
intensively managed and cropped. In recent years,
the commercial value of bison has increased. Numer-
ous private herds are managed as livestock, with

544

Ungulates

American bison (bull, cow, and calf).

meat being sold as a specialty item, and hides, heads,
and other parts finding a ready market.

Natural limiting habitat factors for the American
bison are not of great concern because most of to-
day's herds are intensively managed, which includes
disease control and selective culling. Historically, the
grizzly bear and the wolf were probably the only
natural predators that affected bison (Meagher
1978). Because of the present-day, low population
levels of these two large predators, they no longer
significantly threaten free-ranging bison herds.

Large-scale mortality from adverse winter
weather has been reported historically. Today, only
the herds in Yellowstone National Park suffer any
significant winter losses (Meagher 1978). Winterkill
usually results from the combined effects of forage
availability, climatic stress, and physical condition of
individual animals. These losses usually occur in
late winter and early spring after prolonged severe
winters. Some calf mortality has been reported dur-
ing severe spring weather.

Habitat Requirements and Analysis. The primary
habitat requirements for bison on public lands are
adequate forage, water, and space.

Bison are primarily grazers. Grasses furnish most
of their food over the year; however, sedges make
up a large portion of their diet in both Wood Buffalo
and Yellowstone National Parks. Bison seem quite
similar to cattle in their forage preferences and, as a
result, these two animals often directly compete
with each other. Peden et al. (1974) compared the
diets of bison and domestic cattle in a short-grass
plains area and found that bison prefer warm-season
grasses, feed in different areas, and are less selective.
Bison appear to be able to digest low-protein, low-
quality forage and apparently can eat more of this

type of forage. Bison appear to need water every
day.

If bison are to be maintained as free-ranging
animals, they will need large areas to roam. How-
ever, they can also be successfully kept in small pas-
tures as semidomestic animals.

When available, bison use forested areas for
shade and to escape insects (Meagher 1978). These
areas also furnish forage in deep, crusted snow be-
cause snow is more loosely packed in forested areas.
During severe weather, forest areas and some topo-
graphical features furnish cover (Meagher 1973,
1976).

The bison can thrive in places where no other
large ungulate can. This includes open valleys cov-
ered with deep snow and subject to frequent, severe
wind and blizzard conditions, lasting up to 6 months
(Meagher 1976).

Population Measurements. Bison population lev-
els are best determined by aerial censuses rather
than sample estimates (Meagher 1978). These counts
are most effective when flown during the winter.
Herd and calf production were successfully counted
through the use of aerial photography in Wood Buf-
falo National Park (Stelfox 1976).

Discussion. Free-ranging American bison are found
on only a few areas of public domain land (Utah
and Alaska). Present-day herds are small, easily lo-
cated and tracked, and easily censused. Any standard
vegetation measurement technique can be used to
determine range use and to monitor condition and
trend of bison ranges.

This species is important from a historical and
aesthetic standpoint. Free-ranging populations will
never be large, but proper management of present-
day, free-ranging herds should not be neglected.

Mountain Goat

The mountain goat (Oreamnos americanus) is
one of the least understood of native North Ameri-
can ungulates. Relatively few studies have been con-
ducted on mountain goats because of their limited
geographic range and comparatively low economic
importance. Further, the species has been difficult to
study because its habitats are severe and often
remote.

Most studies of mountain goats are recent. A
large portion of literature on the goat can be found
in the proceedings of a special symposium (Samuel
and MacGregor 1977) and the Biennial Symposium
of the Northern Wild Sheep and Goat Council ( He-
bert and Nation 1978; Hickey 1980; Bailey and
Schoonveld 1982 ).

Ungulates

545

Rocky Mountain goat; mature billy.

Habitat Requirements and Analysis. The basic
habitat requirements of mountain goats are —

• suitable topography for predator evasion along
with a sufficient microclimate,

• forages for all seasons,

• salt, and

• water.

For management purposes, winter ranges, summer
ranges, kidding areas, salt licks, and migration corri-
dors need to be identified.

Mountain goat winter ranges are characterized
by their lack of persistent or melt-crusted snow
along cliffs and steep terrain interspersed with vege-
tation. Suitable winter ranges may be (1) at lower
elevations (Rideout 1974; Smith 1977), including
coastal areas (Hjeljord 1973; Hebert and Turnbull
1977; Fox and Taber 1981) where snow is less abun-
dant and persistent, or (2) on relatively unforested,
steep, mostly south-facing slopes where snow sheds
rapidly (Brandborg 1955; Hjeljord 1973). Mountain
goats mostly use those portions of winter ranges that
are on slopes exceeding 40° (Smith 1977; Smith and

Raedeke 1982 ). Usually, these slopes face south, al-
though Adams and Bailey (1980) described some
goats in a southern population wintering on north
aspects. Overhanging ledges, caves, and abandoned
mine entrances may provide additional shelter for
wintering goats (Richardson 1971; Smith 1977).

In winter, goats become sedentary on tradition-
ally used winter ranges (Smith 1976; Adams and
Bailey 1980; Chadwick 1983), with forages on ledges
or benches within cliff areas or on windswept ridges.
One small population of goats in a forested area
rarely foraged more than 50 m ( 165 ft) from escape
terrain (Smith 1982). This may have been an excep-
tion, related to the small numbers of goats, the poor
visibility in their forested habitat, and the presence
of large predators. In contrast, Fox and Taber (1981)
observed that wintering goats made little use of
areas more than 300 m (990 ft) from steep broken
terrain; Schoen and Kirchoff ( 1981 ) reported over
90% of observed goats were within 400 m ( 1,320 ft)
of cliffs.

At low elevations, particularly near the Pacific
Coast, a conifer canopy may benefit wintering moun-
tain goats by intercepting and redistributing snow
and by providing forage, including arboreal lichens.
In contrast, goats wintering on interior ranges,

546

Ungulates

where snowfall is great, tend to avoid dense stands
of conifers that accumulate snow (Adams and Bailey
1980).

When snow is not an overriding problem, habi-
tat selection is most often determined by needs for
security from predation and for abundant, nutritious
forage. Security is probably a function of distance
from escape terrain, visibility afforded by the habitat,
and group size (Adams et al. 1982a; Risenhoover
and Bailey 1985a). Thompson (1980) noted the aver-
age distance of goats from escape terrain on his
study area was 75 m (248 ft). Schoen and Kirchoff
(1981) found that distance to cliffs was the most im-
portant factor determining goat distribution and that
summering goats made little use of foraging areas
over 400 m (1,320 ft) from cliffs. McFetridge (1977)
also noted that most foraging by nursery bands was
within 400 m (1,320 ft) of escape terrain. Risen-
hoover and Bailey (1985a) suggest that nanny goats
form large nursery bands on alpine tundra, where
visibility is excellent. This may be a strategy for re-
ducing the risk of predation while exploiting forages
far from escape terrain. This strategy is only effective
where goat populations are relatively large and
wolves are absent.

Goats will exploit phenological differences
among habitats to obtain the most nutritious forage.
They often move to lower elevations in spring, seek
green forage at higher elevations in summer, and
feed on north aspects in late summer ( Brandborg
1955; Rideout 1974; Smith 1976; Bailey and Johnson
1977). On some ranges, moist meadows below
cirques or snowfields are attractive foraging sites in
late summer.

Visibility of surroundings seems important to
mountain goats (Chadwick 1983). This may limit
their use of dense stands of conifers far from escape
terrain. In contrast, goats often use sparse stands of
conifers (Smith 1976; Adams and Bailey 1980) and
have used conifers for avoiding people (Singer 1975).
The value of conifers on summer range is unclear,
but may involve associated forage, available water or
salt, or an attractive thermal regime.

Kids are born on the steepest, most rugged
areas of a goat range. Based on a few reports, kidding
areas are usually within winter ranges and used year
after year (Smith 1976; Adams 1981).

Judging by the frequency and tenacity with
which goats use licks, these areas seem to be very
important habitat resources. Peak use of licks occurs
in spring and early summer (Singer 1975). The prin-
cipal attractant at licks probably is sodium (Hebert
and Turnbull 1977). On a Colorado study area, goat
movements seemed limited by licks occurring at
only one end of the area. These summer movements

expanded after a new salt lick was established at
the other end of the area (Bailey, J.A., unpubl. data).

Migration corridors are used between winter
and summer ranges and when goats are visiting salt
licks. When crossing forested areas without escape
terrain, goats repeatedly use the same trails (Singer
1975 ). Presumably the risk of unexpected predator
attack can be great in these areas, since visibility
may be poor and the frequent occurrence of goats
may attract predators. If this is correct, reducing
conifer cover to enhance visibility could improve
goat habitat on migration corridors.

Water is not believed to be a limiting factor on
most mountain goat ranges as these are either in
moist climates or in areas with persistent snowfields.
However, water availability may restrict goat move-
ments and habitat selection in southern ranges
where the species has been introduced.

Food habits of mountain goats have been re-
viewed by Hibbs (1966), Wigal and Coggins (1982),
and Adams and Bailey (1983). Dailey et al. (1984)
used tame animals to evaluate forage preference.
Mountain goats are such adaptable feeders that litera-
ture on their food habits allows few generalizations.
Reported diets have varied depending on locations
and habitats, season, and snow conditions. Grasses
and sedges, forbs, and browse have each been re-
ported as abundant in the diet or as preferred in
both winter and summer studies. Conifers or mosses
and lichens have been important winter forages.

The ability of mountain goats to exploit a wide
variety of forages may be an adaptation compensat-
ing for their narrow habitat preferences, especially
during winter (Adams et al. 1982b). By using all
available forages, goats are able to exist in small
groups on small, snow-shedding areas where they are
protected from predators. Wintering goats seem to
choose areas on the basis of a sedentary life-style
rather than for abundant forage. In contrast, during
summer when nutritional demands for growth and
lactation are high, goats benefit by choosing lush
stands of quality herbaceous forage in alpine grass-
lands and meadows (Risenhoover and Bailey 1985a).

Habitat types used or preferred by mountain
goats on winter and summer ranges have been de-
scribed in several studies in diverse locations (Smith
1976; Adams and Bailey 1980; Fox and Taber 1981;
Schoen and Kirchoff 1981). In these studies, habitat
classifications have involved overstory vegetation,
understory vegetation, and terrain characteristics (as-
pect, steepness, brokenness). While goats use a vari-
ety of vegetation types, their habitat selection has
been most consistently related to steep terrain.

Ungulates

547

Fox et al. (1982) developed a model for predict-
ing presence/absence of mountain goats in a coastal
range of southeast Alaska. The principal predictor
was distance to cliffs. This model apparently has not
been tested in other areas. Useful models of goat
habitat quality, yet to be developed, will probably
depend on the following factors, in order of impor-
tance:

( 1 ) Distance to cliffs or slope characteristics;

(2) Elevation and aspect, including available
ranges of these variables; and

(3) Vegetation type.

The influences of these factors vary among areas
within the geographic range of the species. Another
important factor may be presence and location of
salt licks. While such a model may predict the quan-
tity and quality of seasonal goat ranges, additional
consideration may be needed to identify migration
corridors.

Population Measurements.

Relative Density. Trend counts, especially
from aircraft, are the most commonly used methods
for monitoring mountain goat populations. In
repeated counts, more goats have been observed
from helicopters than from fixed-wing aircraft
(Ballard 1977; Nichols 1980).

Trend counts provide more than annual popula-
tion indexes. They provide herd productivity (age
ratios) and minimum population size indexes that
may be used as bases for conservative harvest strate-
gies. However, the precision of trend counts has
not been evaluated. Many successive surveys of one
population in one season still need to be done. With-
out measuring precision, the significance of a differ-
ence between trend counts over 2 years cannot be
evaluated. Likewise, the probability of detecting a
significant change in population size from trend
counts in consecutive years is uncertain.

A subjective appraisal of Nichols' (1980) data
from replicated aerial counts of local populations,
from July to mid-September and with consistent
weather conditions, suggests that the 90% confi-
dence limits for a single count from a fixed-wing air-
craft might vary by ± 18% of the count. If this is
correct, there is only a 25% probability of detecting,
with 90% confidence, an 18% change in population
size, based on trend counts from consecutive years.
This statement must be used with caution because
Nichols' (1980) study was not designed to measure
precision in this way, and precision may vary greatly
among local populations. However, Nichols' data
illustrate that aerial trend counts of mountain goats
may not be sensitive to fairly significant changes

in goat abundance, even when flying and observing
conditions are selected for consistency. The possibil-
ity that ground-based counts will be equally impre-
cise should be evaluated.

In mostly forested areas, early spring counts of
goats on winter ranges have been favored ( Smith
1976; Bone 1978). Counts are best made after most
snow has melted and goats are easier to observe, but
before the animals leave their winter ranges. Succes-
sive trend counts may be difficult within this short
period.

Midsummer has been preferred for trend counts
in areas where goats form large bands on alpine tun-
dra (Ballard 1977; Hall 1977; Nichols 1980). Large
bands may develop after the kidding season and can
more easily be observed after most snowfields have
melted. Nichols (1980) preferred overcast skies, soft
light, and no turbulence for these aerial counts.
Adult male goats, which are more solitary than other
sex-age classes, may be under-represented in counts
taken in midsummer ( Risenhoover and Bailey 1982a;
Foster 1982).

Census. A known minimum population of
mountain goats may be determined by using the
largest number of goats seen in several successive
counts. If animals are classified, the cohort-comple-
tion method may provide a larger known minimum
population. With this method, the largest number of
goats seen in each age class (rarely, sex-age class)
is determined and the numbers summed (Smith
1976). Known minimum populations are used as
bases for conservative harvest strategies.

Use of the highest count from several successive
counts within a season may enhance precision for
detecting among-years trends in goat numbers (year-
to-year variation among highest counts may be less
than among single counts made each year). In prac-
tice, the precision of the annual highest count
method has been evaluated subjectively. Accuracy is
unknown, but actual populations must be underesti-
mated in almost every case.

Adams and Bailey (1982) used marked goats and
the Peterson Estimator to census a Colorado herd.
With this method, accuracy will be enhanced if ani-
mals from all herd segments are marked. Adult males
and females should be marked proportionally to
their occurrence in the herd (Risenhoover and Bai-
ley 1982a). The possibility of marking goats in more
than one geographic segment of the herd should also
be considered. The second sample used to estimate
the proportion of marked goats in the herd should
be based on observations throughout the study area.

Marking a large proportion of a goat herd will
seldom be possible. Consequently, the precision of a
population estimate will be poor and confidence

548

Ungulates

limits (Bailey 1951) high whenever a single second
sample is used. Precision can be enhanced indefi-
nitely by using numerous second samples, resulting
in replicated population estimates. With replications,
the population is estimated by the mean of succes-
sive estimates, and confidence limits are based on
the number and variance of the estimates, using
common t-statistics.

Animal Condition. Foster (1978) reviewed the
concepts of animal condition and population quality
in relation to mountain goats. Presumably, measures
of animal condition commonly used for ungulates,
such as body size, growth rates, skeletal ratios, and
measures of fat reserves, can be applied to goats.
However, these methods have seldom been used,
partly because large samples of goat carcasses have
seldom been available. It would be useful to obtain
and report such data so that inter- and intrapopula-
tion variation can be evaluated.

Foster (1978) evaluated horn dimensions as
possibly useful measurements of population quality
in mountain goats. Results were largely inconclusive.
The most useful horn dimension measurements may
be (1) horn growth in the first 2 years of life (which
can be measured in goats over 2 years old because
the second horn annulus is distinct) and (2) a ratio
of horn length to ear length for kids and yearlings,
when data are collected in the same month each
year. For 2-year old goats, at least, horn lengths are
different between sexes. Thus, populations should be
compared by using data from one sex for each com-
parison. Lastly, intrapopulation variation of horn
growth may be so large that the horn measurements
for comparing populations will be invalid.

distinguished by their advanced molt in midsummer
(Nichols 1980) but may be under-represented in
samples taken during this time (Foster 1982; Risen-
hoover and Bailey 1982a).

Small, randomly taken samples from a goat herd
will provide highly variable estimates of age struc-
ture. Thus, biologists should not rely on classification
data unless a large proportion of the herd can be
classified.

When data are limited, the simplest age ratio is
the kid to older goat ratio (where older goats in-
clude yearlings, 2-year olds, and adults of both
sexes). This ratio is not a precise indicator of cur-
rent reproductive success since variation in the sex
ratio and in the numbers of nonbreeding older goats
will distort the ratio. However, this ratio has re-
flected herd age (recently transplanted herds having
higher ratios) and winter weather (Bailey and John-
son 1977; Adams and Bailey 1982). The ratio has also
been used to estimate numbers of kids in estimated
herd sizes (Adams and Bailey 1982).

With more detailed classifications of goats in
local herds, population models may be constructed.
Nichols (1980) used aerial surveys in spring to esti-
mate the proportion of yearlings and aerial surveys
in summer to estimate herd size and the proportions
of kids and adult males. He then used data from suc-
cessive years to simulate herd dynamics. Simulations
provided estimates of the numbers of 2 -year olds in
a herd each year. By subtraction, the number of
breeding-age females was estimated, and a kid to
breeding-age nanny ratio determined. Nichols sug-
gested this ratio will be correlated with population
trend. The accuracy of this approach was untested.

Population Structure. Mountain goat females
are not expected to breed before they are 2V2 years
old, nor kid before they are 3 years old. Conse-
quently, classification by sex and by four age classes
is desirable, especially if a ratio of kids to breeding
nannies is to provide an index of current production.
However, such detailed classification is practical
only in intensive, local studies. Not only are year-
lings and, especially, 2-year olds difficult to distin-
guish in some situations, but a lack of strong sexual
dimorphism further complicates classification Ni-
chols (1980), Chadwick (1983), and Smith (unpubl.
ms.) have provided descriptions of the eight sex-age
classes of goats and Foster ( 1978 ) illustrated their
horn characteristics.

The difficulty in distinguishing age classes of
goats varies with season and with the distance and
duration of observation. Often, midsummer is best
for distinguishing kids and yearlings from older
goats. Kids usually become more visible in July, and
yearlings are especially easily distinguished by size
alone until about September. Adult males may be

*j&£

Rocky Mountain goat; nanny and kid.

Ungulates

549

Behavioral Indexes. Petocz (1973) suggested
that conflicts (agonistic behavior) among mountain
goats may indicate the degree of resource depriva-
tion. Risenhoover and Bailey (1985a) speculated that
average group size may be correlated with forage
abundance and continuity.

Discussion. A primary management concern is
overharvest of mountain goats when access to their
habitats has been improved by new roads (Foster
1977; Kuck 1977a,b; Adams and Bailey 1982; Chad-
wick 1983). Appropriate levels of harvest could be
achieved if (1) annual data on population size and
productivity are known, (2) harvest levels can be
regulated on a local basis, and ( 3 ) illegal kill can be
controlled. However, these requirements are often
impractical.

The effects of removing vegetation from moun-
tain goat habitat, such as by logging or fire, are un-
clear. In addition, the effects may differ between
coastal and interior ranges (Hebert and Turnbull
1977), especially on winter ranges, where detrimen-
tal effects would most likely occur. Near the Pacific
Coast, for example, forest removal may be detrimen-
tal to winter ranges.

Interior winter ranges often support sparse
stands of trees or shrubs that are used for forage.
The steep slopes of these winter ranges are often
used by mountain goats for their snow-shedding
characteristics (Kuck 1977a; Adams and Bailey
1980). Removing forage in these areas may affect
forage resources, shelter, or snow-shedding
characteristics.

Some goat populations are small and isolated
(Smith and Raedeke 1982; Smith 1982), and gene flow
among such populations seems infrequent. Any hu-
man activities (e.g., harvest, habitat alteration) may
inhibit goat movements, especially among males,
thus increasing possibilities for inbreeding.

Mountain goats are known to travel long dis-
tances to visit natural licks or artificially established
salt licks (Singer 1975; Hebert and Turnbull 1977;
Thompson and Guenzel 1978; Adams et al. 1982).
The possibility that a sparse distribution of licks may
limit the distribution or productivity of goats has
not been tested.

Muskox

The muskox (Ovibos moschatus) is a boreal
species whose numbers and distribution were greatly
reduced in the 1800s and early 1900s. Since then,
the management of muskoxen has been a conserva-
tion success story. Hunting prohibitions and success-
ful transplants have restored populations to the point
where limited harvests are now possible in Alaska,
Canada, and Greenland. The species has also been a

focus of study in efforts to produce a new domestic
animal for rural economies of circumpolar nations
(Lent 1971, 1978; Coady and Hinman 1984; Wilkin-
son and Teal 1984).

Muskoxen live 20 years or more and were once
thought to have extremely low reproductive rates.
New evidence demonstrates that while rates are
relatively low compared with ungulates that may
produce twins or triplets, they are considerably
higher than was once thought (Lent 1978).

Habitat Requirements and Analysis. The native
range of muskoxen is primarily in the high arctic,
a zone distinguished by very low precipitation, shal-
low snow cover, and patchy distribution of snow
from winds. Many areas where muskoxen have been
introduced are subject to much greater winter snow
accumulation or maritime climates, having icing
conditions in winter or spring (Lono I960; Spencer
and Lensink 1970; Lent 1974b, 1978). Thus identifica-
tion, monitoring, and protection of suitable winter
habitat have been important and will be increasingly
important as introduced populations expand.

Riparian areas with shrubs are preferred summer
habitats in Alaska (Lent 1978; Robus 1984). Wet hab-
itats with sedges (Carex sp.) also receive heavy use
in summer. Both Parker and Ross (1976) and Thing
(1984) contend that such meadows are important to
this species for weight gain and development of fat
reserves in summer. In general, caribou and musk-
oxen rarely compete in areas where they occur to-
gether (Wilkinson et al. 1976; Thomas and Edmonds
1984).

Musk-ox herd on tundra.

550

Ungulates

Few measurements of muskox habitat have con-
clusively related to habitat quality. The importance
of wet meadows and riparian strips has already been
referred to. In the arctic, remote sensing techniques
provided opportunities to map such features on a
large scale (U.S. Department of the Interior, Fish and
Wildlife Service 1983). Home ranges that provide a
variety of habitats in summer seem important be-
cause muskoxen can take advantage of successive
peaks in forage quality.

In winter, forage availability is the key habitat
requirement. Regions with low precipitation and low
and stable temperature regimes are crucial. Methods
for monitoring and mapping tundra vegetation and
for mapping and measuring snow-cover features have
been treated in Chapter 8, Tundra.

Cementum annuli in incisors have been used to
determine age at death. Lent ( 1 974c ) concluded that
this technique was fairly reliable. However, Gron-
quist and Dinneford (1984), working with 18
known-age specimens, half of which were from ani-
mals raised in captivity, concluded that the tech-
nique underestimated true age and had low
reliability. Lent (1974c) and Parker et al. (1975)
have used femur marrow lipid analyses to assess lev-
els of malnutrition and starvation in populations.

Discussion. The most important management con-
sideration for muskoxen is the identification and
protection of suitable winter habitat. This must be
accomplished before muskoxen are transplanted.
Protection of such areas of stable winter conditions
and forage availability is crucial to long-term man-
agement success (Thomas et al. 1981).

Population Measurements. Since the early studies
by Vibe (1958, 1967) and Tener (1965), through the
work of Spencer and Lensink (1970), up to the most
recent censuses (Lassen 1984), the total count
method has been generally used in obtaining muskox
population estimates. The reasons for this are sev-
eral: muskoxen occur in relatively low numbers;
they are large and conspicuous in typically treeless
habitats; and in winter, they are highly clumped in
their distribution and comparatively sedentary. In a
few cases, sampling on transects has been used to
estimate populations (Gunn et al. 1984).

Total counts provide little opportunity to assess
sample error except through successive counts over
short periods. Spencer and Lensink (1970), working
from snow machines on Nunivak Island where snow
conditions lead to extreme clumping of the popula-
tion, believed they counted essentially the entire
population.

Such ground counts and even low-level aerial
surveys, especially with helicopters, are also useful in
obtaining sex and age composition data. After some
training, an observer can accurately identify the sex
of all animals, 1 year and older, and determine, by
year, the age of animals up to 4 years old (Allen 1913;
Smith 1976). Low-level aerial photography has also
been used for this purpose (U.S. Department of the
Interior, Fish and Wildlife Service 1983).

Sex and age structure of groups may be deter-
mined in the field by experienced observers. This is
done by referencing horn development with the
sexual dimorphism associated with horn boss and
curvature. Adults may be aged up to the fifth year.
Line drawings in Smith ( 1976) and photographs
in Pohle ( 1981 ) serve as useful points of reference.
Calf and yearling counts can generally be obtained
with ease, although tight bunching behavior may be
a problem when observing young calves.

Mountain Sheep

The mountain sheep comprise two species, Ovis
canadensis and O. dalli, the bighorn and thinhorn
sheep, respectively (Geist 1971a). Mountain sheep
are distributed in mountains and rugged terrain from
Mexico to Alaska. Their recent research and manage-
ment is reported in the annual proceedings of the
Desert Bighorn Council and the Biennial Symposium
of the Northern Wild Sheep and Goat Council.

Habitat Requirements and Analysis. The evolu-
tionary history, anatomy, and behavior of mountain
sheep, especially bighorn sheep, suggest that they
are best adapted to living in moderately large social
groups in open habitats, on or near steep, rugged
terrain. Specific habitat resources for mountain sheep
are —

• forage of adequate quality and of adequate den-
sity and continuity to support a large group;

• water;

• salt licks;

• good visibility for predator detection and for
visual communication;

• steep, usually rocky, escape terrain; and

• a suitable thermal environment.

Favorable combinations of these resources exist in a
patchy distribution. Consequently, the annual ranges
of mountain sheep populations tend to consist of
up to seven or more seasonal ranges (Geist 1971a)
and their connecting migration corridors (Ough and
deVos 1984). Commonly, an annual range includes
separate winter, spring (green-up), and summer
ranges for rams and ewe-juvenile groups, and a lamb-
ing area. Seasonally critical areas, generally within
these ranges, provide salt licks or, in arid country,
water sources. When seasonal ranges are large and

Ungulates

551

dispersed over a large range of elevation, sheep have
options for avoiding harassment and for selecting
elevations and aspects to improve their foraging and
thermal environments.

Mountain sheep are adaptable foragers. Their
food habits vary greatly among regions and seasons
(Todd 1972; Browning and Monson 1980; Cooper-
rider et al. 1980). In northern areas, grasses tend to
dominate sheep diets, especially during spring, sum-
mer, and fall, and modest amounts of browse are
eaten in winter when snow inhibits use of herbs. By
contrast, desert bighorn, and Rocky Mountain big-
horn at low elevations, may browse abundantly on
twigs or leaves in any season. While species compo-
sitions of mountain sheep diets are inconsistent,
three characteristics are common to quality forage
resources: abundance, continuous distribution, and
low stature. Forage abundance and continuity are
needed to support at least modest group sizes; to al-
low individuals within groups to be dispersed, thus
alleviating social interference with foraging and en-
hancing predator detection; and to allow efficient
rates of forage intake (Dale and Bailey 1982). Low-
growing forage does not interfere with vision, thus
allowing for predator detection and visual communi-
cation among sheep. Grasses are the epitome of
abundant, continuous, low-growing forage, but dense
stands of forbs or low shrubs may also be excellent
forage.

by varying their daily activity pattern and by select-
ing favorable environments. They may use elevation,
aspect, terrain wind-shadow, tree shade (especially
in open forests near or on steep terrain), and rock
overhangs or caves to advantage (Simmons 1969).
Often, selection of elevation or aspect provides ad-
vantages of forage quality or snow scarcity, as well as
thermal advantages (Hebert 1973; Shannon et al.
1975).

Land areas are evaluated as habitat for mountain
sheep when ranges are not occupied by sheep, but
are candidates for a transplanted herd or candidates
for habitat improvement to encourage range expan-
sion by an established herd. Land areas can also be
evaluated when ranges are occupied by sheep if
range condition and trend are to be monitored or
the qualities of ranges are to be compared so disrup-
tive developments or habitat degradations can be
directed toward less valuable sheep range.

If the seasonal ranges, water sources, mineral
licks, lambing areas, and migration corridors for an
existing herd are unknown, first consideration should
be given to a telemetry study for locating these
areas. This information will be needed if range con-
dition is to be monitored. It will also be needed if
good range condition is to be maintained or pro-
tected from development or habitat manipulation.

Water and mineral licks are seasonally critical
resources. Sodium and magnesium seem to be key
components of licks used by ungulates, especially
during plant growth season in moist climates (Jones
and Hanson 1985). Water sources and mineral licks
can be created and maintained in sheep manage-
ment. Mountain sheep are especially favored when
these sources are located in open areas near escape
terrain, when human activity is kept at a distance
and livestock are fenced from the site.

The predator-evasion strategy of mountain sheep
requires visibility and escape terrain (Risenhoover
and Bailey 1985b). These habitat characteristics are
especially important to ewes with lambs. Rams are
more prone to use areas lacking these security fac-
tors. In addition, sheep may forego security to use
critical water sources, salt licks, or necessary migra-
tion corridors. Visibility and escape terrain have
compensating effects on security. Poor visibility is
acceptable on steep terrain; sheep will forage farther
from escape terrain when visibility is good.

Lambing areas are usually the largest areas of
rugged, steep terrain within the annual range of a
population. They are often near water.

In both hot and cold climates, mountain sheep
will avoid extreme temperatures and wind chills

Range Assessments. Hansen ( 1980), Grunigen
(1980), and Holl (1982) published methods for eval-
uating land areas as habitat for bighorn sheep. These
methods have not been verified with test applica-
tions outside the areas where they were developed.
In these methods, habitat components indicating
high value as sheep range include —

( 1 ) Abundance of steep escape terrain;

(2) Abundance of certain vegetation types that
are low-growing or open and known to pro-
vide sheep forage, especially where these
types are on or near escape terrain;

(3) Abundance of southerly aspects (this may
duplicate information from vegetation types);

(4) Annual precipitation and presence of reliable
water sources in arid areas;

(5) Absence of livestock and other human activi-
ties and development that would threaten
sheep or sheep habitat.

Ratings of such habitat components would not
likely be combined in a simple additive model. Ab-
sence of a key factor, such as escape terrain or
water, may severely limit the value of a range, re-
gardless of high ratings for other factors.

552

Ungulates

Desert bighorn sheep on escape terrain.

When the objective of range assessment is to
monitor habitat condition and trend, data on habitat
condition should be obtained primarily from —

( 1 ) Areas known to be used by sheep;

(2) Areas on or within 250 m (825 ft) of steep,
rocky terrain;

(3) The most likely migration corridors between
seasonal ranges (shortest routes, with good
visibility, over escape terrain or along ridges);
and

(4) Areas including potential water sources or
mineral licks.

Habitat Component Assessments. Escape
terrain is the most consistent feature of mountain
sheep habitat. However, most biologists have identi-
fied escape terrain subjectively, usually describing
it as steep (> 80% slope), broken, or rocky terrain,
usually including cliffs. Beasom et al. (1983) offered
a method for quantifying land ruggedness. Mc-
Collough et al. (1980) found the minimum height
and length of cliffs used as escape terrain were 8 m
(26 ft) and 220 m (726 ft), respectively. Lambing
areas are steeper and larger than other acceptable es-
cape terrain. The value of escape terrain will be di-
minished by visibility-obstructing vegetation or by
accumulation of snow on north aspects.

The most important habitats for mountain sheep
are on or near escape terrain. Rams, rather than
ewes and juveniles, and sheep in large groups, rather
than small, will use areas farther from escape terrain.
However, out of 1 3 studies of bighorn sheep sur-
veyed, only one showed more than 10% of the ob-
served groups of sheep over 250 m (825 ft) from
escape terrain (Wakelyn 1984). This review indi-
cates that sheep are more restricted to escape ter-
rain than suggested by Van Dyke et al. ( 1983). Until
more definitive studies are reported, range assess-
ment should emphasize areas on or within 250 m
(825 ft) of escape terrain.

Visibility is an important habitat component for
mountain sheep. It allows for predator detection,
visual communication among sheep, and efficient for-
aging. In addition, open habitats with good visibility
may provide abundant forage. In assessing condition
and trend of sheep habitats, some measurement or
index of visibility should be monitored. However,
few objective methods for measuring visibility have
been applied to sheep habitat, and no standards for
acceptable levels of visibility have been developed.
Most often, visibility has been evaluated by rating
vegetation types or tree or shrub crown cover (with
the more open types or overstories rated as better
sheep habitat). Risenhoover and Bailey (1985b) esti-
mated visibility as the percentage of the compass

Ungulates

553

over which an object, 1 m (3.3 ft) tall, could be
seen at 40 m ( 132 ft). A more objective, but more
expensive, method would involve using a density
board (Gysel and Lyon 1980).

In arid ranges, the distribution of water may
limit bighorn habitat. Best water sources are near es-
cape terrain in areas allowing good visibility, located
or fenced so livestock competition or human activity
did not discourage use by sheep. In dry seasons,
most desert bighorn have been observed within 1.6
to 2.4 km (0.96 to 1.4 mi.) of water (Leslie and
Douglas 1979; Elenowitz 1984).

Forage resources for mountain sheep have been
evaluated at two levels. In the simplest approach,
vegetation types have been numerically rated as
forage resources, based on local experience (Hansen
1980; Van Dyke et al. 1983). These ratings may be
influenced not only by forage values, but also by
expected visibility and proximity to escape terrain
for each vegetation type.

At a higher level of resolution, forage resources
may be evaluated by measuring abundance, species
composition, and plant use on selected areas. Inter-
pretation of species composition data will require
a classification of forage species based on local stud-
ies of sheep food habits and preferences. In the west-
ern U.S., monitoring forage condition and trend on
critical ranges of wild ungulates in this way has been
traditional. These detailed measurements of forage
condition and trend can be useful in managing sheep
when the numbers and productivity of sheep are
limited by forage resources and forage conditions are
determined by manageable factors (either the den-
sity of foraging ungulates or plant succession). In
contrast, sheep populations may be controlled by
factors such as predation, disease, or accidents and
forage conditions may be controlled largely by the
density-independent influences of weather. Expen-
sive systems for detailed measurement of forage con-
dition and trend should not be established without
considering the potential value of the information to
be obtained.

Population Measurements.

Relative Density. Trends in sheep populations
have long been evaluated by counting animals on
key areas (winter ranges, at water holes) or along
transects observed from aircraft, vehicles, or on foot.
Resulting data provide minimum population sizes
and estimates of population sex-age structure. On
key areas or with small herds, counts of animals may
include nearly the entire population, though this
assumption is best tested with marked animals.

Trend counts have not usually been considered
for year-to-year population monitoring. Repeated

Ram, ewe, and lamb herd.

counts within years will be needed to evaluate sam-
pling variations (measure the standard count devia-
tions) so trends occurring over a few years can be
detected. Repeated counts should be taken several
days apart to allow for independence. Selection of
standard weather conditions for making counts may
reduce variation among counts and improve the
ability to detect trends. Without repeated counts
within years, only long-term trends may be evaluated
statistically (Harris 1986).

Trend counts may be biased if changes in ani-
mal-observability are correlated with population
trend. This could occur if some animals alter their
habitat selection or their distribution in relation
to counting areas, in response to changing popula-
tion size.

Census. On key areas, repeated classifications
of sheep may be used to produce a known, mini-
mum population size. Since sheep may be counted
by several sex-age classes, the cohort-completion
method (summing the maximum unduplicated count
for each class) may be advantageous. Largest counts
are obtained when several key areas, such as lambing
areas, winter ranges or rutting areas, or waterholes,
are well-known for the herd.

With a known number of marked sheep in the
population, a herd may be estimated by a Petersen
estimate or a similar method (Furlow et al. 1981).
Unless a large proportion of the herd can be marked,
the precision of a Petersen estimate is best con-
trolled by using several second samples of the
marked to unmarked ratio. Variances among several
resulting estimates may be used to calculate confi-
dence limits. Limits will become narrower as the
number of second samples increases.

Most often, all sex-age classes of sheep cannot
be marked at random nor in proportion to their
numbers in the herd. Consequently, the accuracy of

554

Ungulates

a Petersen estimate will depend on representative
second samples. These samples must be distributed
over the entire area used by the herd. For many
herds, rams and ewes may not be equally sampled
except during the rutting or winter seasons. An alter-
native solution may be to estimate the number of
ewes by the Petersen method and to estimate the
numbers of lambs, yearlings, and rams by lamb to
yearling to ewe and ram to ewe ratios obtained dur-
ing seasons when unbiased ratios are most likely
(as during the rut).

Holl and Bleich (1983) used a variation of the
Petersen Estimator requiring concurrent ground-
based and helicopter-based counts of sheep to iden-
tify duplicate sightings. With this variation, many
ground-based counters were necessary.

Population Structure. Mountain sheep may be
classified by eight sex-age categories: lambs, yearling
males, yearling females, ewes, and four classes of
rams based on V* curl horn increments. However,
yearlings, especially yearling females, may be difficult
to distinguish from ewes when they are viewed from
a distance or from aircraft. Desert bighorns have
relatively asynchronous lambing, and yearling fe-
males may vary by 4, 6, or more months. The oldest
of these yearling females will be difficult to distin-
guish from ewes. Consequently, in desert sheep liter-
ature, many reported lamb to ewe ratios are based
on ewe classes, including nonparous yearlings with
ewes. In contrast, lamb to yearling to ewe ratios are
routinely reported for northern mountain sheep.

Mountain sheep are social animals even though
rams and ewe-juvenile groups use separate ranges
during most seasons. This creates special problems
in obtaining a representative sample of population
sex-age structure. Ram to ewe ratios are best ob-
tained during rutting or perhaps on winter ranges or

near water sources in dry seasons, when the distribu-
tions of rams and ewes are usually less independent.
Sampling the population requires at least one season-
ally used range to be well-known. Due to the social
patterns of mountain sheep, individuals are unlikely
to be classified at random. Consequently, either the
social group or all sheep seen on a census route
or area should be the sampling unit (Bowden et al.
1 984 ). Whenever a few large groups of sheep have
been classified, efforts should continue to find
smaller, peripheral groups having differing sex-age
structures. If a large proportion of a herd is classi-
fied, confidence limits for the resulting sex-age ratios
may be narrowed by using statistical methods appro-
priate for finite populations.

Goodson ( 1978) reviewed 23 studies of Rocky
Mountain bighorn, seeking a correlation between
population age structure and population trend. Lamb
to ewe ratios showed no correlation, but yearling
to ewe ratios were loosely correlated. Yearling to
ewe ratios (mean, range/ 100 ewes) were 9 (0-20)
for decreasing herds; 28 (15-56) for stable herds; 31
(20-43) for slightly increasing herds; and 52 (41-
61) for strongly increasing herds. Since mortality
rates are seldom known, these generalizations must
be used with caution.

Discussion. Habitat loss from expanding human
activities is the most widespread threat to mountain
sheep. In the past, impacts of habitat degradation and
expanding human activity have rarely been miti-
gated. There may be frequent opportunities to miti-
gate for habitat impacts by scheduling threatening
activities to avoid critical periods (e.g., lambing) or
by improving and expanding off-site habitat. Mitiga-
tion, however, remains largely untested. Further-
more, with habitat impacts, opportunities for
poaching are often increased and cannot be
mitigated.

. :t. r ■ :

Bighorn sheep in open habitat.

Ungulates

555

Harvest of mountain sheep is usually conserva-
tive, with only older rams being legal. However,
Heimer et al. ( 1984) and Geist (1971b) questioned
the harvest of rams smaller than full curl.

Disease is a common problem with bighorn
sheep. Rocky Mountain and California bighorn have
suffered spectacular all-age dieoffs from pneumonia
(Stelfox 1971; Feuerstein et al. 1980; Foreyt and
Jessup 1982; Spraker et al. 1984; Onderka and Wis-
hart 1984); scabies caused an all-age dieoff in desert
bighorn (Lange et al. 1980). Chronic lamb mortality
is common in Rocky Mountain and desert bighorn
(Spraker and Hibler 1977, 1982; DeForge et al.
1982). Chronic sinusitis and contagious ecthyma are
additional problems (Bunch et al. 1978; Samuel et
al. 1975; Lance et al. 1981). Factors predisposing
sheep to diseases have been proximity to livestock,
especially domestic sheep (Goodson 1982), and
various stressors resulting from habitat loss, concen-
tration of animals, and harassment from human activ-
ities in critical areas or small ranges. In addition,
mountain sheep may not have had a long evolution-
ary history with many diseases they now face, espe-
cially in southern latitudes and where livestock or
exotic ungulates have been introduced (Geist 1985).

Lack of water is another limiting factor for des-
ert bighorn herds and may also influence the distri-

bution of Rocky Mountain bighorn at low elevations.
Development of artificial water sources is prominent
in desert bighorn management (Graf 1980).

In the forested ranges of Rocky Mountain big-
horn, succession has gradually degraded many big-
horn ranges (Wakelyn 1984; Risenhoover and Bailey
1985). Encroachment by trees, combined with man-
made habitat losses, has reduced the sizes of seasonal
ranges, interrupted migration corridors, and isolated
herds. The result has been many small sedentary
herds with few options to move in response to
weather or human activity. Such small isolated herds
may also suffer from inbreeding and depression of
reproduction and survival (Soule and Wilcox 1980;
Sausman 1982). In forested regions, there should be
abundant opportunities to expand sheep ranges and
to mitigate for habitat losses by vegetation manage-
ment, especially with prescribed fire.

Much effort is expended trapping and trans-
planting bighorn into historic ranges from which
they have been extirpated (Schmidt et al. 1978; Wil-
son et al. 1980; Dodd 1983; Fuller 1984; Bates et
al. 1985). Rocky Mountain bighorn are also baited
and treated for lungworms (Schmidt et al. 1979).
However, long-term effects of these management ac-
tivities have not always been monitored.

Lamb/cwc group.

556 Ungulates

LITERATURE CITED

ADAMS, L. 1981. Ecology and population dynamics of
mountain goats, Sheep Mountain-Gladstone Ridge,
Colorado. M.S. thesis, Colorado State Univ., Fort Col-
lins. 189pp.

ADAMS, LG. and J.A. BAILEY. 1980. Winter habitat selec-
tion and group size of mountain goats, Sheep Moun-
tain-Gladstone Ridge, Colorado. Pages 465-481 in
Hickey, W.O., ed. Proc. Biennial Symp., Northern Wild
Sheep and Goat Council. Salmon, ID.

and . 1982. Population dynamics of moun-
tain goats in the Sawatch Range, Colorado. J. Wildl.
Manage. 46:1003-1009.

and . 1983- Winter forages of mountain

goats in central Colorado. J. Wildl. Manage. 46:1237-
1243.

— , MA. MASTELLER, and J.A. BAILEY. 1982a. Move-
ments and home range of mountain goats, Sheep
Mountain-Gladstone Ridge, Colorado. Pages 391-405
in Bailey, J.A. and G.G. Schoonveld, eds. Proc. Third
Biennial Symp., Northern Wild Sheep and Goat Coun-
cil. Fort Collins, CO.

— , KL. RISENHOOVER, and J.A. BAILEY. 1982b. Eco-

logical relationships of mountain goats and Rocky
Mountain bighorn sheep. Pages 9-22 in Bailey, J.A.
and G.G. Schooneld, eds. Proc. Third Biennial Symp.,
Northern Wild Sheep and Goat Council. Fort Collins,
CO.

ALLEN, A.W., J.G. COOK, and M.J. ARMBRUSTER. 1984.
Habitat suitability index models: Pronghorn. U.S. Dep.
Inter., Fish and Wildl. Serv. FWS/OBS-82/ 10.65.
WELUT, Fort Collins, CO. 23pp.

ALLEN, E.O. 1977. A new perspective for elk habitat man-
agement. Proc. Western Assoc. Fish and Game Comm.
195-205.

ALLEN, J.A. 1913- Ontogenetic and other variations in
muskoxen with a systematic review of the muskox
group, recent and extinct. Memoirs Am. Mus. Nat.
Hist., Vol. 1, Part IV: 103-226.

ANDERSON, A.E., D.E. MEDIN, and DC. BOWDEN 1972.
Mule deer numbers and shrub-yield-utilization on
winter range. J. Wildl. Manage. 36:571-578.

AUTENRIETH, R., ed. 1978. Guidelines for the manage-
ment of pronghorn antelope. Pages 472-526 in Bar-
rett, M.W., ed. Proc. Eighth Biennial Pronghorn
Antelope Workshop. Alberta Recreation, Parks, and
Wildlife, Fish and Wildlife Division.

BAILEY, J.A. 1984. Principles of wildlife management. John
Wiley and Sons, New York, NY. 373pp.

and AY. COOPERRIDER. 1982. Final Report:

Trickle Mountain Research Study. U.S. Dep. Inter.,
Bur. Land Manage., Denver Service Center, Denver,
CO. 137pp.

and B.K JOHNSON. 1977. Status of introduced

mountain goats in the Sawatch Range of Colorado.
Pages 54-63 in Samuel, W. and W.G. MacGregor, eds.
Proc. First Int. Mountain Goat Symp. Kalispell, MT.

and G.G. SCHOONVELD, eds. 1982. Northern wild

sheep and goat council. Proc. Third Biennial Symp.

Fort Collins, CO. 405pp.
BAILEY, N.T.J. 1951. On estimating the size of mobile

populations from recapture data. Biometrika 38:293-

306.
BALLARD, W. 1977. Status and management of the moun-

tain goat in Alaska. Pages 15-23 in Samuel, W. and
W.G. MacGregor, eds. Proc. First Int. Mountain Goat
Symp. Kalispell, MT. 243pp.

BANFIELD, A.W.F. 1962. A revision of the reindeer and
caribou, Genus Rangifer. Natl. Mus. Canada Bull. 177.
Ottawa. 137pp.

BARTMANN, R.M. 1974. Piceance deer study — population
density and structure. Colorado Div. Wildl. Game
Res. Rep., July. Part 2:363-380.

BATES, J W, Jr., J. W. BATES, and J.G. GUYMON. 1985.
Comparison of drive nets and darting for capture of
desert bighorn sheep. Wildl. Soc. Bull. 1 3:73-76.

BEAR, G.D. and R.A. GREEN. 1980. Elk population and
ecology studies. Colorado Div. Wildl., Wildl. Res.
Rep, July. Part 2:221-313.

BEASOM, S.L., E.P. WIGGERS, and JR. GIARDINO. 1983. A
technique for assessing land surface ruggedness. J.
Wildl. Manage. 47:1163-1166.

BELOVSKY, G.E. 1978. Diet optimization in a generalist
herbivore: The moose. Theoretical Population Biol.
14:105-134.

BERGERUD, AT 1963 Aerial winter census of caribou. J.
Wildl. Manage. 27:438-449.

. 1978. Caribou. Pages 83-101 in Schmidt, J. L. and

D.L. Gilbert, eds. Big Game of North America: Ecology
and Management. The Wildl. Manage. Inst., Washing-
ton, DC. and Stackpole Books, Harrisburg, PA.

. 1979. Caribou. Pages 83-102 in Schmidt, J.L. and

D.L. Gilbert, eds. Big Game of North America: Ecology
and Management. Stackpole Books, New York, NY.
494pp.

BISSONETTE, J.A. 1976. The relationship of resource

quality and availability to social behavior and organi-
zation in the collared peccary. PhD dissertation, Univ.
Michigan, Ann Arbor. 137pp.

. 1982. Collared peccary. Pages 841-850 in Chap-
man, J.A. and G.A. Feldhamer, eds. Wild Mammals
of North America. The Johns Hopkins Univ. Press, Bal-
timore, MD.

BLACK, H., R.J. SCHERZINGER, and J.W. THOMAS. 1976.
Relationships of Rocky Mountain elk and Rocky
Mountain mule deer habitat to timber management in
the Blue Mountains of Oregon and Washington. Proc.
Elk-Logging-Roads Symp. Univ. Idaho, Moscow: 1 1-31.

BONE, J.N. 1978. Status of the mountain goat {Oreamnos
americanus) of the Similkameen River, British Co-
lumbia. Biennial Symp., Northern Wild Sheep and
Goat Council 1:123-130.

BOWDEN, DC, A.E. ANDERSON, and D.E. MEDIN. 1984.
Sampling plans for mule deer sex and age ratios. J.
Wildl. Manage. 48:500-509.

BOYD, R.J. 1958. Comparison of air and ground deer and
elk counts. Colorado Dep. Game and Fish Quarterly
Progress Rep. Oct: 145- 146.

1970. Elk of the White River Plateau, Colorado.

Colorado Game, Fish, and Parks Dep. Tech. Publ. 25.
126pp.

1978. American elk. Pages 10-29 in Schmidt, J.L.

and D.L. Gilbert, eds. Big Game of North America:
Ecology and Management. The Wildl. Manage. Inst.,
Washington, DC. and Stackpole Books, Harrisburg, PA.

, T.M. POJAR, and B.D. BAKER. 1975. Deer and elk

management study. Colorado Div. Wildl. Game Res.
Rep, July. Part 2:365-412.

BRANDBORG, S.M. 1955. Life history and management of
the mountain goat in Idaho. Idaho Dep. Fish and
Game Wildl Bull. 2. l42pp.

Ungulates

557

BRASSARD, J.M., E. AUDY, M. CRETE, and P. GRENIER.
1974. Distribution and winter habitat of moose in
Quebec. Naturaliste Canadien 101:67-80.

BROWNING, B.M. and G. MONSON. 1980. Food. Pages 80-
99 in Monson, G. and L. Sumner, eds. The Desert
Bighorn. Univ. Arizona Press, Tucson. 370pp.

BUNCH, T.D., S.R. PAUL, and HE. McCUTCHEN. 1978.

Chronic sinusitis in desert bighorn sheep (Ovis cana-
densis nelsoni). Trans. Desert Bighorn Council
22:16-20.

BURNHAM, K.P., DR. ANDERSON, and J.L. LAAKE. 1980.
Estimation of density from line transect sampling
of biological populations. Wildl. Monogr. 72. 202pp.

CALEF, G. and DC. HEARD. 1980. The status of three

tundra wintering caribou herds in northeastern main-
land Northwest Territories. Pages 582-594 in Rei-
mers, E., E. Gaare, and S. Skjenneberg, eds. Proc.
Second Int. Reindeer/Caribou Symp. Dir. for Vilt og
Ferskvannsfisk, Trondheim.

CAMERON, R.D. and KR. WHITTEN. 1979. Seasonal

movements and sexual segregation of caribou deter-
mined by aerial survey. J. Wildl. Manage. 43:626-633-

CAUGHLEY, G. 1977. Analysis of vertebrate populations.
John Wiley and Sons, New York, NY. 234pp.

. 1979. "What is this thing called carrying capacity?"

Pages 2-8 in Boyce, M.S. and L.D. Hayden-Wing, eds.
North American Elk: Ecology, Behavior, and Manage-
ment. Univ. Wyoming, Laramie. 294pp.

CHADWICK, D.H. 1983- A beast the color of winter. Sierra
Club, San Francisco, CA. 208pp.

COADY, J W. and RA. HINMAN. 1984. Management of
muskoxen in Alaska. Pages 47-51 in Klein, DR., R.G.
White, and S. Keller, eds. Proc. First Int. Muskox
Symp. Spec. Rep., Biol. Pap. Univ. Alaska, Fairbanks.

COLLINS, W.B. and P.J. URNESS. 1979. Elk pellet group
distributions and rates of deposition in aspen and
lodgepole pine habitats. Pages 140-144 in Boyce,
M.S., and L.D. Hayden-Wing, eds. North American Elk:
Ecology, Behavior, and Management. Univ. Wyoming,
Laramie. 294pp.

CONLEY, W. 1978. Population modeling. Pages 305-320
in Schmidt, J.L. and D.L. Gilbert, eds. Big Game of
North America: Ecology and Management. Stackpole
Books, New York, NY. 494pp.

CONNOLLY, G.E. 1981. Trends in populations and har-
vests. Pages 225-243 in Wallmo, O.C, ed. Mule and
Black-tailed Deer of North America. Univ. Nebraska
Press, Lincoln. 605pp.

COOPERRIDER, AY. 1982. Forage allocation for elk and
cattle. Pages 142-149 in Britt, T.L., and DP. Theobald,
eds., Proc. Western States Elk Workshop, February
22-24, 1982. Flagstaff, AZ. 166pp.

and J.A. BAILEY. 1984. A simulation approach to

forage allocation. Pages 525-559 in Developing Strate-
gies for Rangeland Management. Rep. prepared by
the Committee on Developing Strategies for Range-
land Manage. Natl. Res. Council/Natl. Acad. Sciences.
Westview Press, Boulder, CO. 2022pp.

, S.A. McCOLLOUGH, and J.A. BAILEY. 1980. Varia-

tion in bighorn sheep food habits as measured by
fecal analysis. Biennial Symp., Northern Wild Sheep
and Goat Council. Salmon, ID. 2:29-41.

DAILEY, T.V., NT. HOBBS, and T.N. WOODARD. 1984.
Experimental comparisons of diet selection by moun-
tain goats and mountain sheep in Colorado. J. Wildl.
Manage. 48:799-806.

DALE, A.R. and J.A. BAILEY. 1982. Application of optimal

foraging theory for bighorn sheep habitat evaluation.
Biennial Symp., Northern Wild Sheep and Goat Coun-
cil 3:254-261.

DAVIS, J.L. and P. VALKENBERG 1979. Caribou distribu-
tion, population characteristics, mortality and re-
sponses to disturbance in northwest Alaska. Chapter 2
in Lent, PC, ed. Studies of Selected Wildlife and
Fish Populations on and adjacent to the National Pe-
troleum Reserve, Alaska. U.S. Dep. Inter., Fish and
Wildl. Serv., Anchorage, AK 226pp.

DeFORGE, JR., DA. JESSUP, CW. JENNER, and J.E.

SCOTT. 1982. Disease investigations into high lamb
mortality of desert bighorn sheep in the Santa Rosa
Mountains, California. Desert Bighorn Council Trans.
65-76.

DODD, N.L. 1983- Ideas and recommendations for maxi-
mizing desert bighorn transplant efforts. Desert Big-
horn Council Trans. 12-16.

DOERR, J. 1979. Population analysis and modeling of the
western arctic herd with comparisons to other Alas-
kan Rangifer populations. Unpubl. M.S. thesis, Univ.
Alaska, Fairbanks. 34 lpp.

DONALDSON, B. 1967. Javelina. Pages 88-94 in New
Mexico Wildlife Management. New Mexico Dep.
Game and Fish, Santa Fe.

EDDY, T.A. 1961. Food and feeding patterns of the col-
lared peccary in southern Arizona. J. Wildl. Manage.
25(3>248-257.

EINARSEN, AS. 1948. The pronghorn antelope and its

management. Stackpole Books, Harrisburg, PA. 235pp.

ELENOWITZ, A. 1984. Group dynamics and habitat use
of transplanted desert bighorn sheep in the Peloncillo
Mountains, New Mexico. Desert Bighorn Council
28:1-8.

ELLISOR, J.E. and W.F. HARWELL. 1979. Ecology and

management of javelina in south Texas. FA. Rep., Se-
ries 16. Texas Parks and Wildl. Dep., Austin. 25pp.

ERIKSSON, O. 1980. A method for range appraisal using
small aircraft for sampling of vegetation data. Pages
41-46 in Reimers, E., E. Gaare, and S. Skjenneberg,
eds. Proc. Second Int. Reindeer/Caribou Symp. Dir. for
Vilt og Ferskvannsfisk, Trondheim.

FEUERSTEIN, V., RL. SCHMIDT, C.P. HIBLER, and W.H.
RUTHERFORD. 1980. Bighorn sheep mortality in the
Taylor River-Almont Triangle Area, 1978-1979. A case
study. Colorado Div. Wildl. Spec. Rep. 48. 19pp.

FICHTER, E. 1974. On the bedding behaviour of prong-
horn fawns. Pages 352-355 in Geist, V. and F.
Walther, eds. The Behaviour of Ungulates and Its
Relation to Management. IUCN Publ. New Series 24.
94 lpp.

FOREYT, W.J. and DA. JESSUP. 1982. Fatal pneumonia of
bighorn sheep following association with domestic
sheep. J. Wildl. Diseases 18:163-168.

FOSTER, BR. 1977. Historical patterns of mountain goat
harvest in British Columbia. Pages 147-159 in Samuel,
W. and W. MacGregor, eds. Proc. First Int. Mountain
Goat Symp. Kalispell, Ml

. 1978. Horn growth and quality management for

mountain goats. Pages 200-226 in Hebert, D.M. and
M. Nation, eds. Proc. 1978 Northern Wild Sheep and
Goat Conf. Penticton, British Columbia.

1982. Observability and habitat characteristics of

the mountain goat in west-central British Columbia.
M.S. thesis, Univ. British Columbia, Vancouver. 134pp.
FOWLER, CW. and ID. SMITH, eds. 1981. Dynamics of
large mammal populations. John Wiley and Sons, New

558

Ungulates

York, NY. 477pp.

FOX, J.L and R.D. TABER. 1981. Site selection by moun-
tain goats wintering in forest habitat. U.S. Dep. Agric.,
For. Serv., Pacific Northwest For. and Range Experi-
ment Sta, Juneau, AK 55pp.

, KJ. RAEDEKE, and C.A. SMITH. 1982. Mountain

goat ecology on Cleveland Peninsula, Alaska, 1980-
1982. U.S. Dep. Agric, For. Serv., Pacific Northwest
For. and Range Experiment Sta., Juneau, AK. 4 lpp.

FRANZMANN, A.W. 1978. Moose. Pages 67-82 in Schmidt,
J.L. and D.L. Gilbert, eds. Big Game of North America.
Stackpole Books, Harrisburg, PA.

, A. FLYNN, and P.D. ARNESON. 1975. Levels of

some mineral elements in Alaskan moose hair. J.
Wildl. Manage. 39:374-378.

, RE. LeRESCHE, R.A. RAUSCH, and J.L. OLDE

MEYER. 1978. Alaskan moose measurements and
weights and measurement-weight relationships. Cana-
dian J. Zoology 56:298-306.

FRASER, D., E.R. CHAVEZ, and J.L. PALOHEIMO. 1984.
Aquatic feeding by moose: Selection of plant species
and feeding areas in relation to plant chemical com-
position and characteristics of lakes. Canadian J. Zool-
ogy 62:80-87.

FREDDY, D.J. and DC. BOWDEN. 1983a. Efficacy of per
manent and temporary pellet plots in juniper-pinyon
woodland. J. Wildl. Manage. 47:512-516.

and . 1983b. Sampling mule deer pellet-
group densities in juniper-pinyon woodland. J. Wildl.
Manage. 47:476-485.

FULLER, A.F. 1984. Drop net capture of bighorn sheep in
Arizona. Desert Bighorn Council Trans. 28:39-40.

FURLOW, R.C., M. HADERLIE, and R. VAN DEN BERGE.
1981. Estimating a bighorn population by mark-recap-
ture. Desert Bighorn Council Trans. 25:31-33

GASAWAY, W.D., S.D. DUBOIS, and S. HARBO. 1985.

Biases in aerial transect surveys for moose during May
and June. J. Wildl. Manage. 49:777-784.

GEIST, V. 1971a. Mountain sheep: A study in behavior and
evolution. Univ. Chicago Press, IL. 383pp.

1971b. Bighorn sheep biology. The Wildl. Soc.

News 136:61.
— . 1981. Behavior-

-adaptive strategies in mule deer.

Pages 157-224 in Wallmo, O.C., ed. Mule and Black-
tailed Deer of North America. Univ. Nebraska Press,
Lincoln. 605pp.

1985. Pleistocene bighorn sheep: Some problems

of adaptation, and relevance to today's American
megafauna. Wildl. Soc. Bull. 13:351-359.

GILBERT, JR. 1978. Estimating population characteristics.
Pages 297-304 in Schmidt, J.L. and D.L. Gilbert, eds.
Big Game of North America: Ecology and Manage-
ment. Stackpole Books, New York, NY. 494pp.

GILL, R.B., L.H. CARPENTER, and DC. BOWDEN. 1983.
Monitoring large animal populations — the Colorado
experience. Trans. North Am. Wildl. and Nat. Res.
Conf. 48:330-341.

GOODSON, N.J. 1978. Status of bighorn sheep in Rocky
Mountain National Park. M.S. thesis, Colorado State
Univ., Fort Collins. 190pp.

. 1982. Effects of domestic sheep grazing on big-
horn sheep populations: A review. Biennial Symp.,
Northern Wild Sheep and Goat Council 3:287-313.

GRAF, W. 1980. Habitat protection and improvement.

Pages 310-319 in Monson, G. and L. Sumner, eds. The
Desert Bighorn. Univ. Arizona Press, Tucson. 370pp.

GRONQUIST, R.M. and W.B. DINNEFORD. 1984. Age

determination of muskoxen from dental cementum
annuli. Pages 67-68 in Klein, DR., R.G. White, and S.
Keller, eds. Proc. First Int. Muskox Symp. Spec. Rep.,
Biol. Pap. Univ. Alaska, Fairbanks.

GROSS, J. E. 1970. Program ANPOP. A simulation modeling
exercise of the Wichita Mountains National Wildlife
Refuge. Colorado Coop. Wildl. Res. Unit Prog. Rep.
Colorado State Univ., Fort Collins. 133pp

GRUNIGEN, R.E. 1980. A system for evaluating potential
bighorn sheep transplant sites in northern New Mex-
ico. Biennial Symp., Northern Wild Sheep and Goat
Council 2:211-228.

GUNN, A., R. DECKER, and T.W. BARRY. 1984. Pages 67-
68 in Klein, DR., R.G. White, and S. Keller, eds. Proc.
First Int. Muskox Symp. Spec. Rep., Biol. Pap. Univ.
Alaska, Fairbanks.

GYSEL, L.W. and L.J. LYON. 1980. Habitat analysis and
evaluation. Pages 305-327 in Schemnitz, S.D., ed.
Wildlife Management Techniques Manual. The Wildl.
Soc, Washington, DC. 686pp.

HALL, E.R. and KR. KELSON. 1959. The mammals of
North America. Vol. 11:547-1083. The Ronald Press,
New York, NY.

HALL, W.K 1977. Status and management of the Rocky
Mountain goat, Oreamnos americanus, in the prov-
ince of Alberta. Pages 8-14 in Samuel, W. and W.G.
MacGregor, eds. Proc. First Int. Mountain Goat Symp.
Kalispell, MT. 243pp.

HALLS, L.K 1984. White-tailed deer: Ecology and Manage-
ment. Stackpole Books, Harrisburg, PA. 870pp.

HANKS, J. 1981. Characterization of population condition.
Pages 47-74 in Fowler, C.W. and T.D. Smith, eds.
Dynamics of Large Mammal Populations. John Wiley
and Sons, New York, NY. 477 pp.

HANSEN, C.G. 1980a. Habitat. Pages 64-79 in Monson, G.
and L. Sumner, eds. The Desert Bighorn — Its Life
History, Ecology, and Management. Univ. Arizona
Press, Tucson. 370pp.

. 1980b. Habitat evaluation. Pages 320-335 in Mon-
son, G. and L. Sumner, eds. The Desert Bighorn — Its
Life History, Ecology, and Management. Univ. Arizona
Press, Tucson. 370pp.

HANSON, W.R. 1963 Calculation of productivity, survival
and abundance of selected vertebrates from sex and
age ratios. Wildl. Monogr. 9. 60pp.

HARRIS, R.B. 1986. Reliability of trend lines obtained
from variable counts. J. Wild. Manage. 50:165-171.

HEBERT, D.M. 1973. Altitudinal migration as a factor
in the nutrition of bighorn sheep. PhD dissertation,
Univ. British Columbia, Vancouver. 355pp.

and M. NATION, eds. 1978. Proceedings of the

1978 northern wild sheep and goat conference. Pen-
ticton, British Columbia. 412pp.

and W.G. TURNBULL. 1977. A description of south-
ern interior and coastal mountain goat types of British
Columbia. Pages 126-159 in Proc First Inter. Moun-
tain Goat Symp. Kalispell, MT.

HEIB, S.R., ed. 1976. Proceedings of the elk-logging-roads
symposium. Univ. Idaho, Moscow. 142pp.

HEIMER, W.E., S.M. WATSON, and T.C. SMITH. 1984.

Excess ram mortality in a heavily hunted Dall's sheep
population. Biennial Symp., Northern Wild Sheep
and Goat Council 4:425-432.

HESSELTON, W.T. and R.M. HESSELTON. 1982. White-
tailed deer. Pages 878-901 in Chapman, J.A. and G.A.
Feldhamer, eds. Wild Mammals of North America. The
Johns Hopkins Univ. Press, Baltimore, MD.

Ungulates

559

HIBBS, L.D. 1966. A literature review on mountain goat
ecology. Colorado Dep. Game, Fish, and Parks Spec.
Rep. 8. 23pp.

HICKEY, W.O., ed. 1980. Proceedings of the biennial
symposium of the northern wild sheep and goat
council. Salmon, ID. 668pp.

HJELJORD, O. 1973. Mountain goat forage and habitat

preference in Alaska. J. Wildl. Manage. 37(3):353-362.

HOBBS, NT., D.L. BAKER, J.E. ELLIS, DM. SWIFT, and
R.A. GREEN. 1982. Energy- and nitrogen-based esti-
mates of elk winter-range carrying capacity. J. Wildl.
Manage. 46:12-21.

HOLL, S.A. 1982. Evaluation of bighorn sheep habitat.
Desert Bighorn Council Trans. 24:47-49.

and V.C. BLEICH. 1983. San Gabriel mountain

sheep: Biological and management considerations. U.S.
Dep. Agric, For. Serv., San Bernardino National Forest,
San Bernardino, CA.

HOOVER, R.L., C.E. TILL, and S. OGILVIE. 1959. The ante-
lope of Colorado. Colorado Dep. of Game and Fish
Tech. Bull. 4. Denver, CO. 110pp.

HOREJSI, B.L. 1976. Suckling and feeding behavior in

relation to lamb survival in bighorn sheep (Ovis can-
adensis canadensis Shaw). Unpubl. PhD dissertation,
Univ. Calgary. 265pp.

HOUSTON, D.B. 1974. Aspects of the social organization
of moose. Pages 690-696 in Geist, V. and F. Walther,
eds. The Behaviour of Ungulates and Its Relation to
Management. IUCN Publ. New Series 24. 94 lpp.

JENNINGS, W.S. and J.T. HARRIS. 1953. The collared pec-
cary in Texas. Texas Game and Fish Comm. FA. Rep.
Series 12. 31pp.

JONES, R.L. and H.C. HANSON. 1985. Mineral licks, geo-
phagy, and biogeochemistry of North American ungu-
lates. Iowa State Univ. Press. 302pp.

JORDAN, PA., D.B. BOTKIN, AS. DOMINSKI, H.S. LOW
ENDORF, and G.E. BELOVSKY. 1973- Sodium as a
critical nutrient for the moose of Isle Royale. North
Am. Moose Conf. 9:13-42.

JUNIPER, I. 1980. Problems in managing an irrupting

caribou herd. Pages 722-724 in Reimers, E., E. Gaare,
and S. Skjenneberg, eds. Proc. Second Int. Reindeer/
Caribou Symp. Dir. for Vilt og Ferskvannsfisk, Tron-
dheim.

KEISS, RE. 1969. Comparison of eruption-wear patterns
and cementum annuli as age criteria in elk. J. Wildl.
Manage. 33(1)175-180.

KELSALL, J.P. 1965. The migratory barren-ground caribou
of Canada. Dep. of Indian Affairs and Northern Devel-
opment. Canadian Wildl. Serv., Ottawa. 339pp.

. 1968. The migratory barren-ground caribou of

Canada. Canadian Wildl. Serv. Monogr. 3- Ottawa, On-
tario. 340pp.

. 1 969. Structural adaptation of moose and deer for

reindeer on St. Matthew Island. J. Wildl. Manage.

32:350-367.
— . 1982. Fire, lichens and caribou. J. Range Manage.

35:390-395.
— , R.G. WHITE, and S. KELLER, eds. 1984. Proceed-

snow. J. Mammal. 50:302-310.

and E.S. TELFER. 1974. Biogeography of moose

with particular reference to western North America.
Naturaliste Canadien 101:117-130.

KERR, R. 1979. Mule deer habitat guidelines. U.S. Dep.
Inter., Bur. Land Manage. Tech. Note 366. Service
Center, Denver, CO.

KINDSCHY, R.R., C SUNDSTROM, and J.D. YOAKUM.
1982. Wildlife habitats in managed rangelands — the
Great Basin of southeastern Oregon. U.S. Dep. Agric,
For. Serv., Pacific Northwest For. and Range Experi-
ment Sta. Gen. Tech. Rep. 145. 18pp.

KLEIN, DR. 1968. The introduction, increase and crash of

ings of first international muskox symposium. Biol.
Pap. Univ. Alaska, Fairbanks.

KNIPE, T. 1957. The javelina in Arizona. Arizona Game
and Fish Dep. Wildl. Bull. 2.

KONKEL, G.W., ed. 1980. Terrestrial habitat evaluation
criteria handbook — Alaska. Div. Ecol. Serv., U.S. Dep.
Inter., Fish and Wildl. Serv., Anchorage, AK (Original
not seen; cited and partially reproduced in Mule
1982).

KRAUSMAN, PR, JR. MORGART, and M. CHILELLI. 1984.
Annotated bibliography of desert bighorn sheep liter-
ature, 1897-1983. Southwest Nat. Hist. Assoc, Phoe-
nix, AZ. 204pp.

KUCK, L 1977a. Status and management of the mountain
goat in Idaho. Pages 37-40 in Samuel, W. and W.O.
MacGregor, eds. Proc. First Int. Mountain Goat Symp.
Kalispell, MT.

. 1977b. The impact of hunting on Idaho's Pahsime-

roi mountain goat herd. Pages 114-125 in Samuel,
W. and W.O. MacGregor, eds. Proc First Int. Moun-
tain Goat Symp. Kalispell, MT.

KUFELD, R.C., J.H. OLTERMAN, and DC. BOWDEN. 1980.
A helicopter quadrat census for mule deer on Uncom-
pahgre Plateau, Colorado. J. Wildl. Manage. 44:632-
639.

KUROPAT, P. and J.P. BRYANT. 1980. Foraging behavior
of cow caribou on the Utukok calving grounds in
Northwestern Alaska. Pages 64-70 in Reimers, E., E.
Garre, and S. Skjenneberg, eds. Proc. Second Int.
Reindeer/Caribou Symp. Dir. for Vilt og Ferskvanns-
fisk, Trondheim.

KURT, F. 1968. Das Socialverhalten des Rehes. Eines Feld-
studie. P. Parey-Verlag, Berlin. 102pp.

LANCE, W„ W. ADRIAN, and B. WIDHALM. 1981. An

epizootic of contagious ecthyma in Rocky Mountain
bighorn sheep in Colorado. J. Wildl. Diseases 17:601-
603.

LANGE, RE., A. SANDOVAL, and W.P. MELANEY. 1980.
Psoroptic scabies in bighorn sheep in New Mexico. J.
Wildl. Diseases 16:77-82.

LaPERRIERE, A.J., PC. LENT, W.C. GASAWAY, and FA.
NODLER. 1980. Use of Landsat data for moose-habitat
analyses in Alaska. J. Wildl. Manage. 44:881-887.

LASSEN, P. 1984. Muskox distribution and population
structure in Jameson Land, northeast Greenland,
1981-1983. Pages 19-24 in Klein, DR., R.G. White,
and S. Keller, eds. Proc. First Int. Muskox Symp. Spec.
Rep., Biol. Pap. Univ. Alaska, Fairbanks.

LEADER- WILLIAMS, N. 1980. Population ecology of rein-
deer on South Georgia. Pages 664-676 in Reimers,
E., E. Gaare, and S. Skjenneberg, eds. Proc. Second Int.
Reindeer/Caribou Symp. Dir. for Vilt og Ferskvanns-
fisk, Trondheim.

LENT, PC. 1966. The caribou of northwestern Alaska.
Pages 481-517 in Reimers, E., E. Gaare, and S. Skjen-
neberg, eds. Proc. Second Int. Reindeer/Caribou
Symp. Dir. for Vilt og Ferskvannsfisk, Trondheim.

. 1971. Muskox management controversies in North

America. Biol. Conserv. 3:255-263.

. 1974a. Mother-infant relationships in ungulates.

Pages 14-55 in Geist, V. and F. Walthers, eds. The Be-
havior of Ungulates and Its Relation to Management.

560

Ungulates

Vol. I, IUCN Publ. New Series 24. Morges, Switzer-
land. 940pp.

— . 1974b. A review of rutting behavior in moose.
Naturaliste Canadien 101:307-323.

— . 1974c. Ecological and behavioral study of the
Nunivak Island muskox population. Unpubl. Rep. U.S.
Dep. Inter., Fish and Wildl. Serv., Anchorage and
Bethel, AK. 90pp.

— . 1978. Muskox. Pages 135-147 in Schmidt, J.L. and
D.L. Gilbert, eds. Big Game of North America: Ecology
and Management. The Wildl. Manage. Inst., Washing-
ton, DC. and Stackpole Books, Harrisburg, PA.

1980. Synoptic snowmelt patterns in arctic Alaska

in relation to caribou habitat use. Pages 71-83 in
Reimers, E., E. Gaare, and S. Skjenneberg, eds. Proc.
Second Int. Reindeer/Caribou Symp. Dir. for Vilt og
Ferskvannsfisk, Trondheim.

LeRESCHE, RE. 1974. Moose migrations in North Amer-
ica. Naturaliste Canadien 101:393-415.

and A. RAUSCH. 1974. Accuracy and precision of

aerial moose censusing. J. Wildl. Manage. 38:175-182.

, U.S. SEAL, P.D. KARNS, and AW. FRANZMANN.

1974. A review of blood chemistry of moose and
other Cervidae with emphasis on nutritional assess-
ments. Naturaliste Canadien 101 263-290.

LESLIE, DM, Jr. and C.L. DOUGLAS. 1979. Desert bighorn
sheep of the River Mountains, Nevada. Wildl. Monogr.
66:1-56.

LONGHURST, W.M., H.K OH, MB. JONES, and RE. KEP-
NER. 1 968. A basis for the palatability of deer forage
plants. Trans. North Am. Wildl. and Nat. Resour. Conf.
33:181-189.

LONO, O. I960. Transplantation of the muskox in Europe
and North America. Meddel. Norsk Polarinst. 84:3-25.

LOW, W.A. 1970. The influence of aridity on reproduction
of the collared peccary {Dicotyles tajacu [ Linn ] ) in
Texas. PhD dissertation, Univ. British Columbia, Van-
couver. 170pp.

LYON, L.J., T.N. LONNER, J.P. WEIGAND, C.L. MARCUM,
WD. EDGE, J.D. JONES, D.W. McCLEEREY, and L.L.
HICKS. 1985. Coordinating elk and timber manage-
ment. Final Rep. Montana Coop. Elk-Logging Study,
1970-1985. Montana Dep. Fish, Wildlife, and Parks,
Helena.

and A.L. WARD. 1982. Elk and land management.

Pages 443-478 in Thomas, J.W. and D.E. Toweill, eds.
Elk of North America: Ecology and Management.
Stackpole Books, Harrisburg, PA, 698pp.

MACKIE, R.J., K.L. HAMLIN, and D.F. PAC. 1982. Mule
deer. Pages 862-877 in Chapman, J.A. and G.A. Feld-
hamer, eds. Wild Mammals of North America. The
Johns Hopkins Univ. Press, Baltimore, MD.

MacPHERSON, S., F.K. MARTINSON, and AY. COOPER-
RIDER. 1982. Forage allocation between big game
and domestic livestock — new approaches. Proc. West-
ern Assoc. Fish and Wildl. Agencies 62:120-129.

MAUTZ, WW. 1978. Nutrition and carrying capacity.

Pages 321-348 in Schmidt, J.L. and D.L. Gilbert, eds.
Big Game of North America: Ecology and Manage-
ment. The Wildl. Manage. Inst., Washington, DC. and
Stackpole Books, Harrisburg, PA.

McCOLLOUGH, S.A. 1982. Impact of cattle grazing on
bighorn sheep, Trickle Mountain, Colorado. M.S. the-
sis, Colorado State Univ., Fort Collins. 1 19pp.

., AY. COOPERRIDER, and J.A. BAILEY. 1980. Im-
pact of cattle grazing on bighorn sheep at Trickle
Mountain, Colorado. Biennial Symp., Northern Wild

Sheep and Goat Council 2:42-58.

McEWAN, EH. and P.E. WHITEHEAD. 1970. Seasonal

changes in the energy and nitrogen intake in reindeer
and caribou. Canadian J. Zoology 48:905-913

McFETRIDGE, R.J. 1977. Strategy of resource use by
mountain goats nursery groups. Pages 169-173 in
Samuel, W. and W.G. MacGregor, eds. Proc. First Int.
Mountain Goat Symp. Kalispell, MT.

MEAGHER, M.M. 1973. The bison of Yellowstone National
Park. U.S. Dep. Inter., Natl. Park Serv. Sci. Monogr. 1.
161pp.

. 1976. Winter weather as a population regulating

influence on free-ranging bison in Yellowstone Na-
tional Park. Pages 29-38 in Research in the Parks.
Trans. Natl. Park Centennial Symp., AAAS, December
28-29, 1971. Series 1. U.S. Govt. Printing Office.
232pp.

1978. Bison. Pages 123-133 in Schmidt, J.L. and

D.L. Gilbert, eds. Big Game of North America: Ecology
and Management. The Wildl. Manage. Inst., Washing-
ton, DC. and Stackpole Books, Harrisburg, PA. 494pp.

MONSON, G. and L. SUMNER, eds. 1980. The desert big-
horn. Univ. Arizona Press, Tucson. 370pp.

MOULD, E. 1977. Habitat relationships of moose in north-
ern Alaska. North Am. Moose Conf. 13:144-156.

. 1979. Seasonal movements related to habitat of

moose along the Colville River, Alaska. Murrelet
60:6-11.

MULE, R.S. 1982. An assessment of a wildlife habitat evalu-
ation methodology for Alaska. M.S. thesis, Univ.
Alaska, Fairbanks. 215pp.

MURIE, O.J. 1954. A field guide to animal tracks. Hough-
ton Mifflin Co., Boston, MA. 375pp.

NAGY, J.G., H.W. STEINHOFF, and GM. WARD. 1964.
Effects of essential oils of sagebrush on deer rumen
microbial function. J. Wildl. Manage. 28:785-790.

NEAL, B.J. 1959. A contribution on the life history of the
collared peccary in Arizona. Am. Midi. Nat. 61:177-
190.

NEFF, D.J. 1968. The pellet group count technique for big
game trend, census, and distribution: A review. J.
Wildl. Manage. 32:597-614.

NELSON, JR. 1984. A modeling approach to large herbi-
vore competition. Pages 491-524 in Developing Strat-
egies for Rangeland Management. Rep. prepared by
the Committee on Developing Strategies for Range-
land Manage., Natl. Res. Council/Natl. Acad. Sciences.
Westview Press, Boulder, CO. 2022pp.

NELSON, M.E. 1979. Home range location of white-tailed
deer. U.S. Dep. Agric, For. Serv. Res. Pap. NC-173-
10pp.

and L.C. MECH. 1981. Deer social organization and

wolf predation in northeastern Minnesota. Wildl.
Monogr. 77. 53pp.

NICHOLS, L, Jr. 1980. Aerial census and classification of
mountain goats in Alaska. Pages 523-589 in Hickey,
W.O., ed. Proc. Biennial Symp., Northern Wild Sheep
and Goat Council. Salmon, ID.

NPR-A (NATIONAL PETROLEUM RESERVE— ALASKA)
TASK FORCE. 1978. Fish and wildlife resources. Val-
ues and Resource Analysis, Section 6, Vol. 3, Study-
Rep. 2. U.S. Dep. Inter. 224pp.

. 1979. Ecological profile. National Petroleum Re-
serve in Alaska. Study 25, Rep. 4. U.S. Dep. Inter.

OLDEMEYER, J.L. 1977a. Nutritive value of moose forage.
Naturaliste Canadien 101:217-226.

. 1 977b. Impact of LeTourneau tree crushers on

Ungulates

561

moose habitat on the Kenai National Moose Range.
Proc. 28th Annual Conf., Northwest Sections. The
Wildl. Soc, Kalispell, MT.

— , W.J. BARMORE, and D.L. GILBERT. 1971. Winter
ecology of bighorn sheep in Yellowstone National
Park. J. Wildl. Manage. 35:257-269.
— , AW. FRANZMANN, A.L. BRUNDAGE, P.D. ARNE

SON, and A. FLYNN. 1977. Browse quality and the
Kenai moose population. J. Wildl. Manage. 41:533-
542.

ONDERKA, D.K and WD. WISHART. 1984. A major big-
horn sheep die-off from pneumonia in southern Al-
berta. Biennial Symp., Northern Wild Sheep and Goat
Council. Salmon, ID. 4:356-363.

OUGH, WD. and J.C. DE VOS, Jr. 1984. Intermountain
travel corridors and their management implications
for bighorn sheep. Desert Bighorn Council Trans.
32-36.

PARKER, G.R. 1975. A review of aerial surveys used for
estimating the numbers of barren-ground caribou in
northern Canada. Polar Record 17:627-638.

. 1981. Physical and reproductive characteristics of

an expanding woodland caribou population in north-
ern Labrador. Canadian J. Zoology 59:1929-1940.

and R.K ROSS. 1976. Summer habitat use by musk-
oxen (Ovibos moschatus) and Peary caribou (Rangi-
fer tarandus pearyi) in the Canadian High Arctic.
Polarforschung 46:12-25.

, DC. THOMAS, E. BROUGHTON, and DR. GRAY.

1975. Crashes of muskox and Peary caribou popula-
tions in 1973-74 on the Parry Island, Arctic Canada.
Canadian Wild. Serv. Progress Note 56:1-10.

PEDEN, D.G., GM. VAN DYNE, R.W. RICE, and R.M. HAN-
SEN. 1974. The trophic ecology of Bison bison L.
on shortgrass plains. J. Applied Ecol. 1 1 :489-498.

PEEK, J.M. 1974. On the nature of winter habitats of
Shiras moose. Naturaliste Canadien 101:131-141.

, M.D. SCOTT, L.J. NELSON, and DJ. PIERCE. 1982.

Role of cover in habitat management for big game
in northwestern United States. Trans. North Am.
Wildl. Nat. Resour. Conf. 47:363-373-

PEGAU, RE. and J.E. HEMMING. 1970. Caribou report.
Alaska Dep. of Fish and Game, Federal Aid in Wildl.
Restoration Project Rep., Vol. 12. Juneau. 224pp.

PETOCZ, R.G. 1973. The effect of snow cover on the so-
cial behaviour of bighorn rams and mountain goats.
Canadian J. Zoology 51:987-993.

POHLE, C. 1981. Trergartnerische Beobachtungen bei der
Haltung and Zucht von Moschusochsen in Tierpark
Berlin. Zool. Gart. 51:289-322.

POJAR, T.M. 1981. A management perspective of popula-
tion modeling. Pages 241-261 in Fowler, C.W. and
T.D. Smith, eds. Dynamics of Large Mammal Popula-
tions. John Wiley and Sons, New York, NY. 477pp.

and D. STRICKLAND, eds. 1979. A workshop on

the status and application of big game population
modeling. Colorado Div. Wildl., Fort Collins. 53pp.

POLEQUIN, A., B. SCHERRER, and R. JOYAL. 1977. Char
acteristics of winter browsing areas of moose in west-
ern Quebec as determined by multivariate analysis.
North Am. Moose Conf. 13:128-143.

PRENZLOW, E.J., D.L. GILBERT, and FA. GLOVER. 1968.
Some behavior patterns of the pronghorn. Colorado
Dep. Game, Fish, and Parks Spec. Rep. 17. 16pp.

QUIMBY, D.C. and J.E. GAAB. 1957. Mandibular dentition
as an age indicator in Rocky Mountain elk. J. Wildl.
Manage. 21(4):435-451.

REIMERS, E., DR. KLEIN, and R. SORUMGORD. 1983.

Calving time, growth rate and body size of Norwegian
reindeer on different ranges. Arctic and Alpine Res.
15:107-118.

RENECKER, LA. and R.J. HUDSON. 1986. Seasonal forag-
ing rates of free-ranging moose. J. Wildl. Manage.
50:143-147.

RICHARDSON, AH. 1971. The Rocky Mountain goat in
the Black Hills. South Dakota Dep. Game, Fish, and
Parks Bull. 2. 25pp.

RIDEOUT, C.B. 1974. A radio telemetry study of the ecol-
ogy and behavior of the Rocky Mountain goat in
western Montana. PhD thesis, Univ. Kansas, Lawrence.
1 46pp.

RINEY, T. 1982. Study and management of large mammals.
John Wiley and Sons, New York, NY. 552pp.

RISENHOOVER, KL. and J.A. BAILEY. 1982. Social dynam-
ics of mountain goats in summer: Implications for
age ratios. Pages 364-373 in Bailey, J.A. and G.G.
Schoonveld, eds. Proc. Third Biennial Symp., Northern
Wild Sheep and Goat Council. Fort Collins, CO.

and . 1985a. Relationships between group

size, feeding time, and agonistic behavior of mountain
goats. Canadian J. Zoology 63:2501-2506.

and . 1985b. Foraging ecology of mountain

sheep: Implications for habitat management. J. Wildl.
Manage. 49:797-804.

ROBBINS, C.T. 1973. The biological basis for the determi-
nation of carrying capacity. PhD dissertation, Cornell
Univ., Ithaca, NY. 239pp.

ROBINETTE, W.L., CM. LOVELESS, and DA. JONES. 1974.
Field tests of strip census methods. J. Wildl. Manage.
38:81-96.

ROBUS, M. 1984. Summer food habits of muskoxen in
northeastern Alaska. Pages 81-85 in Klein, D.R., R.G.
White, and S. Keller, eds. Proc. First Int. Muskox
Symp. Spec. Rep., Biol. Pap. Univ. Alaska, Fairbanks.

RYEL, L.A. 1971. Evaluation of pellet group surveys for
estimating deer populations in Michigan. PhD disser-
tation, Michigan State Univ., Lansing. 237pp.

SAMUEL, W., G.A. CHALMERS, J.G. STELLOX, A. LOWEN,
and J.J. THOMSEN. 1975. Contagious ecthyma in
bighorn sheep and mountain goat in western Canada.
J. Wildl. Diseases 11:26-31.

and W. MacGREGOR, eds. 1977. Proceedings of the

first international mountain goat symposium. Kalispell,
MT.

SAUSMAN, K 1982. Survival of captive-born Ovis cana-
densis in North American zoos. Desert Bighorn Coun-
cil Trans. 26-31.

SCHMIDT, J.L. and D.L. GILBERT. 1978. Big game of North
America: Ecology and Management. The Wildl. Man-
age. Inst., Washington, DC and Stackpole Books, Har-
risburg, PA. 494pp.

SCHMIDT, R.L., C.P. HIBLER, T.R. SPRAKER, and W.H.

RUTHERFORD. 1979. An evaluation of drug treatment
for lungworm in bighorn sheep. J. Wildl. Manage.
43:461-467.

, W.H. RUTHERFORD, and F.M. BODENHAM. 1978.

Colorado bighorn sheep-trapping techniques. Wildl.
Soc. Bull. 6:159-163.

SCHOEN, J.W. and M.D. KIRCHOFF 1981. Habitat use by
mountain goats in southeast Alaska. Alaska Dep. Fish
and Game, PR Project Rep. W-17-10, 11 and W-21-1,
2. 67pp.

SERGEANT, DE. and D.H. PIMLOTT. 1959. Age determi-
nation in moose from sectioned incisor teeth. J. Wildl.

562

Ungulates

Manage. 23:315-321.

SHANNON, N.H., R.J. HUDSON, V.C. BRINK, and WD.
KITTS. 1975. Determinants of spatial distribution of
Rocky Mountain bighorn sheep. J. Wildl. Manage.
39:387-401.

SIMMONS, N.M. 1969. Heat stress and bighorn behavior in
the Cabeza Prieta Game Range, Arizona. Desert Big-
horn Council Trans. 13:55-63-

SINGER, F.J. 1975. Behavior of mountain goats, elk and
other wildlife in relation to U.S. Highway 2, Glacier
National Park. Completion Rep. U.S. Dep. Inter., Natl.
Park Serv., West Glacier, MT. 98pp.

SINIFF, D.B. and R.O. SKOOG. 1964. Aerial censusing of
caribou using stratified random sampling. J. Wildl.
Manage. 28:39-40.

SKILES J.W., P.T. KORTOPATES, and G.M. VAN DYNE.
1980. Optimization models for forage allocation to
combinations of large herbivores for grazing land
situations: A critical use of proper use factors. Rep. to
U.S. Dep. Inter., Bur. Land Manage. 86pp.

SKOOG, R.O. 1962. Method for estimating caribou herds.
Alaska Dep. of Fish and Game Information Leaflet
20., Juneau. 6pp.

SKOVLIN, J.M. 1982. Habitat requirements and evalua-
tions. Pages 368-413 in Thomas, J.W. and D.E. Tow-
eill, eds. Elk of North America: Ecology and
Management. Stackpole Books, Harrisburg, PA. 698pp.

SMITH, B.L. 1976. Ecology of Rocky Mountain goats in
the Bitterroot Mountains, Montana. M.S. thesis, Univ.
Montana. 203pp

. 1977. Influence of snow conditions on winter

distribution, habitat use, and group size of mountain
goats. Pages 1 74- 1 89 in Samuel, W. and W.G. Mac-
Gregor, eds. Proc. First Int. Mountain Goat Symp.
Kalispell, MT.

SMITH, C.A. and K.J. RAEDEKE. 1982. Group size and
movements of a dispersed, low density goat popula-
tion, with comments on inbreeding and human im-
pacts. Pages 54-67 in Bailey, J.A. and G.G.
Schoonveld, eds. Proc. Third Biennial Symp., Northern
Wild Sheep and Goat Council. Fort Collins, CO.

SMITH, KG. 1982. Winter studies of forest-dwelling
mountain goats of Pinto Creek, Alberta. Pages 374-
390 in Bailey, J.A. and G.G. Schoonveld, eds. Proc.
Third Biennial Symp., Northern Wild Sheep and Goat
Council. Fort Collins, CO.

SMITH, T.E. 1976. Reproductive behavior and related

social organization of the muskox on Nunivak Island.
M.S. thesis, Univ. Alaska, Fairbanks. 1 38pp.

SOULE, ME. and B.A. WILCOX, eds. 1980. Conservation
biology, an evolutionary-ecological perspective. Sin-
auer Assoc, Sunderland, MA. 395pp.

SOWLS, L.K 1978. Collared peccary. Pages 191-205 in

Schmidt, J.L. and D.L. Gilbert, eds. Big Game of North
America: Ecology and Management. The Wildl. Man
age. Inst., Washington, DC. and Stackpole Books.
Harrisburg, PA.

1984. The peccaries. Univ. Arizona Press, Tucson.

251pp
SPENCER, D.L. and C. LENSINK 1970. The muskox of
Nunivak Island. J. Wildl. Manage. 34:1-15.

SPRAKER, T.R. and C.P. HIBLER 1977. Summer lamb
mortality of Rocky Mountain bighorn sheep. Desert
Bighorn Council Trans. 21:11-12.

and . 1982. An overview of the clinical

signs, gross and histological lesions of the pneumonia
complex of bighorn sheep. Proc. Biennial Symp.,

Northern Wild Sheep and Goat Council. Fort Collins,
CO. 3:163-172.
— , , G.G. SCHOONVELD, and W.S. ADNEY.

1984. Pathologic changes and microorganisms found
in bighorn sheep during a stress-related die-off. J.
Wildl. Diseases 20:319-327.

STELFOX, J.G. 1971. Bighorn sheep in the Canadian Rock-
ies: A history 1800-1970. Canadian Field-Naturalist
85:101-122.

. 1976. Wood Buffalo National Park, bison research

1972-76. Section A-I in 1976 Annual Report. Cana-
dian Wildl. Serv. and Parks, Ottawa.

SUNDSTROM, C. 1968. Water consumption by pronghorn
antelope and distribution related to water in Wyo-
ming's Red Desert. Proc. Biennial Antelope States
Workshop 3:39-46.

SWEENEY, J.M. and H.W. STEINHOFF. 1976. Elk move-
ments and calving as related to snow cover. Pages
415-436 in Steinhoff, H.W. and J.D. Ives, eds. Ecologi-
cal Impacts of Snowpack Augmentation in the San
Juan Mountains, Colorado. Colorado State Univ., Fort
Collins.

TABER, R.D. and R.F. DASMANN. 1958. The black-tailed
deer of the chaparral — its life history and manage-
ment in the North Coast Range of California. Califor-
nia Dep. Fish and Game, Game Bull. 8. 163pp.

TANKERSLEY, N.G. and W.C GASAWAY. 1983. Mineral
lick use by moose in Alaska. Canadian J. Zoology
61:2242-2249.

TELFER, E.S. 1974. Logging as a factor in wildlife ecology
in the boreal forest. Forest Chronical 50:186-190.

TENER, J.S. 1965. Muskoxen in Canada. Canadian Wildl.
Monogr. Series 2. Dep. Northern Affairs and Nat.
Resour. Queens Printer, Ottawa. 166pp.

THING, H. 1984. Food and habitat selection by muskoxen
in Jameson Land, Northeast Greenland: A preliminary
report. Pages 69-74 in Klein, D.R., R.G. White, and
S. Keller, eds. Proc. First Int. Muskox Symp. Spec.
Rep., Biol. Pap. Univ. Alaska, Fairbanks.

and B. CLAUSEN. 1980. Summer mortality among

caribou calves in Greenland. Pages 434-437 in Rei-
mers, E., E. Gaare, and S. Skjenneberg, eds. Proc. Sec-
ond Int. Reindeer/Caribou Symp. Dir. for Vilt og
Ferskvannsnsk, Trondheim.

THOMAS, D.C. and J. EDMONDS. 1984. Pages 93-100 in
Klein, D.R., R.G. White, and S. Keller, eds. Proc. First
Int. Muskox Symp. Spec. Rep., Biol. Pap. Univ. Alaska,
Fairbanks.

, F.L. MILLER, R.H. RUSSELL, and G.R. PARKER.

1981. The Bailey Point Region and other muskox
refugia in the Canadian Arctic: A short review. Arctic
34:34-36.

THOMAS, J.W., H. BLACK Jr., R.J. SCHERZINGER, and R.J.
PEDERSEN 1979. Deer and elk. Pages 104-127 in
Thomas, J.W., ed. Wildlife Habitats in Managed For-
ests: The Blue Mountains of Oregon and Washington.
U.S. Dep. Agric, For. Serv. Handbook 553- Washing-
ton, DC.

and D.E. TOWEILL. 1982. Elk of North America:

Ecology and Management. Stackpole Books, Harris-
burg, PA. 698pp.

THOMPSON, R.W. 1980. Population dynamics, habitat

utilization, recreational impacts and trapping of intro-
duced Rocky Mountain goats in the Eagles Nest Wil-
derness Area, Colorado. Pages 459-464 in Hickey,
W.O., ed. Proc. 1980 Biennial Symp., Northern Wild
Sheep and Goat Council. Salmon, ID. 668pp.

Ungulates

563

and R.J. GUENZEL. 1978. Status of the introduced

mountain goats in the Eagles Nest Wilderness Area,
Colorado. Pages 175-197 in Hebert, D.M. and M.
Nation, eds. Proc. 1978 Biennial Symp. Northern Wild
Sheep and Goat Conf. Penticton, British Columbia.

TIMMERMANN, H.R. 1974. Moose inventory methods: A
review. Naturaliste Canadien 101:615-629.

TODD, J. W. 1972. A literature review on bighorn sheep
food habits. Colorado Div. Game, Fish, and Parks
Spec. Rep. 27. 21pp.

TREFETHEN, J.B., ed. 1975. The wild sheep in modern
North America. Winchester Press, New York, NY.
302pp.

. 1975. The Wild Sheep in Modern North America.

Proc. Workshop on the Manage. Biol. North American
Wild Sheep. Univ. Montana, Missoula, June 18-20,
1974. Boone and Crockett Club and the Winchester
Press, New York, NY. 302pp.

TURNER, J.C and RA. WEAVER. 1980. Water. Pages 100-
112 in Monson, G. and L. Sumner, eds. The Desert
Bighorn — Its Life History, Ecology, and Management.
Univ. Arizona Press, Tucson.

U.S. DEPARTMENT OF THE INTERIOR, FISH AND
WILDLIFE SERVICE. 1983. Initial report, baseline
study of the fish, wildlife and their habitats. Arctic
Natl. Wildl. Refuge Coastal Plain Resour. Assessment.
U.S. Dep. Inter., Fish and Wild. Serv., Anchorage, AK
507pp.

VAN DAELE, L.J. and DR. JOHNSON. 1983. Estimation of
arboreal lichen biomass available to caribou. J. Wildl.
Manage. 47:888-890.

VAN DYKE, W.A., A. SANDS, J. YOAKUM, A. POLENZ, AND
J. BLAISDELL. 1983. Wildlife habitats in managed
rangelands, the Great Basin of southeastern Oregon:
Bighorn sheep. U.S. Dep. Agric, For. Serv., Pacific
Northwest Experiment Sta., Gen. Tech. Rep. PNW-
159. 37pp.

VIBE, C. 1958. The muskox in east Greenland. Mammalia
22:168-174.

. 1967. Arctic animals in relation to climatic fluctua-
tions. Meddel om Gronland 171:1-227.

WAKELYN, LA. 1984. Analysis and comparison of existing
and historic bighorn sheep ranges in Colorado. M.S.
thesis, Colorado State Univ., Fort Collins. 274pp.

WALLMO, O.C., ed. 1981. Mule and black-tailed deer of
North America. Univ. Nebraska Press, Lincoln. 605pp.

. 1981. Mule and black-tailed deer distribution and

habitats. Pages 1-25 in Wallmo, O.C., ed. Mule and
Black-tailed Deer of North America. Univ. Nebraska
Press, Lincoln.

, L.C. CARPENTER, W.L. REGELIN, R.B. GILL, and

D.B. BAKER. 1977. Evaluation of deer habitat on
a nutritional basis. J. Range Manage. 30:122-127.

, R.B. GILL, L.H. CARPENTER, and D.W. REICHERT.

1973- Accuracy of field estimates of deer food habits.
J. Wildl. Manage. 37:556-562.

and J.W. SCHOEN. 1981. Forest management for

deer. Pages 434-448 in Wallmo, O.C., ed. Mule and
Black-tailed Deer of North America. Univ. Nebraska
Press, Lincoln. 605pp.

WHITE, G.C., DR. ANDERSON, KP. BURNHAM, and D.L.
OTIS. 1982. Capture-recapture and removal methods
for sampling closed populations. Los Alamos Natl.
Laboratory, Los Alamos, NM. 235pp.

WIGAL, RA. and V.L. COGGINS. 1982. Mountain goat.
Pages 1008-1020 in Chapman, J.A. and GA. Feld-
hamer, eds. Wild Mammals of North America. Johns
Hopkins Univ. Press, Baltimore, MD. 1 147pp.

WILKINSON, P.F., C.C. SHANK and D.E. PENNER. 1976.
Muskox-caribou summer range on Banks Island,
Northwest Territories. J. Wildl. Manage. 40:151-162.

and P.N. TEAL. 1984. The muskox domestication

project: An overview and evaluation. Pages 162-166
in Klein, DR., R.G. White, and S. Keller, eds. Proc.
First Int. Muskox Symp. Spec. Rep., Biol. Pap. Univ.
Alaska, Fairbanks.

WILSON, L.W., J. BLAISDELL, G. WELSH, R. WEAVER, R.
BRIGHAM, W. KELLY, J. YOAKUM, M. HINKS, J.
TURNER, and J. DeFORGE. 1980. Desert bighorn
habitat requirements and management recommenda-
tions. Trans. Desert Bighorn Council 24:1-7.

WISDOM, M.J., L.R. BRIGHT, C.G CAREY, WW. HINES,
R.J. PEDERSEN, DA. SMITHEY, J.W. THOMAS, and
F.W. WITMER. 1986. A model to evaluate elk habitat
in western Oregon. U.S. Dep. Agric, For. Serv., Pacific
Northwest Region. 36pp.

WITMER, G.W., M. WISDOM, E.P. HARSHMAN, R.J. AN-
DERSON, C. CAREY, MP. KUTTEL, ID. LUMAN, J.A.
ROCHELLE, R.W. SCHARPF, and D. SMITHEY. 1985.
Deer and elk. Pages 231-258 in Brown, E.R., ed. Man-
agement of Wildlife and Fish Habitats in Forests of
Western Oregon and Washington. U.S. Dep. Agric,
For. Serv. Pub. R6-F&WL-192-1985. Portland, OR.

WOOD, J.E., T.S. BICKLE, W. EVANS, J.C. GERMANY, and
V.W. HOWARD, Jr. 1970. The Fort Stanton mule deer
herd (some ecological and life history characteristics
with special emphasis on the use of water). New
Mexico State Univ., Agric. Experiment Sta. Bull. 567.
32pp.

YOAKUM, J.D. 1978. Pronghorn. Pages 103-122 in

Schmidt, J.L. and D.L. Gilbert, eds. Big Game of North
America: Ecology and Management. The Wildl. Man-
age. Inst., Washington, DC. and Stackpole Books,
Harrisburg, PA. 494pp.

. 1980. Habitat management guides for the Ameri-
can pronghorn antelope. U.S. Dep. Inter., Bur. Land
Manage. Tech. Note 347. Denver Federal Center,
Denver, CO. 77pp.

564

Ungulates

IV HABITAT MEASUREMENTS

26 Soils

27 Terrestrial Physical Features

28 Aquatic Physical Features

29 Hydrologic Properties

30 Water Quality

31 Vegetation

32 Macroinvertebrates

m

■■■

26

SOILS

James E. Stone

U.S. Bureau of Land Management
Washington, DC 20240

". . . there is a remarkable correlation between game
supply and soil fertility throughout North America."

— Aldo Leopold, from Game Management

Editor's Note: Aldo Leopold noted the relationship
between soils and wildlife populations. Since then,
many wildlife biologists and soil scientists have
observed the linkage between soils, vegetation, and
wildlife populations.

While the concept of relating soils to vegetation
and wildlife is well established, the predictive rela-
tionships are not. This chapter brings together in-
formation available on these relationships.

INTRODUCTION

Soil is a vital part of wildlife habitat. Indirectly,
soil affects animals through its influence on the com-
position, amount, palatability, and nutritive value of
vegetation. In addition, soil properties directly affect
such things as the ease of digging, burrow stability
and depth, maintenance of body temperature and
moisture levels, availability of suitable reproductive
and resting sites, mobility, and body coloration.

Until recently, the consideration of soils on
wildlands has been hampered by the lack of available
soil information, but this situation is rapidly chang-
ing. Millions of acres of lands administered by the
U.S. Bureau of Land Management and the U.S. Forest
Service have now been surveyed, and ongoing pro-
grams are in place. Acquiring the soil information,
however, is only the beginning. Putting this informa-
tion to use is the ultimate goal.

Several years ago, Klemmedson ( 1970) wrote of
the communication gap between those who make
soil surveys and those who could and should be
using the information. Although progress has been
made, the problem still exists. Effective application
of soils data in habitat management requires a coop-
erative effort on the part of both soil scientists and
wildlife biologists. Soil scientists must recognize the
needs of other disciplines and tailor surveys and
associated soil interpretations to meet those needs.
Wildlife biologists in turn must acquire at least a
working knowledge of soil terminology, kinds and
sources of soil information, and soil-animal relation-
ships so as to fairly assess application opportunities
as well as constraints.

Soils

567

The intent of this chapter is to provide a work-
ing-level discussion of the use of soil information
in habitat characterization and management.

SOIL DESCRIPTION

Soil is a mixture of solids and pores. The solid
fraction consists of mineral particles from the weath-
ering of rocks and organic material from the deposi-
tion and decay of plant and animal tissue. The
intervening pore spaces contain either air or an
aqueous solution with substances either dissolved or
in suspension. The nature and relative amounts of
these constituents are described according to spe-
cific standards set forth in the Soil Survey Manual
(U.S. Soil Conservation Service 1981 ). A representa-
tive soil description is given in Figure 1; descriptive
terminology and standards are summarized below for
selected soil characteristics.

Parent Material

Parent material is the unconsolidated and more
or less chemically weathered mineral material from
which soils develop. Parent materials may have been
formed in place or transported by water, wind, ice,
or gravity. In any event, parent material exerts a
strong influence on soil development and associated
plant and animal communities. For example, soils
derived from serpentine parent materials are gener-
ally low or lacking in several essential plant nutrients
and are therefore highly selective for native plant
species (Kruckeberg 1969). Notable plant-soil rela-
tionships are also evident on other parent materials
(Marchand 1973). In addition, the nature of the par-
ent rock itself influences the availability of special-
ized habitats such as caves, lava tubes, and talus
(Maser et al. 1979).

DEPTH, DRAINAGE
PARENT MATERIAL

HORIZONATION
ROOTS

TEXTURE
STRUCTURE
ROCK CONTENT

COLOR
CONSISTENCE

REACTION (pH)

Kech series

The Kech series consists of shallow, well-drained soils that formed
in place and from locally transported sediment from sandstone and
interbedded shale. Kech soils are on upland hills and ridges. Slopes
range from 3 to 40%. Average annual precipitation is about 14 inches,
and average annual air temperature is about 44°F.

Kech soils are similar to Progresso and Potts soils and are near
Scholle, Mesa, and Agua Fria soils. Progresso soils have sandstone
bedrock at a depth of 20 to 40 inches. Potts, Scholle, Mesa, and Agua
Fria soils are more than 60 inches deep. Mesa soils are dry. Agua
Fria soils have more than 35 percent clay in the B horizon.

Typical pedon of Kech loam, in an area of Kech-Progresso loams, 3
to 15 percent slopes, SW1/4 Sec. 13, T. 15 S., R. 93 W., in Delta
County:

Al — 0 to 4 inches: brown (7.5YR 5/2) loam, dark brown (7.5YR 4/2)
moist; moderate very fine granular structure; soft, friable, slightly
sticky and slightly plastic; few fine and medium roots: 5 to 1 5%
sandstone channers; neutral; clear smooth boundary . (3 to 5
inches thick).

B2t — 4 to 12 inches; brown (7.5YR 5/-t ) clay loam, dark brown
(7.5YR 4/4) moist; moderate medium subangular blockv
structure parting to moderate fine subangular blocky; hard,
friable, sticky and plastic; few fine and very fine roots; moderate
continuous clay films on peds; 5 to 10% sandstone channers:
noncareous; mildly alkaline; clear wavy boundary. (4 to 9 inches
thick).

B3ca — 12 to 19 inches; pinkish white (7.5YR 8/2) channery loam,
pink (7.5YR 7/4) moist; weak medium subangular blocky
structure parting to weak fine subangular blocky; hard, firm,
slightly sticky and slightly plastic; 25 to 30% sandstone channers;
few thin patchy clay films; soft masses of secondary lime; mildy
alkaline; clear irregular boundary (4 to 8 inches thick)

R — 19 inches; partially weathered calcareous sandstone.

The A horizon has value of 4 to 6 dry and 3 to 5 moist and has
chroma of 2 or 3 dry and moist. The B horizon has hue of 7.5YR or
10YR, value of 5 or 6 dry and 4 or 5 moist, and chroma of 3 or 4 dry
and moist. Texture of the B horizon is clay loam. Clay content ranges
from 27 to 35% in this horizon, and sandstone channers range from
0 to 15%. Reaction is neutral or mildly alkaline. The B3 horizon
contains visible secondary lime and has as much as 30% sandstone
channers. Weathered sandstone bedrock is at a depth of 10 to 20
inches.

h

SETTING

ASSOCIATED
SOILS

TYPE
LOCATION

PROFILE
DESCRIPTION

RANGE OF
CHARACTERISTICS

Figure 1. Soil description.

568

Soils

Horizonation

Vertical sections through the soil (soil profile)
often reveal the presence of distinct layers or hori-
zons (Figure 2). The nature, arrangement, and thick-
ness of horizons (or their absence) are important
for differentiating soils and for interpreting soil be-
havior. Each soil horizon, designated by a particular
symbol (e.g., Al — 0 to 4 in.), is described separately.

••> ri-- '■%<&*$;> ' "J, '- i -'r' * ' ■ ;i i -~

mm*

Clay(%)

Silt(%)

Sand(%)
Figure 3. Soil texture.

pendence upon soil texture of 18 races of rodents.
Feldhamer (1979) reported a direct relationship
between the population density of pocket mice (Per-
ognathus parvus) and the percentage of sand in the
soil; an inverse relationship existed with the percent-
age of clay.

Figure 2. Soil horizonation.

Texture

Texture refers to the size distribution of mineral
particles within the soil. Those particles smaller than
2 mm (0.08 in.) in diameter, termed the fine earth
fraction, are described separately from larger rock
fragments. Soils are assigned textural class names
(e.g., sandy loam, silty clay) based on the amounts of
sand-, silt-, and clay-sized particles in the fine earth
fraction (Figure 3).

Small ground-dwelling animals often show an
affinity for particular soil textures. Stuart (1932)
related the distribution of several lizard species to
soil texture; Hardy (1945) observed a varying de-

Mojave fringed-toe lizard. Completely restricted to fine,
loose, windblown sand.

Rock Fragments

Rock fragments, those mineral particles greater
than 2 mm (0.08 in.) in diameter, are described
according to their size and shape (Figure 4). The
adjective form of a rock fragment class is often used
to modify the textural class name (e.g., channery
loam ). Rock content in the profile is described as a
percentage by volume.

Rock fragments in the soil profile reduce mois-
ture storage as well as the volume of soil material

Soils

569

Fragment

Centimeters

Diameter

Length

Inches

(f\p

Gravel

0.2-7.6

0.1-3

Cobble

7.6-25

3-10

Channer
(flat)

0.2-15

0.1-6

~^ Flagstone
(flat)

15-38

6-15

Stone

25-60

10-24

Boulder

60

24

aA single fragment of this size is called a "pebble.
Figure 4. Rock fragments.

that plant roots and animals can occupy. Fragments
on the surface can provide specialized habitats for
animals in areas that might otherwise be unsuitable.

Structure

Soil structure refers to the organization of indi-
vidual mineral particles into clumps or aggregates.
Some soils, such as loose sands, are structureless.
Structured soils are described by the stability, size,
and shape of the aggregates (e.g., moderate, medium,
subangular blocky structure [Figure 5]). Structure
affects the size of pores and the total amount of pore
space in the soil, thus influencing air and water rela-
tions. Unlike texture, soil structure can be altered
rapidly by use and management practices, thereby al-
tering habitat suitability for some animals.

Consistence

Soil consistence characterizes the cohesion
among soil particles and the adhesion between soil

and other substances. One element of consistence is
soil strength, the degree of resistance to breaking
or crushing when force is applied. The standard
terms used to describe soil strength are specific to
moisture condition (Table 1).

Soil strength is important when considering
habitat suitability for small burrowing animals. For
example, Feldhamer ( 1979) found a direct correla-
tion between the population density of chipmunks
(Eutamias minimus) and soil strength.

Color

Soil color is described by comparison with
standard Munsell color charts (Munsell 1975). The
Munsell System uses three elements (hue, value,
chroma) to make up a specific color notation. For
example, 7.5 YR 8/2 is the hue and value/chroma
notation for pinkish white (Figure 6). Colors of both
moist and dry soil are commonly recorded in the
soil description.

570

Soils

Structural Type

General Connotation

Granular

Found in surface horizons; formation promoted by organic
matter; encourages air and water transmission.

Platy

Found in surface and subsurface horizons; severely
restricts air and water transmission.

Blocky

Found in fine-textured subsurface horizons; most common
in humid climates; restricts air and water transmission.

Prismatic

Found in fine-textured subsurface horizons; most common
in arid and semiarid climates; restricts air and water
transmission.

Columnar

Found in fine-textured subsurface horizons; most common
in arid and semiarid climates; often indicates high sodium
levels; restricts air and water transmission.

Figure 5. Soil structure.

Table 1. Soil consistence (U.S. Soil Conservation Service 1981, Chap. 4).

Descriptive Term (by moisture condition)

Connotation

Air Dry

Field Capacity (moist)

Loose

Loose

No specimen can be obtained.1

Soft

Very friable

Crushes or breaks under slight force ap-
plied by thumb and forefinger.2

Slightly hard

Friable

Crushes or breaks under moderate force by
thumb and forefinger.

Slightly hard

Firm

Crushes or breaks under moderate force by
thumb and forefinger.

Hard

Very firm

Crushes or breaks under strong force by
thumb and forefinger.

Very hard

Extremely firm

Cannot be crushed or broken by thumb
and forefinger; can be broken by squeezing
slowly between hands.

Extremely hard

Extremely hard

Cannot be crushed or broken in hands; can
be crushed or broken underfoot

1A specimen for testing strength is normally a cube approximately 1 cm on a side

2Under laboratory conditions, strength classes are defined quantitatively by measured force in newtons

Soils

571

MUNSELL- SOIL COLOR CHART

5YR

3/

2.5/

■ ■■■

/I / /3 A /6 /8

CHROMA >■

Figure 6. Soil color.

Several researchers have found a distinct rela-
tionship between soil color and the pelage colora-
tion of various mammals (Dice and Blossom 1937;
Hardy 1945). For woodrats (Neotoma sp. ) and chip-
munks (Eutamias sp.) in the South Dakota badlands,
the color relationship was strongest on sites largely
devoid of vegetative cover (Stebler 1939). Soil color
can also exert a strong influence over surface
temperatures.

Depth

Soil depth is measured from the surface to a
restricting or contrasting layer. The nature of the
layer in question is generally specified (e.g., depth to
hardpan ). Where the nature of the layer is not given,
it is understood to be consolidated bedrock. Stand-
ard classes for depth are given in Figure 7.

Soil depth and soil texture influence the size
and distribution of pocket gophers (Thomomys sp.,
Geomys sp.) in the western U.S. (Davis 1938; Davis
et al. 1938). Hardy (1945) contrasted the soil depth

Chipmunk.

Depth Class

Very
shallow

Shallow

Moderately
deep

Deep

Very
ik deeP

Centimeters Inches

< 25

25-50

50-100

100-150

> 150

< 10

10-20

20-40

40-60

> 60

Figure 7. Soil depth.

requirements of two species of kangaroo rats (Dipo-
domys sp.). Additionally, soil depth can affect ani-
mals indirectly through its influence on the kind and
amount of vegetation ( McColley and Hodgkinson
1970; Passey et al. 1982).

572

Soils

,

Table 2. Root quantity and size classes (U.S. Soil
Conservative Service 1981, Chap. 4).

Ord's kangaroo rat.

Roots and Animal Traces

The quantity and size of plant roots in each soil
layer are routinely recorded in soil descriptions (Ta-
ble 2 ). Evidences of animal activity are also de-
scribed when apparent, although no specific
descriptive standards have been developed. Such
evidences range from surface features (e.g., ant
mounds, burrow openings) to actual structures
within the soil (e.g., tunnels, krotovinas ). Krotovinas
are irregular, tubular streaks or spots in the soil
caused when abandoned animal tunnels become
filled with contrasting materials.

The presence or absence of roots and animal
traces is useful for assessing habitat suitability for
plants and animals. Root abundance can be used to
make inferences about soil moisture conditions as
well as aeration and temperature (Weaver 1977;
Lunt et al. 1973; Daubenmire 1972). The abrupt ter-
mination of roots at a particular level generally de-
fines a physical or chemical barrier. That which
restricts root growth will likely inhibit many burrow-
ing animals and other biological activity.

Root Quantity

Descriptive Term

< 1 per unit area1
1-5 per unit area
> 5 per unit area

Few

Common

Many

Very Fine
Fine
Medium
Coarse

Root Size

(diameter)
< 1 mm
1-2 mm
2-5 mm
> 5 mm

Unit area for very fine and fine roots is 1 square centimeter;
unit area for medium and coarse roots is 1 square decimeter

Drainage

The times and depths at which a soil is wet
exerts a strong influence on biological activities as
well as chemical processes. Soil drainage classes
(Table 3) serve to summarize soil wetness. As gener-
ally defined, however, drainage classes reflect an
agricultural and a geographic (humid-temperate)
bias. Local refinements are often necessary to adapt
the drainage class concept to other climates and land
uses. Drainage is a product of climate, slope, and
landscape position along with soil characteristics.

Reaction

Soil reaction is expressed as pH. The pH influ-
ences the presence of toxic ions in the soil as well
as nutrient availability to plants. Plants growing in
soils with a pH less than 5.2 are likely to be affected
by aluminum toxicity. Soils with a pH greater than
8.5 often have excessive amounts of sodium which
can destroy soil structure and thereby reduce soil
moisture and aeration. In addition, sodium can be
toxic to some plants.

Table 3. Soil drainage classes (adapted from U.S. Soil Conservation Service 1981, Chap. 4).

Descriptive Term

General Connotation

Excessively drained, somewhat excessively
drained

These soils have high to very high rates of water transmission
and retain little moisture for plant use.

Well-drained

These soils have intermediate water transmission rates and reten-
tion capacities.

Moderately well-drained, somewhat poorly
drained

These soils are commonly wet near the surface for at least a
portion of the year.

Poorly drained, very poorly drained

These soils are commonly wet at or near the surface for a con-
siderable part of the year.

Soils

573

Figure 8. Soil reaction and nutrient availability.

Several essential nutrients become less available
to plants as the pH decreases below about 6 or in-
creases above about 7 (Figure 8). This effect can
indirectly affect wildlife through the forage they
consume. For example, many arid soils of the West
have pH values between 7 and 8.4. Within this
range, phosphorus availability is reduced through the
formation of insoluble calcium phosphates. Short
( 1979) suggested that declines in herds of mule deer
(Odocoileus hemionns) in the southwestern U.S.
might be caused in part by the low available phos-
phorus levels in the soils.

Soil pH and associated potassium levels partially
account for variation in peak densities in Microtus
pennsylvanicus (Krebs et al. 1971). The distribution
of several species of salamanders has also been re-
lated to soil pH (Mushinsky and Brodie, Jr. 1975;
Batson 1965).

Other Soil Characteristics

Additional soil properties and characteristics,
although not specifically stated in a standard soil
description, are nevertheless commonly determined
during the course of a soil survey. Many of these
characteristics are not directly observed on a soil
profile, but can be inferred from other properties or
measured through laboratory analyses of soil
samples.

Available Water Capacity. Within the plant-rooting
zone, the amount of water that can be held between
1/3 bar and 15 bars of tension (1 bar = 0.99 atm)
is the water retention difference. The amount of
moisture at 1/3 bar, termed the field capacity,
approximates the moisture that would remain in the
soil after natural drainage occurred in response to
gravity alone. At this point, the larger soil pores
would be empty while the smaller pores would be
filled with water. Soil water held at tensions of more
than 15 bars (wilting point) is unavailable to many
plants; hence, the moisture content between 1/3 and
15 bars is commonly referred to as the available
water capacity (AWC). Some plants, however, most
notably greasewood, saltbush, shadscale, and other
salt desert species, can extract soil moisture at
tensions far exceeding 1 5 bars.

The available water capacity varies with soil
texture, structure, rock fragment content, and other
soil properties. In general, however, fine-textured
soils such as clays, clay loams, or silt loams have a
high AWC whereas coarse-textured soils such as
sands or sandy loams have a low AWC. The AWC
does not reflect actual soil moisture content, but
rather the potential to retain moisture within the de-
fined tension limits. Thus, a sandy loam in New Mex-
ico and New York might well have similar available
water capacities, yet contain vastly different amounts
of moisture at any given time due to different pre-
cipitation regimes. Nevertheless, the AWC can be
useful in evaluating wildlife habitat on a local basis
(Figure 9).

So/7 Series
LEAPS

PARLIN
DUFFSON

SPRING
CREEK

FOLA

AVAILAE

0-10"
10-48"

silty clay loam
silty clay

0-11"

11-31"
31-48"

channery loam
channery clay loam
very stony loam

0-8"

8-16"

16-30"

30"

loam

clay loam
loam
bedrock

0-9"

9-19"

19"

St

b€

Dny loar
avelly Ic
sdrock

n

•am

0-6"
6-48"
I I

cc

ve

I

bbly sandy loam
ry cobbly sandy loam
I 1 I I I I

012345678

3LE WATER CAPACITY (inches)

Figure 9.

The potential of selected soils to retain
moisture.

574

Soils

Seasonal Water Table. A seasonal water table is
marked by the upper limit of the soil that is
saturated with water for a specified period during
most years. For example, a particular soil may be
characterized as having a seasonally high water table
at a depth of 2 feet, commonly during May and June.
The presence or absence of a seasonally high water
table can alter the suitability of a site for burrowing
animals. It can also affect the vertical distribution
within the soil of certain animals such as
salamanders (Taub 1961).

Hydrologic Soil Group. The hydrologic soil group
provides an interpretation of soil behavior in terms
of water infiltration, transmission, and runoff.
Infiltration is the rate at which water enters the soil
surface, transmission is the rate of water movement
within the soil, and runoff is the amount of water
that moves over the surface to depressions or
defined channels without entering the soil. For
standardized conditions of bare soil after prolonged
wetting, individual soils are assigned to one of four
hydrologic soil groups, A, B, C, or D (Figure 10).

be suitable for the Merriam kangaroo rat (Dipodo-
mys merriatni). He speculated that this was due to a
salt-derived surface crust.

Sodium Adsorption Ratio. The sodium adsorption
ratio (SAR) relates the amount of sodium in the soil
to the amount of calcium and magnesium. As noted
earlier, high sodium can be toxic to plants as well as
harmful to the physical condition of the soil. The
SAR is particularly useful in assessing reclamation
potential of mined lands or other severely disturbed
areas. Most State and Federal reclamation guidelines
for topsoil and overburden suitability include the
SAR as an evaluation factor. In general, SAR values of
12 or more suggest unsuitability or at least severe
limitations for successful reclamation (Fisher and
Deutsch 1983).

Erodibility. Soil erosion is influenced by several
non-soil factors such as slope, precipitation, wind
velocity, and vegetative cover. Nevertheless, some
soils erode more readily than others even when all
other factors are the same.

Group

Water
Infiltration

Water
Transmission

Water
Runoff

A

(HIGH)

(HIGH)

(LOW)

B

C

D

\
(LC

)W)

(LC

)W)

(Hl(

3H)

Figure 10. Hydrologic soil groups.

Soils in Group A with low runoff potential and
correspondingly high infiltration rates consist chiefly
of deep, well to excessively drained sands and grav-
els. In contrast, soils in Group D, having high runoff
potential and low infiltration, are often clayey, have a
permanent high water table, or are shallow over
impervious or nearly impervious material.

Salinity. Salinity refers to the amount of salts in the
soil, the standard measure of which is the electrical
conductivity (EC) of an extract of the soil solution.
In many situations, however, salinity is inferred and
expressed as salinity classes covering a range of EC
values.

Soil salinity exerts a strong influence over vege-
tation; as might be expected, the higher the salinity,
the greater the effect (Table 4). Salts in the soil can
also affect certain animals directly. Hardy (1945)
noted the lack of digging activity on sites thought to

The soil erodibility factor (K) reflects this inher-
ent susceptibility of a soil to erode under the action
of raindrop impact and water flowing over the sur-
face (sheet and rill erosion). The wind erodibility
group ( WEG) reflects the susceptibility of a soil to
erode under wind action. Although inherent soil
erodibility varies with several different soil proper-
ties, a rough relationship exists between soil texture
and both the K factor and the WEG (Figure 11).

Erosion presents dual problems in habitat man-
agement. First, soil loss can result in reduced forage
production because of the removal of fertile topsoil.
Second, soil lost from one site is eventually depos-
ited elsewhere. During transport, suspended soil
particles and attached ions can degrade water qual-
ity. Upon deposition, eroded soil material can further
affect aquatic habitats. For example, the distribution
of two species of frogs (Rana sp. ) varied with soil
texture, apparently due to different erodibilities and
associated sediment burdens in adjacent streams
(Lynch 1978).

Soil Fertility. Soil fertility is the inherent capability
of a soil to supply nutrients to plants in adequate
amounts and in suitable proportions (Buckman and
Brady 1969). As discussed earlier, soil pH serves
as an indicator of nutrient availability. Soil fertility is
more specifically characterized through laboratory
analyses of extractable bases (e.g., Ca, Mg, Na, K),
phosphorus, nitrogen, and determinations of cation
exchange capacity, and percentage-base saturation
(see Sources of Soil Information section).

Soil fertility has long been known to influence
the palatability of plants. For example, the volatile oil

Soils

575

Table 4. Soil salinity classes (U.S. Soil Conservation Service 1983, Part 603).

Electrical

Conductivity

(millimhos/cm)

Salinity Class

Connotation1

0-2

Nonsaline

No effect on plants.

2-4

Very slightly saline

Effect on plants is mostly negligible; only most sensitive
plants affected adversely.

4-8

Slightly saline

Many plants are adversely affected but diversity of adapted
species is guite high (e.g., four-wing saltbush, winterfat,
galleta).

8-16

Moderately saline

Plant communities are dominated by salt-tolerant species
(e.g., greasewood, mat saltbush, western wheatgrass).

> 16

Strongly saline

Few plants are adapted (e.g., shadscale, iodine bush, alkali
sacaton).

1U.S. Salinity Laboratory Staff 1954, McArthur et al. 1978; Thornburg 1982

content of sagebrush leaves seems to vary with soil
depth and fertility among other non-soil factors
(Powell 1970; Nagy 1966). This is of particular im-
portance because high volatile oil content inhibits
the activity of rumen microorganisms (Nagy et al.
1964). Some of the more dramatic palatability effects
are observed when commercial fertilizers are added
to nutrient deficient soils. Gessel and Orians (1967)
observed that certain rodents selectively fed on ni-
trogen-fertilized trees while avoiding adjacent unfer-
tilized trees or those fertilized with potassium.

In a series of studies spanning several years in
Missouri, the abundance, distribution, health, and
size of several animals including wild turkeys, rab-
bits, raccoons, muskrats, opossums, and squirrels
were linked to soil fertility (Denny 1944; Crawford
1950; Albrecht 1957). Low soil fertility, particularly
levels of nitrogen and phosphorus, has previously
been noted as a possible cause of the decline in
herds of mule deer in the Southwest (Short 1979).

Relationships between animals and soil fertility
are not always predictable. For example, a pocket
gopher (Thomomys bottae) has been found in abun-
dance on serpentine-derived soils in contrast to adja-
cent non-serpentine soils (Proctor and Whitten
1971 ). As noted previously, serpentine-derived soils
are generally low or lacking in several essential plant
nutrients. This apparent anomaly was explained by
the presence of Brodiaca sp., a serpentine-tolerant
plant whose corm serves as a primary food source
for the gophers.

SOIL CLASSIFICATION

Soil taxonomy, although foreign to most biolo-
gists, is as much a scientific necessity as the taxon-
omic systems used to classify plants and animals. The
soil taxonomic system used in the U.S. groups soils
according to measurable characteristics such as
depth, texture, temperature and moisture regimes,
and chemical properties. Like botanical and zoologi-
cal systems, soil taxonomy provides a uniform means
to identify a particular soil and facilitates the transfer
of both soils information and that which can be re-
lated to soils.

The system is hierarchical, consisting of six lev-
els or categories including order, suborder, great
group, subgroup, family, and series. At the highest or
most general level there are 10 soil orders. At the
lowest and most specific level, over 12,000 series
have been identified in the U.S. alone, and the num-
ber is increasing annually. Table 5 shows the classifi-
cation at all levels for one particular soil. The
nomenclature applied conveys progressively more
information from the order through the family level.
Each term is specifically defined in a handbook out-
lining soil taxonomy (Soil Survey Staff 1975). At
the lowest level, the series name itself has no spe-
cific technical meaning, often being derived from a
place or local landmark near where the soil was first
described. A series description, however, refines
family attributes by further narrowing the allowable
range of individual soil properties ( Figure 1 ).

Soil taxonomy is useful in making habitat gener-
alizations. For example, Kantrud and Kologiski

576

Soils

.10

.17

.20

K FACTOR

.24 .32

.37

.43

.64

WATER

ERODIBIUTY

=>

Sand

Loamy

Clay

Sandy Loam

Sandy Clay,

Loam,

Silt Loam

Silt

Sand

Silty Clay

Clay Loam,
Sandy Clay
Loam

Silty Clay
Loam

LOW

■►HIGH

WIND

ERODIBIUTY

=>

Wet Soils,

Silt

Clay Loam

Loam,

Clay,

Sandy

Loamy

Sand

Rocky

Silt Loam

Sandy Clay,

Silty Clay,

Loam

Sand

Soils

Sandy Clay
Loam

Silty Clay
Loam

6 5 4 3

WIND ERODIBIUTY GROUP

Figure 11. Soil erodibility.

Table 5. Taxonomic system for soil classification (Soil Survey Staff 1975).

Taxonomic
Category

Number in
System'

Example

Order

10

Aridisol

Suborder

45

Argid

Great Group

187

Haplargid

Subgroup

990

Lithic Ustollic Haplargid

Family

5,603

Loamy, mixed, mesic, Lithic Ustollic Haplargid

Series

12,002

< 35% clay in zone of clay
accumulation

"iixea mineral'
mineral other than quartz)

vo of any one

mean annual soil temp, between 8° and 15° C
(at 50 cm or bedrock, whichever shallower)

depth to bedrock is < 50 cm (20 in.)

is wetter and has more organic matter than normal for
the great group

has no cemented pan or high sodium content
has a horizon of silicate clay accumulation

is dry in plant-root zone for at least half of growing season

as of October 1980

Soils

577

(1982) found that average bird density and species
richness were slightly higher on soils of the order
Mollisol than on soils of the order Aridisol. Within
the order Mollisol, significantly different species
densities were found at the soil subgroup level. In
another study, Leopold and Dalke (1941) found that
a single soil series supported over 90% of the wild
turkey population in Missouri.

SOIL MAPPING

Soil surveys are designed and conducted at dif-
ferent intensities to meet different needs. Accord-
ingly, not all surveys provide information at the same
level of detail nor of equal applicability in habitat
management. Failure to recognize this fact can and
has led to the misuse of soil survey information.
Potential users of soil information should be in-
volved in the design of soil surveys. Where soil sur-
veys are already completed, users of the data should
be sufficiently knowledgeable to recognize both the
strengths and the limitations of the soil map.

Mapping Variables

Four main variables, map unit components,
kinds of map units, field procedures, and map scales,
are used in designing soil surveys. The same varia-
bles are useful in assessing the applicability of exist-
ing soil survey information.

Map unit components are bodies of soil that can
be identified as being a member of some class
(taxon) in soil taxonomy or bodies of nonsoil that
can be identified as any of 22 types of miscellaneous
areas (e.g., rock outcrop, badland, riverwash; U.S. Soil
Conservation Service 1981). In practice, soil taxons
and miscellaneous areas are often further refined
by one or more phase modifiers. Phases reflect var-
ious use-related soil and nonsoil attributes such as
surface texture and slope (see examples in Figure
12).

The kind of map unit depends on the number
and arrangement of the individual components on
the landscape (Figure 12). Single-component map
units, termed consociations, are dominated by one
soil taxon or miscellaneous area. Multicomponent
map units, termed either associations or complexes,
contain two or more dominant taxons or miscella-
neous areas. In an association, the named compo-
nents generally lie in a regular geographic pattern
and can be located in the field by visible landscape
features (e.g., hill slopes vs. valley bottoms, north
slopes vs. south slopes). In contrast, the named com-
ponents in a complex either lie in a very intricate
pattern that cannot be determined by visible land-
scape features or are simply too small to be sepa-
rated. The choice of map unit depends on the soil
itself as well as the purpose of the survey and the

level of detail needed. Regardless of the map unit,
however, there will always be some areas of soil or
nonsoil material that do not conform to the named
components. Such areas, termed inclusions, occur
because of map scale limitations or errors in the
placement of map unit boundaries. As a general rule,
inclusions that would differ significantly in use and
management from the named components should
not exceed 15% of the map unit.

The initial step in a soil survey is often to de-
velop a preliminary soil map based on ancillary data
(e.g., geology, landform, adjacent soil maps) and a
knowledge of soil formation and its relationship to
other landscape features. These preliminary map
units are then field checked and revised as necessary.
The specific procedures used in the field work deter-
mine the precision with which map unit components
are identified and described, and the accuracy of
map unit boundaries. Common field procedures in-
clude explicit soil sampling schemes, observations of
exposed soils (e.g., roadcut), landscape features on
the ground, and interpretations of aerial photographs
or other remotely sensed images.

Map scale determines the smallest sized unit
that can be practically delineated on a map. Gener-
ally this is a 1/4 x 1/4 in. square or a circular area of
about 1/16 square inch. By using these guidelines,
the smallest unit that can be delineated at a scale of
1:24,000 (7 1/2'min quadrangle) is about 6 acres
(Table 6). Smaller areas, if important, can be shown
by spot symbols. Such a practice should be used
to indicate special habitat features such as small
ponds, or wet depressions, springs, and escarpments.

Table 6. Guide to map scales and corresponding
minimum size delineations (U.S. Soil
Conservation Service 1981, Chap. 2).

Minimum

Delineation

Scale

Inches/Mile

(acres)

1

7,920

8.0

0.6

1

15,840

4.0

2.5

1

24,000

2.6

6.0

1

31,680

2.0

10.0

1

63,360

1.0

40.0

1

126,720

0.5

160.0

Mapping Intensity

Based on the four variables above, soil surveys
can be grouped by levels of intensity and general
application. Five such levels are commonly recog-
nized ranging from Order 1 surveys, the most inten-

578

Soils

sive, to Order 5 surveys, the least intensive. Any
given survey, however, need not be of a single inten-
sity. In fact, a mixture of intensities will often best
meet user needs most economically. Soil surveys on
BLM-administered lands often contain elements of
Order 2, 3, and 4 mapping intensities. Figure 13
shows how a single area might differ when mapped
at each of these intensity levels.

SOURCES OF SOIL INFORMATION

Perhaps the most obvious source of soils infor-
mation is a soil scientist. Currently, the Bureau of
Land Management employs about 75 soil scientists
distributed among its various field offices, the BLM
Service Center, and the Headquarters Office in
Washington, DC.

map units

single-component

multi-component

components
separable

components
not separable

CONSOCIATION

ASSOCIATION

COMPLEX

CONSOCIATIONS

Avalon loam, 0 to 3 percent slopes

(series taxon with texture and slope phase modifiers)
Torrifluvents, frequently flooded

(great group taxon with flooding phase modifier)
Gullied land

(miscellaneous area with no phase modifier)

ASSOCIATIONS

Progresso — Potts association, 1 to 6% slopes

(series taxons with slope phase modifier)
Typic Haplargrids — Typic Camborthids association

(subgroup taxons with no phase modifier)

COMPLEXES

Lazear — Rock outcrop complex, 3 to 30% slopes

(series taxon and miscellaneous area with slope phase modifier)
Fughes — Curecanti loams, 10 to 40% slopes

(series taxons with texture and slope phase modifiers; where textural

phase is same among components, plural texture phase is used without

the term complex)
Orthents — Rock outcrop complex

(suborder taxon and miscellaneous area with no phase modifier)

Figure 12. Kinds of soil-map units.

Soils

579

1:24,000

ORDER 2

| Rock outcrop

2 Alpha gravelly loam
] Beta gravelly loam

ORDER 3

\^ Alpha-rock outcrop
association

1 :63,360

~] Beta gravelly loam

ORDER 4
XI

b^

Alpha-beta rock
outcrop association

1:126,720

Figure 13. Soil mapping intensity.

Published soils information is concentrated in
soil survey reports. Such reports are generally pub-
lished by the U.S. Soil Conservation Service (SCS) as
part of a formal series, often in cooperation with
State Agricultural Experiment Stations and other
State and Federal agencies. These reports have a
standardized format, the primary elements of which
are listed in Table 7. A nationwide list of published
soil surveys is prepared and periodically updated by
the SCS (U.S. Soil Conservation Service 1982).

Interim reports, of variable content and formal-
ity, are often prepared before formal publication.
This is true on much of the public land administered
by the BLM where field mapping is still in progress
or only recently completed. These reports are avail-
able in the respective District or Resource Area Of-
fices. Special purpose surveys, often of limited areal
extent (e.g., mining and mineral exploration tracts,
research areas), may also be available through pri-
vate companies, universities, or other Federal agen-
cies.

Soil laboratory data in much greater detail than
that commonly found in soil survey reports are pub-
lished as Soil Survey Investigations Reports (SSIR).
They have been issued for many individual States
and groups of States ( U.S. Soil Conservation Service
1972). Similar information is also published by State
Agricultural Experiment Stations in various forms

(e.g., Science Monographs, Miscellaneous Reports,
Special Resource Series).

APPLICATION OF SOIL INFORMATION

A knowledge of soils and their variability across
the landscape can aid in delineating habitat units,
in assessing habitat suitability, and in developing and
evaluating habitat management alternatives.

Habitat Delineation

An initial objective in wildlife inventory is to
delineate land units with the same environmental at-
tributes. The "habitat site" constitutes this unit in
the BLM's Integrated Habitat Inventory and Classifi-
cation System (IHICS; U.S. Bureau of Land Manage-
ment 1982). Once delineated, similar although not
necessarily adjacent habitat sites are aggregated into
so-called standard habitat sites (SHS). Data on animal
occurrence from a few representative habitat sites
are then extrapolated to characterize an entire SHS.
The reliability of this procedure rests on the assump-
tion that any given habitat site is uniform in terms
of those site factors that influence animal
occurrence.

By definition, a habitat site is an area of homoge-
neous (present) vegetation, landform, soils, and cli-
mate (U.S. Bureau of Land Management 1982). In
practice, however, habitat sites and standard habitat
sites are often delineated almost exclusively on the
present vegetation. This approach can result in non-
homogeneous sites which in turn can lead to erro-
neous conclusions on animal occurrence and distri-
bution. For example, overgrazing or other land-use
practices can produce relatively uniform vegetation
over broad areas. With vegetation as the main habitat
criterion, the illusion of habitat uniformity exists
even though the underlying soils may be variable.
For animals sensitive to soil variability, subsequent
habitat characterizations are incomplete at best.

In most instances, neither soils nor vegetation
( nor any other single factor ) can adequately describe
habitat. When all relevant factors are considered
together, however, they are complementary and pro-
vide an added dimension to the understanding of
the habitat unit.

Habitat Suitability

The suitability of a standard habitat site for a
particular wildlife species depends on whether exist-
ing habitat features meet the needs of the animal.
When soil information is used in the delineation and
aggregation of habitat sites, suitability assessments

580

Soils

Table 7. Major sections of a soil survey report (U.S. Soil Conservation Service 1983, Part 605).

Section

General Content

General nature of the survey area

Commonly contains a discussion of the history, climate, physiography,
drainage, and natural resources of the survey area

General soil map and map unit
description

Includes a general soil map (common scale of 1:250,000 or smaller) and
associated map unit descriptions; narrative descriptions commonly give a
brief characterization of the included soils, their relative proportions, and
major inclusions, along with information on elevation, landform, topogra-
phy, average precipitation and air temperature, plant communities, and
dominant use and management considerations.

Use and management

Provides interpretive information, mainly in tabular form, on soil productive
capability, potential plant communities, and suitability and/or limitations
for major land uses and practices.

Soil properties

Commonly includes tabular data on soil texture, rock content, water reten-
tion difference, pH, salinity, wind and water erosion factors, hydrologic
soil group, and other features.

Classification of soils

Gives the taxonomic classification for each soil.

Soil series and morphology

Contains the detailed descriptions for each soil series (or other category if
applicable); Figure 1 was taken from this section of a soil survey report.

can be strengthened. To do so, relationships be-
tween animals and soils must be characterized. More
specifically, the soil-related affinities and tolerances
must be identified for individual animal species or
groups of species.

Applicable information on soil-animal relation-
ships does exist, although it is widely scattered and
poorly developed for actual application. The results
of a limited literature review are shown in Table 8.
The task of accumulating and arraying such informa-
tion in a usable form should be a joint effort be-
tween soil scientists and wildlife biologists. The
logical first step would be an exhaustive review of
pertinent literature. The next step would be to de-
velop a data format to enable practical application. A
sample format patterned after traditional soil inter-
pretations is shown in Table 9. Once identified, rela-
tionships between soil properties and animals can
be used to interpret soil map units or components of
map units. For example, the Kech loam, 3 to 15%
slopes, is a soil map unit whose primary component
is the Kech soil series. If the Kech series exhibits
properties that are suited to a particular animal spe-
cies, all areas mapped as such could be considered at
least potentially suitable for that species. In conjunc-
tion with the evaluation of other factors (e.g., vegeta-
tion), such analyses can be useful in predicting
animal presence or absence, determining probable
ranges, and selecting optimum sampling sites for
supporting studies.

Even when relationships between soils and ani-
mals are unknown, there are opportunities to make
use of existing soils information. For example, where

estimates of wildlife occurrence have been estab-
lished through previous sampling or other studies,
presumed distribution patterns can be superimposed
on soil maps. The objective is to identify any appar-
ent associations between wildlife distribution and
soil map units. At this level of analysis, it is not nec-
essary to know the exact nature of the relationship,
but simply that a relationship is probable. This
knowledge alone can aid in extrapolating data from
sampled to unsampled areas.

For certain animal species, particularly large
ungulates with broad ranges, habitat suitability may
be closely related to the intermingling of various
soils and associated vegetation producing so-called
edge effects or ecotones. Wertz ( 1966) provided an
example of such an analysis from Wisconsin. Individ-
ual soils were characterized by vegetation type and
cover. Natural edge conditions were identified where
contrasting soils occurred together on the landscape.
Ratings based on the number of miles of soil-associ-
ated natural edge were then used to assess the suita-
bility of different management areas for various
wildlife species.

Searching for soil-animal relationships by analyz-
ing several layers of data is a tedious job. This, in
part, accounts for the general lack of species-specific
soil interpretations for wildlife habitat management.
With the increasing availability of computerized soils
and vegetational data bases along with the analytic
capabilities of geographic information systems (GIS),
such analyses should be greatly facilitated.

Soils

581

Table 8. Relationships between soils and wildlife.

Soil Property

Wildlife Characteristic

Wildlife Species

Reference

Texture

Distribution

Kangaroo rats (Dipodomys sp.)

Dale 1939

Texture

Distribution

Pocket mice (Perognathus sp.)

Hardy 1945

Texture

Distribution

Deer mouse {Peromyscus manicula-

tus)
S. grasshopper mouse {Unchomys

Hardy 1945

Texture

Distribution

Hardy 1945

torridus)

Texture

Distribution

Various lizard species

Stuart 1932

Texture

Population density

G. Basin pocket mouse (Perogna-
thus parvus)

Feldhamer 1979

Texture

Population density

Least chipmunk (Tamias minimus)

Feldhamer 1979

Texture

Dust-bathing behavior

Bobwhite quail (Colinus virginianus)

Borchelt and Overmann
1974

Texture

Denning habits

Red fox (Vulpes vulpes)

Sheldon 1950

Texture

Burrow location

Woodchuck (Marmota monax)

Moss 1940

Texture & erodibility

Distribution

Leopard frogs (Rana sp.)

Lynch 1978

Texture & depth

Distribution, animal size

Pocket gophers {Thomomys sp.)

Davis 1938

Texture & depth

Distribution, animal size

Pocket gophers (Geomys sp.)

Davis et al. 1938

Depth

Distribution

Kangaroo rats (Dipodomys sp.)

Hardy 1945

Depth

Distribution

W. harvest mouse (Reithrodontomys

Hardy 1945

megalotis)

Hardy 1945

Depth & strength

Population density

Least chipmunk (Tamius minimus)

Feldhamer 1979

Strength

Various desert rodents

Rosenzweig and Winakur
1969

Drainage

Burrowing habits

Black-tailed prairie dog (Cynomys
ludovicianus)

Sheets et al. 1971

Water table depth

Vertical distribution

Red-backed salamander (Plethodon
cinereus)

Taub 1961

Moisture

Distribution

Voles (Microtus sp.)

Hodgson 1972

Moisture

Distribution

Sagebrush vole (Lagurus curtatus)

O'Farrell 1972

Moisture

Distribution

Ground squirrels (Spermophilus sp.)

Turner 1972

Moisture

Distribution

Kangaroo rats (Dipodomys sp.)

Bienek and Grundmann

1971
Spight 1967

Moisture

Rehydration (water

Various salamander species

balance)

Color

Pelage coloration

Various small mammals

Dice and Blossom 1937

Color

Pelage coloration

Various small mammals

Stebler 1939

Reaction (pH)

Substrate preference

Various salamander species

Mushinsky & Brodie, Jr,
1975

Reaction (pH)

Population density

Meadow vole (Microtus pennsylvan-
icus)

Krebs et al. 1971

Nitrogen level

Feeding habits

Red-backed vole (Clethnonmys

sp.)
Mule deer (docoileus hemionus)

Gessel and Orians 1967

Available phosphorus

Fawn recruitment

Short 1979

Sodium level

Lick preference

Mountain goat (Oreamnos amen-

canus)
Elk (Cervus elaphus)

Stockstad et al. 1953

Sodium level

Lick preference

Stockstad et al. 1953

Sodium level

Lick preference

Mule deer (Odocoileus hemionus)

Stockstad et al. 1953

Sodium level

Lick preference

White-tailed deer (Odocoileus virgi-

Stockstad et al. 1953

nianus)

Stockstad et al. 1953

Sodium level

Peak population density

Microtine rodents

Aumann 1965

Calcium availability

Feeding habits

Ring-necked pheasant (Phasianus
colchicus)

Harper and Labisky 1964

Overall fertility

Distribution, animal weight

Raccoon (Procyon lotor)

Crawford 1950

Overall fertility

Pelt size and quality

Opossum (Didelphis virginianus)

Crawford 1950

Overall fertility

Animal weight, bone
strength

E, cottontail (Sylvilagus flondanus)

Crawford 1950

Overall fertility

Fecundity

E. cottontail (Sylvilagus flondanus)

Williams and Caskey 1965

Overall fertility

Litter size

E. cottontail (Sylvilagus flondanus)

Hill 1972

Overall fertility

Breeding potential

White-tailed deer (Odocoileus virgi-
nianus)

Crawford 1953

Parent material (Ser-

Distribution

Botta's pocket gopher (Thomomys

Proctor and Whitten 1971

pentine)

bottaie)

Soil classification

series
Soil classification

Distribution

Turkey (Meleagns gallopavo)

Leopold and Dalke 1943

Forage preference

Black-tailed deer (Odocoileus Hem-

Whitaker 1965

series

ionus columbianus)

Subgroup

Species richness

Various bird species

Kantrud and Kologiski
1982

582

Soils

Table 9. Sample format for soil interpretations of habitat suitability for a particular animal species.

Soil Property or Soil-
Related Habitat Feature

HABITAT SUITABILITY

Optimum

Marginal

Unsuitable

Texture

Loamy sand, Sandy loam,
Sandy clay loam, Loam

Silt loam, Silt, Silty clay
loam, Clay loam

Clay, Silty clay, Sandy clay,
Sand

Depth (inches)

>40

2CM0

<20

Drainage

Well-drained

Moderately well-drained,
Somewhat poorly drained

Poorly drained, Very poorly
drained, Excessively drained,
Somewhat excessively
drained

Consistence (moist)

Very friable, Friable

Firm

Loose, Very firm, Extremely
firm, Extremely hard

Salinity (mmhos/cm)

0-4

4-8

>8

Depth to high water table (ft)

—

—

<2

Rock fragments 3 in. (wt %)

>15

—

—

Hydrologic soil group

B

C

A, D

Salt-derived surface crust

—

Present

Habitat Management

Habitat managers are naturally concerned with
existing habitat conditions. Management objectives,
however, may well call for more than maintenance
of the status quo. Effective management often neces-
sitates change and with it the need to predict the
likelihood of success, risks involved, and any special
practices that may be desirable. This requires a
knowledge of the present, but equally important, a
knowledge of habitat capability. Habitat capability as
embodied in the concepts of habitat type (Dauben-
mire 1968; Dyksterhuis 1983), and ecological
(range) site (U.S. Soil Conservation Service 1976;
Society for Range Management 1983) is expressed as
the potential plant community. Where sites have
been subjected to little disturbance, potential can
often be described directly or deduced from the
existing plant community. As a result of past land
use, however, present vegetation may bear little
resemblance to the potential. In such situations, soils
may provide the only identifiable link between pres-
ent and potential (Munn et al. 1978; Passey et al.
1982). With soil as the common denominator,
knowledge of potential plant communities can be
extrapolated from a few relict sites to numerous
disturbed or altered sites. In addition, arraying vege-
tational data from several sites on the same or simi-
lar soils reveals the variety of developmental stages,
presumably leading to the same potential plant com-
munity. Comparison of the present with the poten-
tial provides a valuable clue as to site responsiveness
and thereby aids in evaluating opportunities and
choosing among various management alternatives.

The concept of site capability goes beyond an
interest in potential plant communities. Equally im-
portant are soil-related limitations in the selection of
habitat improvement methods and materials and in
the establishment of reasonable use restrictions to
prevent habitat degradation. Some soils have few
limitations. They are often deep, flat-lying, well-
drained, and are free of surface stones or boulders.
Other soils may have yearlong or perhaps seasonal
limitations due to wetness. Surface stones may pre-
clude the use of certain types of seeding equipment
and brush choppers. Hardpans and soils shallow to
bedrock can obviously affect projects requiring exca-
vation. Failure to observe such limitations can dam-
age equipment, increase construction and
maintenance costs, reduce the effective life of an
improvement, and result in unnecessary habitat
degradation.

Soil survey reports commonly rate soils for shal-
low excavations (e.g., pipelines), pond and reservoir
areas, embankments, and as sources of construction
materials (e.g., sand, gravel, topsoil). Many other
interpretations applicable to habitat management are
being developed although there has been insufficient
testing and validation to warrant widespread use.
Examples include soil limitations for rangeland
equipment (e.g., drills, discs), tree planting, fencing,
prescribed burning, and the use of off-road vehicles.
A strong commitment to testing and refining inter-
pretations is essential if they are to be applied with
confidence.

Soils

583

LITERATURE CITED

ALBRECHT, W.A. 1957. Soil fertility and biotic geography.
Geog. Rev. 47(1):86-105.

AUMANN, G.D. 1965. Microtine abundance and soil so-
dium levels. J. Mammal. 46(4):594-604.

BATSON, J.D. 1965. Some comparative behavior studies
on three genera of salamanders. Trans. Ky. Acad. Sci.
25:120-128.

BIENEK, G.K and AW. GRUNDMANN. 1971. Burrowing
habits of two subspecies of Dipodomys merriami
in California and Utah. Great Basin Nat. 3 1(3): 190-
192.

BORCHELT, PL. and S.R. OVERMANN. 1974. Develop-
ment of dust-bathing in bobwhite quail: I. Effects of
age, experience, texture of dust, and social facilita-
tion. Dev. Psychobiol. 7(4):305-313.

BUCKMAN, HO. and N.C. BRADY. 1969. The nature and
properties of soils. The Macmillian Company. New
York, NY. 567pp.

CRAWFORD, W.T. 1950. Some specific relationships be-
tween soils and wildlife. J. Wildl. Manage. 14(2):1 15-
123.

. 1953- Relationships of soils and wildlife. Pages 10-

18 in Nagel, W.O. ed. Wildlife and the Soil. Missouri
Conservation Commission, Jefferson City, MO.

DALE, F.H. 1939- Variability and environmental responses
of the kangaroo rat, Dipodomys heermanni saxatilis.
Amer. Midi. Nat. 22:703-731.

DAUBENMIRE, R. 1968. Plant communities: A textbook of
plant synecology. Harper and Row. New York, NY.
300pp.

. 1972. Annual cycles of soil moisture and tempera-
ture as related to grass development in the steppe
of eastern Washington. Ecol. 53(3):4l9-424.

DAVIS, W.B. 1938. Relation of size of pocket gophers to
soil and altitude. J. Mammal 19:338-342.

, R.R. RAMSEY, and J.M. ARENDALE. 1938. Distribu-
tion of pocket gophers (Geotnys breviceps) in rela-
tion to soils. J. Mammal. 19:412-418.

DENNY, AH. 1944. Wildlife relationships to soil types.
Trans. N. Am. Wildl. Conf. 9:316-322.

DICE, L.R. and P.M. BLOSSOM. 1937. Studies of mammal-
ian ecology in southwestern North America with
special attention to the colors of desert mammals.
Carnegie Inst. Wash. Publ. No. 485:1-129.

DYKSTERHUIS, E.J. 1983. Habitat-type: A review. Range-
lands 5(6):270-271.

FELDHAMER, GA. 1979. Vegetative and edaphic factors
affecting the abundance and distribution of small
mammals in southeast Oregon. Great Basin Nat.
39(3):207-218.

FISHER, S. and P. DEUTSCH. 1983. The soil resource: Its
importance in the West and its role in coal develop-
ment and reclamation in Coal Development: Col-
lected Papers. Papers presented at Coal Development
Workshops in Grand Junction, Colorado, and Casper,
Wyoming. Sponsored by U.S. Dep. Inter., Bur. Land
Manage. Vol. 11:845-984.

GESSEL, S.P. and G.H. ORIANS. 1967. Rodent damage to
fertilized silver fir in western Washington. Ecol.
48(4):694-697.

HARDY, R. 1945. The influence of types of soil upon the
local distribution of small mammals in southwestern
Utah. Ecol. Monog. 15( 1 ):7 1-108.

HARPER, J.A. and R.F. LABISKY. 1964. The influence of
calcium on the distribution of pheasants in Illinois. J.
Wildl. Manage. 28:722-731.

HILL, E.P., HI. 1972. Litter size in Alabama cottontails as
influenced by soil fertility. J. Wildl. Manage. 36:1199-
1209.

HODGSON, J.R. 1972. Local distribution of Microtus mon-
tanus and M. pennsylvanicus in southwestern Mon-
tana. J. Mammal. 53(3):487-499.

KANTRUD, HA. and R.L. KOLOGISKI. 1982. Effects of
soils and grazing on breeding birds of uncultivated
upland grasslands of the Northern Great Plains. U.S.
Dep. Inter., Fish and Wildl. Serv. Wildl. Res. Rep. 15.
Washington, DC. 33pp.

KLEMMEDSON, JO. 1970. Needs for soil information in
the management of range resources. J. Range Manage.
23(2): 139- 143.

KREBS, C.J., B.L. KELLER, and J.H. MYERS. 1971. Microtus
population densities and soil nutrients in southern
Indiana grasslands. Ecol. 52(4):660-663.

KRUCKEBERG, A.R. 1969. Soil diversity and the distribu-
tion of plants with examples from western North
America. Madrono 20:129-154.

LEOPOLD, AS. and P.D. DALKE. 1943. The 1942 status of
wild turkeys in Missouri. J. Forest. 4l(6):428-435.

LUNT, OR., J. LETEY, and SB. CLARK. 1973- Oxygen
requirements for root growth in three species of
desert shrubs. Ecol. 54(6):1356-1362.

LYNCH, J.D. 1978. The distribution of leopard frogs {Rana
blairi and R. pipiens) (Amphibia, Anura, Ranadae) in
Nebraska. J. Herpetol. 12(2): 157- 162.

MARCHAND, D.E. 1973. Edaphic control of plant distribu-
tion in the White Mountains, eastern California. Ecol.
54(2):233-250.

MASER, C, J.M. GEIST, DM. CONCANNON, R. ANDER-
SON, and B. LOVELL. 1979. Wildlife habitat in man-
aged rangelands — the Great Basin of southeastern
Oregon: geomorphic and edaphic habitats. U.S. Dep.
Agric, For. Serv. Gen. Tech. Rep. PNW-99. Portland,
OR. 84pp.

MCARTHUR, ED., A.P. PLUMMER, and J.N. DAVIS. 1978.
Rehabilitation of game range in the salt desert. Pages
23-50 in Wyoming Shrublands, Proc. Seventh Wyo-
ming Shrub Ecology Workshop, Rock Springs, WY.
Range Manage. Div., Univ. WY. Laramie.

MCCOLLEY, P.D. and H.S. HODGKINSON. 1970. Effect of
soil depth on plant production. J. Range Manage.
23(3): 189- 192.

MOSS, A.E. 1940. The woodchuck as a soil expert. J. Wildl.
Manage. 4:441-443

MUNN, L.C., GA. NIELSEN, and W.F. MUEGGLER. 1978.
Relationships of soils to mountain and foothill range
habitat types and production in western Montana. Soil
Sci. Soc. Amer. J. 42:135-139.

MUNSELL COLOR. 1975. Munsell Soil Color Charts. Mac-
beth Division, Kollmorgen Corp. Baltimore, MD.
21218.

MUSHINSKY, H.R. and E.D. BRODIE, Jr. 1975. Selection of
substrate pH by salamanders. Amer. Midi. Nat.
93(2):440-443.

NAGY, J.G 1966. Volatile oils and antibiosis of Artemisia.
PhD. dissertation. Colorado State Univ. Ft. Collins.
73pp.

, H.W. STEINHOFF, and G.W. WARD. 1964. Effects

of essential oils of sagebrush on deer rumen microbial
function. J. Wildl. Manage. 28:785-790.

584

Soils

O'FARRELL, T.P. 1972. Ecological distribution of sage-
brush voles, Lagurus curtatus, in south-central Wash-
ington. J. Mammal. 53(3):632-636.

PASSEY, H.B., V.K. HUGIE, E.W. WILLIAMS, and D.E. BALL.
1982. Relationships between soil, plant community,
and climate on rangelands of the Intermountain West.
U.S. Dep. Agric, Soil Conserv. Serv. Tech. Bull. 1669-
123pp.

POWELL, J. 1970. Site factor relationships with volatile
oils in big sagebrush. J. Range Manage. 23( 1 ):42-46.

PROCTOR, J. and K. WHITTEN. 1971. A population of the
valley pocket gopher (Thomomys bottae) on a ser-
pentine soil. Amer. Midi. Nat. 85( 2 ):5 17-521.

ROSENZWEIG, M.L. and J. Winakur. 1969. Population
ecology of desert rodent communities: Habitats and
environmental complexity. Ecol. 50(4):558-572.

SHEETS, R.G., R.L. LINDER, and R.B. DAHLGREN. 1971.
Burrow systems of prairie dogs in South Dakota. J.
Mammal. 52(2):451-453.

SHELDON, W.G. 1950. Denning habits and home range of
red foxes in New York State. J. Wildl. Manage.
14:33-42.

SHORT, H.L. 1979. Deer in Arizona and New Mexico:

Their ecology and a theory explaining recent popula-
tion decreases. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. RM-70. 25pp.

SOCIETY FOR RANGE MANAGEMENT. 1983. Guidelines
and terminology for range inventories and monitor-
ing. Report of the Range Inventory Standardization
Committee (RISC). 13pp.

SOIL SURVEY STAFF. 1975. Soil taxonomy: A basic sys-
tem of soil classification for making and interpreting
soil surveys. U.S. Dep. Agric, Soil Conserv. Serv.,
Agric. Handb. 436. 754pp.

SPIGHT, T. M. 1967. The water economy of salamanders:
Exchange of water with the soil. Biol. Bull.
132(1): 126- 132.

STEBLER, A.M. 1939. An ecological study of the mammals
of the Badlands and the Black Hills of South Dakota
and Wyoming. Ecol. 20:382-393.

STOCKSTAD, D.S., M.S. MORRIS, and EC. LORY. 1953.
Chemical characteristics of natural licks used by big
game animals in western Montana. Trans. N. Am.
Wildl. Conf. 18:247-257.

STUART, L.C. 1932. The lizards of the middle Pahvant
Valley, Utah: Materials for a study in saurian distribu-
tion. Occ Papers Mus. Zool., Univ. Mich. 244:1-33-

TAUB, F.B. 1961. The distribution of the red-backed sala-
mander, Plethodon C. cinereus, within the soil. Ecol.
42(4):681-698.

THORNBURG, AA. 1982. Plant materials for use on sur-

face-mined lands in arid and semiarid regions. U.S.
Dep. Agric, Soil Conserv. Serv. SCS-TP-157. 88pp.

TURNER, L.W. 1972. Habitat differences between Spermo-
philus beldingi and Spermophilus columbianus in
Oregon. J. Mammal. 53(4):9l4-917.

U.S. BUREAU OF LAND MANAGEMENT. 1982. Inte
grated Habitat Inventory and Classification System.
U.S. Dep. Inter., Bur. Land Manage. Manual 6602.
Washington, DC.

U.S. SALINITY LABORATORY STAFF. 1954. Diagnosis
and improvement of saline and alkali soils. U.S. Dep.
Agric, Agric. Handb. 60. Washington, DC. 160pp.

U.S. SOIL CONSERVATION SERVICE. 1972. Soil Survey
Laboratory Methods and Procedures for Collecting
Soil Samples. Soil Survey Investigation Report (SSIR)
No. 1. U.S. Dep. Agric, Soil Conserv. Serv. Washing-
ton, DC. 20013-

Additional reports, containing soil survey labora-
tory data for specific states include:
SSIR No. 7 Montana
SSIR No. 8 Wyoming
SSIR No. 10 Colorado
SSIR No. 1 1 Oklahoma
SSIR No. 22 Alabama and Florida
SSIR No. 23 Nevada
SSIR No. 24 California
SSIR No. 28 Arizona

. 1976. National range handbook. NRH-1. U.S. Dep.

Agric, Soil Conserv. Serv. Washington, DC. 20013-

. 1981. Soil survey manual SSM-430-V. U.S. Dep.

Agric, Soil Conserv. Serv. Washington, DC. 20013

. 1982. List of published soil surveys: January 1982.

U.S. Dep. Agric, Soil Conserv. Serv. Washington, DC.
20013.

. 1983. National soils handbook. NSH-430-VI. U.S.

Dep. Agric, Soil Conserv. Serv. Washington, DC.
20013.

WEAVER, T 1977. Root distribution and soil water re-
gimes in nine habitat types of the northern Rocky
Mountains. Marshall, J. ed. The Below Ground Ecosys-
tem. Range Sci. Dep. Sci. Series 26. Colorado State
Univ., Fort Collins. 351pp.

WERTZ, W.A. 1966. Interpretation of soil surveys for wild-
life management. Amer. Midi. Nat. 75(1 ):2 11-224.

WHITAKER, G.A. 1965. Deer-soil relationships in the oak-
grasslands. M.S. Thesis, Humboldt State College. Ar-
eata, CA. 87pp.

WILLIAMS, C.E. and A.L. CASKEY. 1965. Soil fertility and
cottontail fecundity in southeastern Missouri. Amer.
Mid. Nat. 74( 1 ):21 1-224.

Soils

585

27

TERRESTRIAL

PHYSICAL

FEATURES

Allen Cooperrider

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

"With a broad awareness and understanding of par-
ticular physiographic features, one begins to appreci-
ate them as wildlife habitat, their relationship to
land mangement, and the need to account for them
in land-use planning."

— Chris Maser et al., Geomorphic and Edaphic
Habitats

Editor's Note: The wildlife biologist instinctively
recognizes the importance of physical features for
wildlife — the nesting cliff for falcons, the talus
slope for marmots, the south-facing snow-free win-
ter range for mule deer. However, habitat invento-
ries and evaluations often concentrate exclusively
on vegetation and ignore these important habitat
components and attributes. Formal systems for
classifying and naming such features as wildlife
habitat are virtually non-existent for terrestrial
features. This chapter covers the classification and
measurement of terrestrial features.

INTRODUCTION

Physical features of the environment are just as
important as vegetation in determining animal abun-
dance and distribution. Many wildlife species have
adapted to using certain physical features of the en-
vironment to the point where they cannot survive
and reproduce without them. Consider such obligate
relationships as use of cliffs for breeding by prairie
falcons and many other raptors, or use of steep, rug-
ged topography for escape cover by bighorn sheep
(Ovis canadensis).

Biologists and naturalists have long recognized
these dependencies and use of physical features by
individual species. However, little work has been
done to develop general methodology for categoriz-
ing or quantifying the features themselves. Several
reasons account for this neglect. First, some of the
most important attributes of physical features are
spatial or geographic. For example, the most impor-
tant attribute about a feature, such as a seep in a
limestone outcropping, is its location. If biologists
have marked the location on a map, they have re-
corded the data needed to predict, for instance, that
a species of salamander may be present. Measure-
ments of height, slope, etc., may be quite irrelevant.
Even with features such as cliffs, where attributes
such as height are important, broad categories, such
as "10 to 30 m" (33 to 99 ft) and "over 30 m" (99
ft), are adequate for most purposes.

Second, most biologists only work with a lim-
ited number of species and they know from experi-
ence what a suitable physical feature looks like. For
example, bighorn sheep biologists know that "steep,
rugged, rocky terrain" is an essential component of
sheep habitat. However, since they easily recognize
such habitat, few bother to attempt to quantify
"steep," "rugged," or "rocky." This lack of quantifica-
tion of terrestrial physical features sharply contrasts
with aquatic features, for which well-developed
measurement systems have been in use for years.

However, the need for more quantification or at
least more standardized categorization has become
evident. Biologists must deal with many more spe-

Terrestrial Physical Features

587

cies than in the past, and it is useful to have quanti-
tative measures or standard categories of physical
features that they can use with unfamiliar species.
Furthermore, biologists must document their work
more thoroughly. The subjective judgment that "it
looks like bighorn habitat to me" is being replaced
or corroborated by simple measurements that can be
compared to a standard description to confirm such
judgment.

The limited literature on terrestrial physical
features is confusing due to slightly different termi-
nology and varying approaches in determining physi-
cal features from biological ones. Is a snag, for
example, physical or biological? A concrete dam is
certainly a man-made physical feature, but what is a
beaver dam? Similarly, physical structures, such as
playas, are sometimes considered landforms (Peter-
son 1981) and other times considered geomorphic
features (Maser et al. 1979a).

Much of this confusion is caused by biologists
borrowing terms and definitions from other fields
such as soil science and geology. These definitions
have been developed for quite different purposes.
For the purposes of wildlife biology, how a feature is
classified (e.g., landform or geomorphic feature) is
probably not too important, as long as biologists
understand what the terms mean and know how to
use such features to predict species occurrence. If
biologists understand what a playa is and what ani-
mal species depend on such a feature, it does not
matter if they call it a landform because of its large
size or a geomorphic feature because it is a small,
distinct, localized habitat.

For the purposes of this chapter, terrestrial
physical features are all features of the terrestrial
environment, including man-made, that are not com-
posed of living vegetation, but are useful in predict-
ing animal abundance and distribution. Therefore,
this chapter identifies those features that are com-
monly termed landforms (Peterson 1981 ), geo-
morphic and edaphic habitats (Maser et al. 1979a),
special habitat features (Hamilton et al. 1983), man-
made habitats (Maser et al. 1979b), and features
derived from vegetation as snags and litter. Soils also
fall into this category. However, because of the im-
portance of soils in determining both plant and ani-
mal distribution, an entire chapter in this publication
has been devoted to this relationship (Chapter 26).
Soils are only briefly discussed in this chapter. This
chapter emphasizes "wildland" environments, those
habitats that are not used for urban, suburban, or
agricultural purposes, but which may be managed for
forestry, livestock production, or other purposes.

The dearth of literature on the subject has been
mentioned. Several publications are available, how-
ever, that must be considered as pioneering efforts
in the field, and they will be cited and quoted fre-

quently in the following discussions. The biologist
who is seriously involved in measuring terrestrial
physical features should obtain copies of these
publications.

The paper by Maser et al. (1979a) on "Geo-
morphic and Edaphic Habitats" is one of the first at-
tempts to list, categorize, and quantify these habitats.
Although it focuses on the Great Basin of southeast-
ern Oregon, much of the information and suggested
approaches can apply to different regions. Similarly,
the paper by Maser et al. (1979b) on "Man-Made
Habitats" published in the same series may have
widespread application. Biologists have been aware
for years of the benefits of man-made habitat features
such as nest boxes and gallinaceous guzzlers that
were developed specifically for wildlife. In the latter
paper, the importance to many species of wildlife
of certain man-made features, such as mine shafts,
abandoned buildings, and telephone poles, that were
developed for quite different reasons, is also
described.

A paper by Short and Burnham ( 1982) repre-
sents one of the first attempts to systematically de-
fine structural components of habitat that are used
for feeding and breeding. They use both physical
strata and vegetation structure for this purpose. The
former provides one of the first systematic efforts
to provide a comprehensive listing of physical fea-
tures used by vertebrate species.

Finally, Hays et al. ( 1981 ) have provided a com-
prehensive and systematic listing of measurement
techniques for habitat variables. Many procedures,
such as estimating height, can be applied to a variety
of habitat features, such as cliffs, trees, power poles,
etc. I refer to this publication frequently for guid-
ance on making these measurements or estimates,
and for information on considerations such as preci-
sion, accuracy, cost, efficiency, equipment, and train-
ing associated with different sampling techniques.

FEATURES AND THEIR MEASUREMENT

Physical features are discussed here under seven
general categories: ( 1 ) landforms, (2) geomorphic
habitats, (3) soils, (4) edaphic habitats, (5) vegeta-
tion-derived features, (6) animal-made features, and
(7) man-made features. These categories represent a
convenient grouping based primarily on two factors:
(1 ) derivation of the feature and (2) size of the fea-
ture. Some features such as a beaver dam, can fall
easily into two categories: vegetation derived and an-
imal made. It may be classified as either one or both
depending on the purpose of the investigation.
Where authors have categorized or defined features
slightly differently, I point out such differences.

588

Terrestrial Physical Features

Landforms

A landform is "a three-dimensional part of the
land surface, formed of soil, sediment, or rock, that is
distinctive because of its shape, that is significant for
land use or to landscape genesis, that repeats in var-
ious landscapes, and that also has fairly consistent
position relative to surrounding landforms" (Peter-
son 1 98 1 ). Alluvial fans and rock pediment are ex-
amples of landforms. Landforms are the basic
building blocks of the landscape and, thus, have a
dominating influence on the soils, vegetation, and an-
imal life that develop on their surfaces. For this rea-
son, landform is quite useful in predicting animal
occurrence.

The following two aspects of landforms are usu-
ally considered:

Classification. Classification of landforms is the
process of delineating an area on a map as a
landform type. A landform is thus always something
with area; this contrasts with most other physical
features used, which tend to be small and localized
and typically treated as points on maps. Landform
can usually be classified by simple visual obser-
vation or by interpretation of appropriate aerial
photography by a biologist who is familiar with the
concept and the local landforms. Unfortunately, few
if any landform classifications have been developed
for use by wildlife biologists. In fact, landform
classifications for any purpose are not available for
most areas. Therefore, in many cases, classifications
will have to be developed.

Peterson (1981 ) developed a landform classifica-
tion for the Basin and Range Province for purposes
of soil surveys (Table 1 ), and this has been used
by many range conservationists, botanists, and wild-
life biologists for their own purposes. This is a hier-
archical classification in which major landforms are
broken down into component landforms, landform
elements, and slope components. A biologist not
wanting to use all levels of the classification may find
the lower levels of classification ( landform elements,
slope components) useful in predicting animal oc-
currence. However, Peterson's classification should
be useful in developing classifications for other areas.

The U.S. Department of Interior, Bureau of Land
Management's (BLM's) Integrated Habitat Inventory
and Classification System (IHICS) (1982) lists land-
forms used by biologists; these represent ad hoc
definitions, and new categories are being added con-
tinuously. However, most of the initial definitions
came from Peterson ( 1981 ).

Attributes. Biologists frequently find that attributes
of landforms, such as slope and aspect, are useful
in predicting animal occurrence. These types of
measurements can be made anywhere and treated as
point measurements (i.e., a characteristic of an exact
location). Since measurements such as slope are
usually averaged and used to categorize a landform
or some other area of land, they are treated here
as landform attributes. The three most common such
measurements are slope, aspect, and elevation. Hays
et al. ( 1981 ) describe the first two measurements.

1

LU
00

E
<

o

h-
IX

LU

> <

^4 5°

HORIZONTAL RUN

SLOPE OF 45°OR 100%

SLOPE

SOUTH
ASPECT

ASPECT

ELEVATION

Terrestrial Physical Features

589

Table 1. Landform Classification System for the Basin and Range Province (from Peterson 1981).

Landtorms

Parts of Landforms

1

II

III

IV

V

Major

Component

Physiographic Part

Major Landform

Landform

Landform Element

Slope Component

Bounding mountains

(not defined)

Piedmont slope

Mountain-valley

Erosional fan

Summit

fan

remnant

QiH£iQlr*r~i£>

Shoulder

OlUt-olU[Jt;

Backslope

Footslope

Partial ballena

Crest
Shoulder
Backslope
Footslope

Inset fan

Channel
Channel

Rock pediment

Rock-pediment

Qi i mmit or _

Crest

OUilllllll Ul

remnant

sideslope

Channel

Shoulder
Backslope

Footslope

Ballena

Crest
Shoulder

Backslope

Footslope

Channel

Inset fan

Channel

Alluvial fan

Fan collar

Channel

Erosional fan

Summit

remnant

QiHoclnno

Shoulder

OlUcblupt;

Backslope

Footslope

Partial ballena

Crest
Shoulder
Backslope
Footslope

Channel

Inset fan

Channel

Fan pediment

Erosional fan

Summit

remnant

QiHoclnno -

Shoulder

OlUcolUJJt;

Backslope

Footslope

Partial ballena

Crest

Shoulder

Backslope

Channel

Inset fan

Channel

Fan apron

Channel

Nonburied fan

Channel

remnant

Beach terrace

Pediment

Pediment remnant

Summit

QiHoclnno -

Shoulder

OIUfc.olU|JC

Backslope

Footslope

Channel

Fan okirt

Channel

I c*l I or\l I I

Beach terrace

590

Terrestrial Physical Features

Table 1. Landform Classification System for the Basin and Range Province (concluded).

Landforms

Parts of Landforms

1

Major

Physiographic Part

II
Major Landform

III

Component

Landform

IV
Landform Element

V
Slope Component

Basin floor (Bolsom
or semi-Bolsom
floor)

Alluvial flat

Relict alluvial flat
Recent alluvial flat

Channel
Channel

Alluvial plain

Sand sheet

Sand dune
(Parna dune)

Interdune flat

Beach plain
(Bolson floor only)

Offshore bar
Barrier bar
Lagoon

Channel

Lake plain
(Bolson floor only)

Lake-plain terrace

Channel

Playa (Bolson floor
only)

Floodplain playa

Channel

Axial-stream flood-
plain

Floodplain playa
Stream terrace

Channel
Summit

Shoulder

Backslope

Footslope

Slope is the angle between the horizontal and
the plane of the ground surface. It is usually ex-
pressed in either angle degrees or percent. In the
latter case, percent is the vertical rise per horizontal
run. Thus, a 45-degree angle would be 100% so
that slopes of over 100% can occur. Slope can be
measured using a clinometer (Hays et al. 1981:87).
Usually an area will vary quite a bit in slope due
to microtopography. Since a biologist is usually inter-
ested in the average slope of the entire area, many
slope measurements are usually taken and then aver-
aged. Certain types of "badland" areas may have
very steep slopes even though there is little net
slope in any direction from one boundary of the area
to another. In such cases, average slope would be
close to zero and thus meaningless. However, the
variance of slope measurements may be a useful
measure of the ruggedness of the topography.

Aspect is the downhill direction of a slope or
the direction a slope faces. It is usually measured in
conjunction with slope by using a compass (Hays
et al. 1981 :87 ). As with slope, average aspect for an
area, rather than a point measurement, is usually
determined. On flat or relatively flat areas of little
measurable slope, aspect is of negligible interest and
is often not measured.

Elevation is the height above sea level. Although
elevation can be measured with an altimeter, since
most areas in the lower 48 states have already been
surveyed, the elevation can be read from a topo-
graphic map.

Slope, aspect, and elevation are extremely im-
portant determinants of animal distribution. For ex-
ample, in northern areas, south-facing slopes are
substantially warmer in winter and contain less
snow; therefore, animals congregate in these areas.
Similarly, the distribution of many animal species
is limited by elevation. However, if one has topo-
graphic maps, measurement of these attributes is not
necessary since they can be read (elevation) or
measured (slope and aspect) from such maps (Hays
et al. 1981:85). Furthermore, if information is being
stored in a geographic information system (GIS),
then digitized topographic information for many
areas is available on tapes. This information can be
purchased from the National Cartographic Informa-
tion Center (NCIC) of the U.S. Geological Survey.
Most GIS software has commands for making such
calculations, so the biologist does not need to do the
process manually.

Soils

Soils provide a substrate for vegetation to grow
and are a major determinant of the type and abun-
dance of vegetation that can grow on a site. How-
ever, soil also provides habitat for many species of
fossorial animals such as moles, ground squirrels,
pocket gophers, badgers, and many species of am-
phibians and reptiles. Soils provide habitat for —

• feeding (e.g., American woodcock [Scolopax
minor], starnose mole \Condylura aistata].
California legless lizard [Aniella pulchra));

Terrestrial Physical Features

591

• breeding or nesting (e.g., desert tortoise [Go-
pherus agassizii], white-tailed prairie dog [Cy-
nomys leucurus], burrowing owl [Athene
cunicularia]);

• cover (e.g., California ground squirrel [Spermo-
philus beecheyi], Coachella Valley fringe-toed
lizard [Utna notata}).

The type, texture, and other properties of the
soil will all determine the suitability of an area for
animals such as those mentioned above. For more in-
formation see Soils, Chapter 26 in this book; for fur-
ther reading, see Bailey (1984:68-78) and Robinson
and Bolen (1984.125-146).

contrast markedly with the surrounding area." They
described one edaphic habitat that supports a desert
shrub community of saltbush (Atriplex sp.), sage-
brush {Artemisia sp.), and rabbitbrush (Chrysotham-
nus sp. ). Although they recognize that edaphic
habitats have universal occurrence, they stated that
it is of special interest in the northern Great Basin
because it is the primary habitat of the white-tailed
antelope squirrel (Ammospermophilus leucurus)
and the only habitat of several native endemic plants.

Edaphic habitats form the basis for many wildlife
habitat classifications. As such, no particular measure-
ments are required but rather edaphic habitats are
identified as either areas or points on maps and class-
ified accordingly.

Edaphic Habitats

An edaphic habitat is one in which existing or
potential vegetation is determined primarily by soils
rather than climate. To a certain extent, this is true
of all habitats. Thus, Odum (1959) states that for
a given region it is convenient to recognize ( 1 ) a
single climatic climax, which is in equilibrium with
the general climate, and ( 2 ) a varying number of
edaphic climaxes, which are modified by the local
conditions of the substrate. Theoretically, given
enough time, the edaphic climaxes would converge
upon the climatic climax. In practice, however, biol-
ogists work with edaphic climaxes in terms of identi-
fying potential natural vegetation. Thus the concept
of edaphic habitat is used commonly in vegetation
and wildlife habitat classifications.

Edaphic habitats have been used in a more re-
strictive sense, however, by Maser et al. (1979b).
They considered edaphic features (1979a:2) as "local
distinctive soils that, along with their vegetation,

Geomorphic Habitat Features

Geomorphic habitat features are products of
geologic or geomorphic processes and include cliffs,
caves, talus, lava flows, sand dunes, and playas. There
is some overlap between what some consider a land-
form and a geomorphic feature. For example, a playa
is considered a geomorphic feature by Maser et al.
( 1979a) and a landform by Peterson ( 1981 ). In gen-
eral, however, geomorphic features represent parts
of the larger landform and do not necessarily occur
in a predictable position relative to surrounding
landforms or geomorphic features. For example, low
cliffs may be scattered throughout a larger landform
such as a fan piedmont. They do not necessarily bear
any consistent relationship to each other or to other
landforms. Similarly, geomorphic features such as
caves are often relatively small and localized and
treated as points on maps rather than features with
areas. Landforms, on the other hand, always have
area.

m v'ili;:
mmSKmm

Sand dunes are examples of geomorphic habitats.

Playas are considered either a geomorphic feature or a
landform.

592

Terrestrial Physical Features

Maser et al. (1979a) described six geomorphic
features of particular use in predicting the occur-
rence of vertebrate species in the Great Basin of
eastern Oregon: cliffs, caves, talus, lava flows, sand
dunes, and playas. Other geomorphic features are
used in the BLM IHICS efforts. In both systems, most
of these features are treated as point data, that is,
the only information recorded during an inventory is
the location. The area of a rock shelter, for example,
is considered to be negligible. Others are treated as
point data or an area depending on size. For exam-
ple, a cliff that is very localized and occupies less
than a hectare (2.5 a.), may be conveniently treated
as a point on a map, whereas in other situations,
where a series of cliffs are found for many kilometers
along a river system, they are more usefully re-
corded on the map as polygons with area.

Most descriptions of geomorphic features for
purposes of wildlife habitat surveys are limited to
mere notation of occurrence. However, systems of
classification or measurement have been developed
for several of these features. These systems are de-
scribed below.

Cliffs. Maser et al. (1979a) described features of
igneous rock cliffs such as facial cracks or joints,
ledges and occasional caves, and the shelflike
features and holes or pockets in sedimentary rock
cliffs as being very important for wildlife species in
the Great Basin. Cliffs are equally important in other
regions.

In addition to noting presence and possibly area,
cliffs can be classified according to parent rock and
resistance (Table 2) and according to structural
characteristics of height class, length class, joint
spacing, and joint width (Table 3). Any of these vari-
ables such as height can be measured according to
methods described in Hays et al. (1981).

Table 2. General resistance of cliffs to weathering
in Great Basin Region of southwestern Oregon
(from Maser et al. 1979a).

Parent Rock

Group

1

II

Basalt

•

Rhyolite

•

Andesite

•

Nonwelded tuff

o

Welded tuff

•

Sedimentary rock

o

(weakly consoli-

dated)

more resistant

0= less resistant

Caves. Caves are extremely important for many
species of vertebrates because they provide shelter
from weather extremes, stable environments,
darkness, and seclusion. Generally caves are
recorded as points on maps. Like cliffs, they can be

Igneous rock cliff's are important to wildlife because they
have long-lived features such as cracks, ledges, and deep
caves.

Caves in igneous rock are more important to cave dwell-
ing animals than sedimentary rock cliffs because they are
usually deeper and more resistant to weathering.

Terrestrial Physical Features

593

classified according to the type of parent rock, type
of formation ( Figure 1 ), as well as by structural
characteristics such as diameter of opening and
horizontal depth (Table 4).

Talus. Talus is the accumulation of rocks on or at
the base of steep slopes, usually cliffs. Talus affords
protection for reproduction and hibernation and
includes a stable environment. Talus slopes are
frequently recorded as point locations, although
Maser et al. ( 1979a) indicate that length and width
of deposits as well as depth are important
determinants of animal use. A slope can be noted as
a point location, while measurements are recorded
as attributes. In addition, talus can be classified by
structural characteristics such as type and size
classes of rocks (Table 5 and Figure 2).

Lava Flows. Lava flows have been classified into two
general categories, "old" and "new," by Maser et al.
(1979a). New flows are distinguished from old flows
by the lack of soil and vegetation development. Lava
flows, particularly old flows, create areas of rock
rubble of a size similar to talus and would afford the
same advantages to various wildlife species as talus
areas.

Sand Dunes. Sand dunes form unique habitats for
many species of vertebrates. Because the habitat
is usually unique and dunes tend to be rather

localized and few in number, many threatened,
endangered, and sensitive species of vertebrates as
well as plants and invertebrates occur on sand
dunes. They are usually mapped out. In addition,
Maser et al. (1979a) distinguish between active and
stabilized dunes.

Playas. Playas are shallow desert basins into which
water drains after snow melt or rain storms. They do
not have natural drainageways and water is
accumulated seasonally. These areas that contain
water are important feeding and nesting habitat for
some water bird species. They also form important
resting areas for migrating waterfowl in the spring
(Maser et al. 1979a). Because of salts left after
evaporation, playas have distinctive vegetation
communities. Playas are usually mapped as
intermittent water areas on maps and labeled as
playas.

Vegetation-Derived Habitat Features

Vegetation-derived habitat features refer here to
any habitat feature that is composed of dead vegeta-
tion. For convenience, these features are frequently
measured or located during the course of vegetation
surveys. For example, density of snag trees can usu-
ally be measured during the course of a timber in-
ventory without much additional effort. Similarly,

Table 3. Classification of structural characteristics of cliffs that produce habitat attributes that
can be exploited by adapted species of wildlife (from Maser et al. 1979a).

Structural Characteristic

Size Class or Description

Height class

3 m (10 ft)
3-10 m (10-33 ft)
10-30 m (33-100 ft)
30 m (100 ft)

Length class

30 m (100 ft)
30-100 m (100-330 ft)
100-500 m (330-1 ,640 ft)
500-1 ,500 m (1 ,640-4,920 ft)
1,500 m (4,920 ft)

Joint spacing

0.3 m (1 ft)

(called netted jointing — fine network of small cracks)

0.3-1.5 m (1.5 ft)

1 .5-7.5 m (5-25 ft)

7.5-30 m (25-100 ft)

30 m (100 ft)

Joint width (opening size)

0.15 m (0.5 ft)
0.15-0.30 m (0.5 ft)
0.3-0.6 m (1-2 ft)
0.6-1.5 m (2.5 ft)
1 .5 m (5 ft)

594

Terrestrial Physical Features

CO

CD

a
>^

i-

co

CD

co

D

"O

c
CO

CO

c

CO
LL

L^2>^

t/if ^

fillip®

Shallow caves

Cliff-face caves

Lava tubes

Mine shafts

Mice
Wood rats

Bats

Bats

Bats

Shelter, reproduction,
nesting

Roosting, reproduction,
hibernation

Roosting, reproduction,
hibernation

Roosting, reproduction,
hibernation

Coyotes
Bobcats

Mice
Wood rats

Mice
Wood rats

Shelter, feeding

Shelter, reproduction,
nesting

Shelter, reproduction
nesting

Figure 1. Types of caves and some of the wildlife species that use them.

Table 4. Structural characteristics of caves that
produce habitat attributes which can be
exploited by adapted species of wildlife (from
Maser et al. 1979a).

Structural
Characteristic

Size Class or Description

Opening
(diameter)

0.3 m (1 ft)
0.3-1 m (1-3.3 ft)
1 m (3.3 ft)

Depth of cave
(horizontal)

1 m (3.3 ft)

1-3 m (3.3-10 ft)

3 m (10 ft)

Origin

Natural

Man-made
— abandoned mine shaft
— abandoned railroad
tunnel

Table 5. Talus class, based on the predominant or
most common rock size in the talus field (from
Maser et al. 1979a).

Talus Class

Rock Size (Diameter)

1

II
III
IV

0.5 m (< 1.6 ft)
0.5-1 m (1.6-3.3 ft)
1-2 m (3.3-6.5 ft)

2m(< 6.5 ft)

density or thickness of litter is often estimated dur-
ing range surveys. The features are described here
because they are sometimes omitted from vegetation
surveys; those conducting such surveys are primarily
concerned with live vegetation and often do not
understand the importance of dead material to
wildlife.

Dead vegetation as a habitat feature is extremely
important to many species. Furthermore, many vege-
tation management practices such as livestock graz-
ing, timber harvesting, fire management, and
firewood cutting and collecting are done in a man-
ner that tends to minimize the amount of dead vege-
tation left in the habitat. The quantity and quality of
vegetation-derived habitat features may thus change
drastically in a few years. By contrast, physical fea-
tures such as cliffs, caves, and other geomorphic
habitats are largely unaffected by human activities
and are primarily quantified during inventories but
rarely monitored. Measurement of dead vegetation
should not be overlooked or dismissed in inventory-
ing or monitoring habitat features.

There are literally thousands of habitat features
formed by dead vegetation that are relevant to indi-
vidual vertebrate species. Three major categories are
discussed here: ( 1 ) litter and mulch, (2) dead and
down woody material, and (3) snags.

Terrestrial Physical Features

595

1

ZONE OF

ZONE WITH

LITTLE PLANT LIFE

SOME PLANT LIFE

»|

~~,

* *

"1

> m
- <

< w
in z

,

_l

v ^S&SSrc-S^fc'- V>_

> ->

— CQ

< <

UJ ^

W^^jmUL '" *' JiNHw

CC

_

ZONE OF

ZONE OF

LEAST ANIMAL USE

GREATEST ANIMAL USE

Figure 2. Structure of a talus field and its use by
wildlife.

Litter and Mulch. Litter or mulch is the dead
material from herbaceous or woody plants, primarily
grasses, forbs, and shrubs. Range conservationists
make the distinction between persistent and non-
persistent litter. Persistent litter is defined as
"undecomposed organic debris on or near the soil
surface with expected decomposition rates
exceeding 2 years" (U.S. Department of the Interior,
Bureau of Land Management 1985).

Non-persistent litter is "undecomposed organic
debris on or near the soil surface with expected
decomposition rates of 2 years or less" (U.S. Depart-
ment of the Interior, Bureau of Land Management
1985). It is composed primarily of herbaceous mate-
rial.

Litter is commonly measured during range vege-
tation surveys through step-point transects or other
techniques. It is usually expressed as percent cover.
In certain areas, particularly the Mediterranean, an-
nual grasslands of California, and deciduous riparian
habitats, litter is so abundant and important that
thickness is important and is measured. The term
mulch is often used for this type of litter in annual
grasslands.

Dead and Down Woody Material. Dead and down
woody material refers to fallen trees in various stages
of decay. The importance of such material as wildlife
habitat in forests has received considerable attention
in recent years (Maser et al. 1979c; Maser and
Trappe 1984), because many management practices,
such as chipping and slash-burning, tend to reduce
the amount of such material.

Biologists have rarely measured amounts of dead
and down woody material. However, Maser et al.
(1979c) have suggested several approaches that
might be useful and practical. The number of dead
and down logs in an area can be sampled and ex-
pressed as number per unit area without any mea-
surements of characteristics of individual logs. In
addition, if a more refined approach is desired, they
have suggested classifying of logs in terms of their
stage of decomposition, which is relevant to their
value for wildlife species (Table 6).

Snags. A snag has been defined for wildlife purposes
as "any dead or partly dead tree at least 10.2 cm (4
in.) in diameter at breast height (DBH) and at least
1.8 m (6 ft) tall" (Thomas et al. 1979). This
definition is based on the minimum diameter and
height of trees for cavity nesting birds. Snags are
important for many species of wildlife not only for
nest or den sites but for many other biological
functions such as feeding and roosting (Davis et al.
1983; Thomas et al. 1979).

Snags in an area are generally counted or esti-
mated by using a standard sampling procedure and
expressed as a density (i.e., the number of snags per
hectare or acre). Thomas et al. (1979) have sug-
gested that snags may be usefully classified as hard
or soft, depending on the degree of decay and dete-
rioration. Hard and soft snags can be classified in the
field by striking them with an axe (Gale 1973). If
the axe sinks in with difficulty, the snag is hard,
whereas if it penetrates the wood easily, the snag is
soft. Hard snags usually have many dead branches
and an intact top, whereas soft snags usually have
broken tops and few limbs (Gale et al. 1973).

Snags can be further classified based on DBH.
The U.S. Fish and Wildlife Service uses many varia-
tions of snag measurements in their habitat suitabil-
ity index (HSI) models.

Animal-Made and Man-Made Features

Litter is an important habitat component for
many species of small mammals, reptiles, and am-
phibians. It provides habitat for many invertebrates
that are food sources for vertebrates, and it provides
thermal and hiding cover for many of these species.

In many cases, animals create distinctive fea-
tures in a habitat which are then used by the same
or other species. Many of these are small and local-
ized, such as the tree cavities created by woodpeck-
ers that may subsequently be used by birds lacking

596

Terrestrial Physical Features

Table 6. A classification of decomposed logs (from Maser et al. 1979c).

Decomposition
Class

Log Characteristic

Bark

Twigs

3 cm (1.18 in.)

Texture

Shape

Color of
Wood

Portion of Log
on Ground

1

Intact

Present

Intact

Round

Original
color

Log elevated
on support
points

2

Intact

Absent

Intact to
partly soft

Round

Original
color

Original log
elevated on
support points
but sagging
slightly

3

Trace

Absent

Hard, large
pieces

Round

Original
color to
faded

Log is sag-
ging near
ground

4

Absent

Absent

Small, soft,

blocky

pieces

Round
to oval

Light brown
to faded
brown or
yellowish

All of log on
ground

5

Absent

Absent

Soft and
powdery

Oval

Faded to
light yellow
or gray

All of log on
ground

Decaying logs furnish cover for many species of wildlife.

Snags are used by many species of cavity
using animals.

Terrestrial Physical Features

597

the ability to excavate a cavity. However, some ani-
mal-made features are large and important enough to
be noted during habitat surveys. Examples of these
types of features are beaver dams and elk wallows.
These are usually noted as present or absent without
any measurement or further categorization of the
feature.

The importance of man-made features for wild-
life has received increased recognition. Wildlife man-
agers have developed structures such as guzzlers
and nesting platforms for wildlife for many years and
their value to wildlife is understood. In recent years,
the value of structures built for other purposes has
received attention. These include features such as
bridges and culverts used for nesting by swallows
and other birds, mine shafts used by many bat spe-
cies, and rock walls used by many species of reptiles
and small mammals. However, rarely has a man-made
feature been solely responsible for the occurrence
of a species in an area.

On the other hand, many man-made features
such as roads, powerlines, and pipelines, have ob-
vious detrimental effects on wildlife. These detrimen-
tal features need to be recorded during wildlife
habitat inventories as they are also an important
component of the habitat. Maser et al. (1979b) de-
scribes in detail important man-made features for the
Great Basin of eastern Oregon. However, few fea-
tures in other regions have been described.

Both animal- and man-made features are gener-
ally not measured but simply recorded as being pres-
ent in habitat inventories. Since these types of
features change very slowly over time, monitoring
them is rarely a high priority.

SPECIAL HABITAT FEATURES AND THEIR
USE

The above description of terrestrial physical
features categorizes habitat components according to
their origin in a series, ranging from those caused
by geological processes to those created by humans.
Although useful in subdividing a large set of compo-
nents, in practice, the origin of a habitat feature is
probably of little concern to the animals that use
them. If the environmental conditions are otherwise
similar, an abandoned mine shaft is equally accept-
able as a natural cave to a roosting bat. Thus a biolo-
gist conducting an inventory of such features does
not need to be overly concerned with categorizing
origin.

The BLM's IHICS makes extensive use of "spe-
cial habitat features," including any of the above

categories (geomorphic, edaphic, soils, vegetation-
derived, animal-made, man-made). In most cases,
these are habitat features that are too small to be
mapped as units with area, i.e., that are mapped as
points on maps. In addition, these features include
small patches of vegetation that are too small to be
mapped as separate vegetation types, such as a small
stand of trees around a desert spring (Table 7).

Since the number of special habitat features is
almost unlimited, time and money do not permit
a complete inventory of them. Thus, the biologist
must choose those that are relevant to wildlife man-
agement. BLM biologists have found this approach
of recording terrestrial physical features to be quite
useful.

The special habitat features listed in Table 7 for
IHICS are those that were in use as of July 1985.
However, these are not fixed. Biologists may add
other features as required. Undoubtedly features
listed in Table 7 that are used in other systems or
described in other publications, will be added as
necessary. As with all habitat components, the biolo-
gist must determine which ones are most useful
based on the objectives of the inventory or monitor-
ing effort.

DISCUSSION

Wildlife use physical features in many different
ways with varying degrees of dependence. First,
some species in an area increase their abundance
through the presence or addition of a feature. For
example, many lizard populations will increase
through the addition of a wooden fence, but the
populations do not depend on fences for survival.
Most species respond to man-made features of this
type.

Second, some species that are mobile will use a
site if a necessary feature is there, but the feature
may not be limiting the population. For example,
many cliff-nesting raptors use a limited number of
appropriate cliff sites; presence of an appropriate site
may determine local use, but may not limit popula-
tion density or productivity in the region. If the site
is destroyed, the birds simply use alternative nest
sites with no detrimental effects on the population.
Unfortunately, many man-made structures, including
some that have been constructed deliberately for
wildlife, fall into this category. For example, many
gallinaceous guzzlers have been constructed for quail
and chukars in desert areas that were already fairly
well-watered. Although quail will concentrate around
such waters during dry periods and drink from them,
there is no evidence that they increase population
densities in such areas.

598

Terrestrial Physical Features

Table 7. Special habitat features used in the Bureau of Land Management's Integrated Habitat Inventory and
Classification System (IHICS).

Natural Features

Man-Made Features

Avalanche-slide area

Bridge

Cave

Fence

Cave ice

Underpass

Cave lava

Salting area

Cliff

Goose nesting platforms

Cone, volcanic

Artificial nesting boxes

Dike, volcanic

Small seedlings

Dune, sand

Buffer strip

Insect mounds

Building

Overhang

Bird ramp

Salting area

Berm

Seep

Culvert

Cold spring

Dock

Sinkhole

Dredged area

Snag or group of snags

Exclosure, study

Talus slope

Fish migration barrier

Talus field

Gauging station, water

Elk wallow

Mining activity

Waterfall

Poles, electrical/telephone

Wasteland

Perches

Island (too small to be typed as habitat site)

Road

Log jam

Trail

Downed timber

Stream improvement structure

Bluff

Railroad

Beaver dam

Stream crossing

Muskrat house

Shelter (overnight)

Cataracts (stream)

Recreation area

Barren lands

Feeding stations

Hot springs

Fire break

Blowouts

Seismographic trail

Mudflow

Oil sump pit

Temporary pond

Windmill

Small natural pond

Irrigation diversion or ditch

Small grp trees or shrubs

Water gap

Small grp trees or riparian

Stock water pond

Dry meadow (not typed as vegetation type)

Corral or loading chute

Dry wash

Artificial wildlife water

Stream bank gravel

Domestic water source

Raptor nest tree

Artesian well

Buffalo wallow

Oil well

Boulder or rock outcrop

Gas well

Rodent colony

Pipeline

Beaver lodge

Material site

Otter slide

Airfield

Perennial snowfield

Breakwater

Rocky crags

Dam

Alpine fell field

Wilderness camp

Pingo

Winter trail

Gravel bar

Burn

Sand bar

Mine shaft

Ocean cliff

Mine tunnel

Stack

Stock water tank

Glacier

Disposal site (active)

Spit

Disposal site (inactive)

Barren

Wrecked ships

Burn

Abandoned homesites

Booming, dancing, or strutting ground

Relay stations

Wet meadow (not typed as vegetation type)

Pump jack

Brushy openings (too small to be typed as habitat site)

Brush pile/rows

Terrestrial Physical Features

599

Table 7. Special habitat features used in the Bureau of Land Management's Integrated Habitat Inventory and
Classification System (IHICS) (concluded).

Natural Features

Man-Made Features

Snake den

Roost

Kipuka

Stream (too small to be mapped as habitat site)

Mineral spring

Rock formation raptor nests

Raptor cliff nest-stick

Raptor cliff nest-scrape (no nest material)

Raptor cliff perch

Raptor nest on pinnacle (rock or earthen)

Raptor nest (ground or hillside)

Raptor nest (shrub)

Raptor nest (ground burrow)

Rookery

Oxidation ponds (oxidation ponds and evaporation

playa)
Spring, wildlife (developed & useable by wildlife only)
Spring, livestock (developed & useable by livestock

only)
Spring, wildlife/livestock (developed & useable by

wildlife & livestock)
Exclosure (vegetation protection for wildlife)
Vegetation Manipulation (for wildlife; too small to be

typed as a habitat site
Artificial catchment
Reservoir

Gallinaceous guzzler
Rain gauge
Raptor nest (artificial)
Raptor nest (earth cut)

A third category involves species that are totally
dependent on a feature within an area. Examples of
this relationship are species of salamanders found
only at limestone seeps and the many desert fishes
found at isolated desert springs. Most of these spe-
cies are sedentary, such as amphibians and reptiles.
Many species are remnant or relict species from
earlier ages when conditions favorable to their exis-
tence were more widespread. Desert pupfish (Cypri-
nodon sp. ) are good examples of such a situation.

Biologists should consider these three types of
relationships when planning and setting priorities to
inventory or measure physical features. Priority
should normally be given to inventory features in
the third category and subsequently verify species
presence around them. The second category requires
that the biologist determine the extent to which the
features are limiting and at what level (site, regional,
statewide, etc.). Returning to the raptor example,
the biologist must determine if suitable nest sites are
a factor limiting the population at the state or re-
gional level. If it is not, then attention to inventory
and management of such sites is not warranted un-

less the management objective is to attract birds
to the local site. Special features of the first type are
not normally a high priority for inventory or man-
agement, unless they increase the abundance of high
priority species, such as threatened, endangered, or
game species.

SUMMARY

Physical features are important habitat compo-
nents for many wildlife species. These include land-
forms, geomorphic and edaphic features, soils,
vegetation-derived features, animal-made features,
and man-made features. Formal systems for invento-
rying and describing such features have not been
developed to the same degree as systems for survey-
ing vegetation. In many cases, simply noting or map-
ping their presence is adequate. For some special
features, measurements may be required; for others,
classifying them by size or other characteristics is
useful. The biologist must determine which physical
features to inventory, describe, or measure depend-
ing on the objectives of the project.

600

Terrestrial Physical Features

LITERATURE CITED

BAILEY, J. A. 1984. Principles of wildlife management. John
Wiley & Sons. New York, NY 373pp.

DAVIS, J.W., G.A. Goodwin, and R.A. Ockenfels, tech.

coords. 1983 Snag habitat management: Proceedings
at the symposium. U.S. Dep. Agric, For. Serv., Gen.
Tech. Rep. RM-99. 226pp.

GALE, R.M. 1973- Snags, chainsaws and wildlife: One as-
pect of habitat management. Page's 97-1 12 in
Yoakum, J. ed. Cal-Neva Wildlife 1973, Transactions.
Reno

, W.F. KELLEY, and J.A. LORENZANA. 1973. Snag

management: Coordination guidelines for wildlife
habitat. U.S. Dep. Agric, For. Serv., California Region.
13pp.

HAMILTON, C.KL, R.M. KERR, and LA. PETERSON. 1983.
IHICS, the Bureau of Land Management's habitat in-
ventory system. Presented at National Workshop
on Computer Uses in Fish and Wildlife Programs — a
State of the Art Review. Virginia Polytechnic Institute
and State University, Blacksburg, VA. 1 3pp

HAYS, R.L., C. SUMMERS, and W. SEITZ. 1981. Estimating
wildlife habitat variables. U.S. Dep. Inter., Fish and
Wildl. Serv., FWS/OBS-81/47. 111pp.

MASER, C, R.G. ANDERSON, K. CROMACK, Jr., J.T. WIL-
LIAMS, and RE. MARTIN. 1979c. Dead and down
woody material. Pages 78—95 in Thomas, J.W. ed.
Wildlife Habitats in Managed Forests — The Blue
Mountains of Oregon and Washington. U.S. Dep.
Agric, For. Serv., Agric. Handbook 553 512pp.

, J.M. GEIST, DM. CONCANNON, R. ANDERSON,

and B. LOVELL. 1979a. Wildlife habitats in managed
rangelands — The Great Basin of Southeastern Ore-

gon— Geomorphic and edaphic habitats. U.S. Dep.
Agric, For. Serv., Gen. Tech. Rep. PNW-99. 84pp.

— , J.W. THOMAS, ID. LUMAN, and R. ANDERSON.
1979b. Wildlife habitats in managed rangelands — The
Great Basin of southeastern Oregon — man made habi-
tats. U.S. Dep. Agric, For. Serv., Gen. Tech. Rep. PNW-
86. 84pp.

— and J.M. TRAPPE, eds. 1984. The seen and unseen

world of the fallen tree. U.S. Dep. Agric, For. Serv.,
Gen. Tech. Rep. PNW-164. 56pp.

ODUM, E.P. 1959. Fundamentals of ecology. 2nd Edition,
W.B. Saunders Co., Philadelphia, PA. 546pp.

PETERSON, F.F. 1981. Landforms of the Basin and Range
Province. University of Nevada, Tech. Bull. 28, Reno.
52pp.

ROBINSON, W.L and E.G. BOLEN. 1984. Wildlife ecology
and management. MacMillan Publishing Co., New
York, NY. 478pp.

SHORT, H.L. and K.P. BURNHAM. 1982. Techniques for
structuring wildlife guilds to evaluate impacts on
wildlife communities. U.S. Dep. Inter., Fish and Wildl.
Serv., Spec. Sci. Rep. — Wildlife 244. 34pp.

THOMAS, J.W., R.G. ANDERSON, C MASER, and E. BULL.
1979. Snags. Pages 60-77 in Thomas, J.W. ed. Wildlife
Habitats in Managed Forests — the Blue Mountains of
Oregon and Washington. U.S. Dep. Agric, For. Serv.,
Agric. Handbook 553- 512pp.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF

LAND MANAGEMENT. 1982. Integrated habitat inven-
tory and classification system. U.S. Dep. Inter., Bur.
Land Manage., Manual 6602, unnumbered.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF

LAND MANAGEMENT. 1985. Rangeland monitoring—
analysis, interpretation, and evaluation. U.S. Dep.
Inter., Bur. Land Manage., Tech. Reference
4400-7. 69pp.

Terrestrial Physical Features

601

28

AQUATIC

PHYSICAL

FEATURES

Paul Cuplin

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

Editor's Note: Although biologists have done little
to develop standardized methods or measurements
for terrestrial physical features, there has been more
work with aquatic systems. Fisheries biologists and
limnologists have long recognized the importance
of physical features of streams such as streamflow,
stream gradient, and water temperature. Whereas
the terrestrial biologist must, for the most part,
develop ad hoc systems for quantifying physical
features, the aquatic biologist has some reasonably
well -developed measurements and classification
systems for aquatic systems. This chapter describes
physical features that are important determinants
or predictors of the quality of the aquatic habitat
for fish and other vertebrates, and systems for meas-
uring or classifying such features.

INTRODUCTION

Aquatic biologists have collected many types of
data to describe aquatic habitat — sometimes with
uncertainty as to their value. Accuracy, precision,
and bias were not generally considered; usually the
ease of data collection was most important. More
attention is now being directed toward repeatable
results, recognizing fewer variables that can be meas-
ured with higher precision and accuracy. Examples
of these variables are water temperature, stream
width, stream depth, lake depth, surface acreage,
stream gradient, and turbidity. Examples of variables
that cannot be measured with high precision are
pool quality, stream channel stability, streambank sta-
bility, and the amount of various-sized particles in
the streambed.

HABITAT FEATURES AND MEASUREMENT
TECHNIQUES

The variables that are most important for de-
scribing stream and lake habitat are discussed in this
chapter, regardless of the precision with which they
can be measured.

Precision and accuracy of measurements of
aquatic physical features vary considerably. Platts
(1981) rates the precision that can be applied to
variables related to stream habitat (Table 1 ). Hamil-
ton and Bergersen (1984) discuss precision and accu-
racy of many of these variables.

Streamflow Pattern

Characterization of streamflow pattern as peren-
nial, intermittent, or ephemeral streamflow can be
determined by observing the stream and measuring
streamflow during high and low precipitation pe-
riods. It can also be determined from U.S. Geological

Aquatic Physical Features

603

Survey (USGS) topographic maps as different sym-
bols are used tor perennial, intermittent, and ephem-
eral stream reaches.

Streamilow

Streamflow is measured to determine stream
size for baseline information. Streamflow volume re-
lates to the amount of living space available for fish
and aquatic life.

Historic records of streamflow (mean, record
high and low, etc.) can be obtained from the U.S.
Geological Survey.

Streamflow can be measured by using the for-
mula R = WDCV (Robins and Crawford 1954):

where R = flow in cubic feet per second ( cfs )
W = average width of stream in feet
D = average depth of stream in feet
C = a constant for stream bottom
0.8 - rough bottom
0.9 - smooth bottom
V = velocity in feet/second

Two people are generally required to make
these measurements at each sampling station. Ap-
proximately half an hour is required to measure
stream depths, widths, and velocities which are
needed to calculate streamflow in cubic feet per
second. Measurement of each of these is described
separately. The minimum equipment required is a
watch with second hand, yardstick or meter stick,
float stick, and 100-foot measuring tape.

Installing a streamflow measuring gauge.

Table 1. The average expected repeatability, precision, and confidence intervals of water column and

streambank measurement means from six selected streams in Idaho and Nevada. Confidence intervals are
at the 95% level and expressed as percent of the mean (from Platts 1981).

Confidence

Repeatability

Measurement

Interval

Precision

over Time

Water Column

Stream width

5.4

©

©

Stream depth

8.2

©

©-•

Streamside water depth

16.6

©

©-©

Pool (percent)

10.3

©

O

Pool (quality)

8.0

©

o-©

Riffle (percent)

12.5

©

o

Sun arc angle

1.1

•

©

Bank to bank width

Very wide

o

o

High water stream width

Very wide

O

o

Streambank

Soil alteration

12.3

©

©-©

Vegetative stability

3.1

•

©-©

Undercut

18.5

©

©

Angle

4.4

•

©

Rock content

Very wide

o

o

Excellent

Good Q= Fair Q= poor

604

Aquatic Physical Features

Stream Velocity. Stream velocity can be measured
in numerous ways. The simplest method, and the
one commonly used in conjunction with streamflow
measurement, is the float method.

A straight section of stream, 100 feet long, is
measured. A 6-inch wooden stick or table tennis ball
is used as a float. Some people recommend using an
object with the specific gravity of water (1.0), such
as an orange or water-filled prescription bottle.

Streamflow velocity is more precisely measured
by using a flow meter to determine feet-per-second
flow. A minimum of 20 measurements are taken
across a stream transect and averaged.

For a more detailed discussion of velocity and
its measurement, see Hamilton and Bergersen
(1984).

Stream Width. Stream width is the average distance
between water's edges transversely across the
stream. On smaller streams, measurements can be
made with a tape or estimated. Several
measurements should be taken and the average
calculated for a given stream reach.

For larger streams, an optical range finder can
be used in the field and a map measurer can be use-
ful in making these measurements in the office. Hays
et al. (1981) describe use of these techniques.

Stream Depth. Water depth is the vertical distance
from the surface to the bottom. Water depth
measurements may be required for either streams or
lakes; however, the techniques used are often
different.

In addition to its use in streamflow calculations,
measuring the depth of a body of water provides a
check on siltation rate and a basis for classifying and
typing lakes and reservoirs.

For streams, water depth can be measured with
a calibrated rod. Several depth recordings are used
to calculate average stream depth.

Lakes, ponds, and reservoirs and larger streams
or rivers can be sounded with a weighted line or
recorded with depth-sounding equipment from a
boat. A sounding line, marked in 5-foot increments,
will provide adequate depth information for cross
sections and verification and correction of topo-
graphic maps for small lakes and reservoirs.

A depth recorder, with the transducer mounted
on the boat keel, provides an accurate timesaving
depth-measuring tool. The recorder should be cali-
brated in feet or meters and equipped with a record-
ing tape. A minimum of four depth-sounding cross

sections are located on the topographic map — a tran-
sect from the backwater to the outlet in the center
of the lake and three or four transects perpendicular
to the backwater-outlet transect. Irregularly shaped
reservoirs and lakes will require additional transect
soundings. The depth-sounding tapes keyed to the
topographic map can be processed by an engineer to
develop a new or current topographic map.

Hamilton and Bergersen (1984) provide a de-
tailed description of techniques for measuring water
depth.

Instream Flow Needs

The amount of water needed to maintain stream
habitat on a year-round basis is termed "instream
flow needs." Instream flow needs are assessed to
identify streamflow requirements for resident aquatic
life and identify the amount of habitat loss that
would occur at reduced rates of streamflow.

Instream flow needs can be determined by sev-
eral methods as summarized by Cuplin et al. (1979).
Where detailed, precise, legally defensible informa-
tion is required, the multiple transect — incremental
flow method — is recommended. This method is de-
scribed in detail in Bovee and Milhous (1978).

For rapid assessment, a single transect method is
available (Cuplin et al. 1979). This technique re-
quires less time for field work and data analysis but
it produces less precise results than the incremental
flow methods.

Computer programs have been developed for
analyzing stream channel cross-section survey data
collected in conjunction with instream flow assess-
ments (Parsons and Hudson 1985). A single transect
method can be used for rapid assessment using ADP
programs. A computer analysis program is available
for U.S. Bureau of Land Management offices through
the Denver Service Center, Division of Resources
(Parsons and Hudson 1985).

Streambank and Shoreline Stability

Streambank and shoreline stability refers to the
resistance of bank materials to erosion from flowing
water. Streambank and shoreline stability measure-
ments are used to classify stream/riparian habitat
quality and to monitor changes in stability caused by
the land use.

Streambank stability can be classified into four
condition classes from excellent to poor (Table 2).
This classification is designed for addressing stability
as a function of livestock grazing. Other land-use
impacts can also be assessed by ocular surveys. All

Aquatic Physical Features

605

Table 2. Condition classes for streambanks and shorelines (U.S. Department of the Interior, Bureau of Land
Management 1979).

Category

Rating

Description

1

•

No negligible use/damage; well-rooted vegetation,
primarily grasses, sedges, and forbs; sod intact;
very little if any erosion from vegetation areas; less
than 5% bare soil showing along shoreline.

II

9

Some use/damage; vegetation generally well-rooted;
sod mostly intact; soil showing in places, 6 to 1 5%
bare soil showing overall; some surface erosion
evident.

III

9

Use or damage close to sod; vegetation shallowly
rooted; moderate surface erosion, 16 to 25% bare
soil showing overall.

IV

O

Heavy to severe use/damage; vegetation generally
grazed down to the soil; considerable soil showing,
over 25% with serious sod damage; active surface
erosion is a serious problem.

I = Excellent Q = Good Q = Fair Q = Poor

estimates are subjective and thus the accuracy of this
method depends on the ability and experience of
the surveyor.

Stream Channel Stability

Stream channel stability is rated as part of the
overall stream habitat condition. Existing natural
conditions and land use govern stream channel
stability.

Stream channel stability is evaluated by ocular
classification. One method that can be used classifies
channel stability in four categories — excellent, good,
fair, and poor (Table 3). This method has limited
application to desert streams that are generally ero-
sional in nature.

Pfankuch ( 1975) and Duff and Cooper ( 1978)
describe a very detailed channel stability evaluation
procedure. This method requires visually categoriz-
ing numerous bank and bottom attributes individ-
ually and then assigning them a point value (Figure
1 ). An overall score is determined by summing
scores for individual attributes. This system works
best for timbered mountainous streams; I have had
poor results trying to apply it to arid land sedimen-
tary streams. For the latter, the simpler visual system
(Table 3) works better.

Stream Gradient

Stream gradient is a general slope or rate of
change in vertical elevation per unit of horizontal
distance of water surface of a flowing stream. Stream

gradient can be measured in the field using a clino-
meter or Abney level or calculated in the office from
measurements on topographic maps.

In the field, gradient is estimated by measuring
the angle from the horizontal on two selected points
on the ground with a clinometer or Abney level.
Use a rod with a mark at the same height as the
measuring crew member's eye level to measure an-
gle from the horizontal. One crew member holds the
marked rod at one of the sample points and the
other measures the angle of the target from the hori-
zontal using the clinometer. One person can com-
plete the required measurements, but it is more
efficient with two. Hays et al. (1981:87-90) describe
this technique in detail.

In the office, gradient can be calculated off USGS
topographic maps using the following formula:

Em (map scale fraction)
G ~ d

where G = gradient (in percent)

E = difference (in feet) in elevation on

ground between two points
m = the map scale, decimal fraction
d = distance (in inches) on the map

between the two points

Hays et al. (1981:85-87) describe this technique in
detail.

If digital elevation model tapes are available and
a stream course has been digitized, then gradient
can be calculated automatically for any stream reach
using a geographic information system (GIS).

606

Aquatic Physical Features

Table 3. Stream channel stability (from Cuplin 1978).

Category

Rating

Description

1

•

None or negligible lateral channel movement and
bank erosion (cutting) (5% or less), scour, or
changing channels.

II

Q

Some lateral channel movement and bank erosion
(5 to 10%); minor channel scour or changing
channel within streambed.

III

©

Frequent lateral channel movement (10 to 15%);
moderate channel scour or channel change within
streambed.

IV

O

More than 20% lateral channel movement and bank
cutting; changing channels and severe scour
evident; source of extreme sedimentation.

I = Excellent Q = Good Q = Fair Q = Poor

Pool Quality

Pools are the deeper, slow-moving sections of
streams which provide resting areas for fish. A vari-
ety of pool classes will provide the greatest diversity
and fulfill the requirements for various stages of fish
life.

Pool quality is an estimate of the pool's capabil-
ity to provide habitat for the various life stages of
fishes. A method of rating pools has been developed
that considers pool size, depth, and cover ratings
(Table 4). By using these measurements, pools are
categorized into five pool classes of which one is the
best. Nevertheless, pools with low ratings may have
some value. For example, a pool that receives a low
rating of 1 for depth and 1 for size may still provide
essential habitat for the survival of young fish. Hamil-
ton and Bergersen (1984) describe pool rating sys-
tems in detail.

Riffles

Riffles are the steeper gradient, higher velocity
stream sections. They generally consist of coarse
gravel, rubble, or larger substrate. Riffles are used for
spawning and also produce macroinvertebrates — a
primary source of fish food. Headwater streams are
often primarily composed of riffles with few pools.

Riffles are not usually rated the way pools are.
However, riffle length and width can be measured
with a tape or ocular estimate. Pool-to-riffle ratio is
often used as an indicator of stream habitat
condition.

Streambed

Streambed material can be classified by particle
size according to Table 5. A stream transect or cross

Meandering, low-gradient stream.

Steeper gradient, as indicated by riflles.

Aquatic Physical Features

607

Table 4. Pool quality rating system (after Hamilton and Bergersen 1984).

Category

Rating1

Description

Pool size

3
2
1

Pool larger or wider than average width of stream
Pool as wide or long as average width of stream
Pool shorter or narrower than average width of stream

Depth

First through Fourth
Order streams

Fifth Order streams
or larger

3
2

1

3
2

1

Over 3 feet deep
2-3 feet deep
Under 2 feet deep

Depth 20% of average stream width or greater
Depth 1 0-20% of average stream width
Depth less than 10% of average stream width

Cover

3
2
1

Abundant cover
Partial cover
Exposed

1Pool size, depth, and cover ratings are summed to provide an overall pool rating according to the following scheme:

Total Ratings

Pool Class

8-9

1

7

2

5-62

3

4-5

4

3

5

2Sum of 5 must include 2 for depth and 2 for cover.

Table 5. Classification of streambed material (from
Duff and Cooper 1978).

An electronic depth finder.

Material

Particle Size

Bedrock

Exposed solid rock

Boulder

+ 12 in. (305 mm)

Large rubble

6-12 in. (152-305 mm)

Small rubble

3-6 in. (76-152 mm)

Coarse gravel

1-3 in. (25-76 mm)

Fine gravel

0.1-1.0 in. (2.5-25 mm)

Sand

(0.074-2.5 mm)

Clay

(0.074 mm)

Organic muck
(sapropel)

Decomposed organic
matter

Organic debris (detritus)

Undecomposed organic
matter

608

Aquatic Physical Features

oo c F

3 O CO

C <= c

flj (1) 0)

5- E -

g^-6 E

" « p

03 CD

<» TT

-C >>

o c

*-" co

(D c

5 E

CD O

■O "O

O CD

TO

" _5

cd o °
5 O m
0) Q.]£
"*~ en CO

en 2

3 MT3

q. y "

-tr E to
</> . 3
o §

-f|

o go o <"
V ca -d E

a)

■9 °
cd to
> ■-
OQ

CT°
CD E

73 E

c

CD

E «
0) ">

£ o

8^?

C Q_ cj
03 03 o

£ ™ E

en <D

CD ^

CD

o*

3 u

>. C

o

a -SC-
SI'S

w C

CM

03 03

V

CD -C

tfl

LL "D

>. O

to o

P. 0)

3 ■« §

§"- 5-

— ■* 5

to CM <5

o ,_ €

F B o

4? > >

< O O

y e

tr o_

— 03 O

o g. aj

Q. >• CO

LU Q.<

E <D

O to

~° 8

Q. CO

8 S|
-I O

CD >■

"O ^

> 03

CD CD
O) w

c cr.

■5 CD

o E

03 tr
Q. o

° CO

Z 03

2 .05

5 2

13 <?

T3

CD CD

-£ 23

03 03
2 CO

*n

^

CO GO cr

o

O

CD - CD

Q. c oo

X

>fl) U)

)

— CD J-

CO

**-

m

« B3a

o

c
o

C i CO

c g -Q

b

2
to

CO
13

Pere
Yello
may

^

CO

>s

t£ CD

cd x:

Sj to a>

O" o sz

CD Q. a)

- to c

CD 5 §

03 2 -o

® CD Jr,

•D p 2

■5 ° §

CD CD

> £ "O oo

i s « «

o o 5 E

CD 5 03 O

T) or 3

I T] O O

o c M «

IT) 03 03 "O

CO oo
CD T3
Q. o

*- O

if OQ

™ 2 5

Eo0£

gS 03

o "5

6 2

CM "O

CO

=1 o

cr d

CD >=

.£= oo

5 E

Sj o

o JX v>

E £ o

^ 03 O

oo X3 Q.

I?°

CD 00 O)

= =3 C

CD CO =

"D O ^z

i2 2 CD

3 "E T3

o g> >

„ o <"

c _ ci)

CO CO c

^Er

^ „ D)

g> o o

co tr go

5 CO >=

— oo CO

M cd ■Q

CD CO J

X) O c

CD ° g

?!? §

CD CD 00

73 £ T3

o CO u

5 O) to

§.9

CD "

08 .1

00 X3
CD O

E 5

E

oo M

00 »-

CO CD

O CD

£6

CD CM
O) jo
05 «
o CD

D

n

o

n

u

10

CO

u

JJ

-

c CO

CO

T3 -Q

O CO

^55

■= o

to C3)

o ff o

Q. 03 *"

CO ,_ QJ

-Q I E

^ -^ CO

¥ ^ E

cp

_z 13
CO =3

i;

CD

2 5

«^

„ CO

C CD _

CD SI ro

CD 00 CD

•£■§ °

rr

Li

-i

i )

c/>

o

m

*

U

n

ni

13)

<u

ii

0)

u

00

5

o

CD UJ

> TT

O co

03 "O

ol co

cs o o) S
cic^g

Is- CO 03 t

n

c

iz=

01

CD

n

CO

CO

a

r

r

c

00

u

E
"a

o

tj

3

c
to

u

CO

CO

ni

^_r

(71

iJ

<

C_
CD

c
CD

u

CD

C

U)

c/)

i

LI

<)

0)

r

CO
CO

o

o

u
CB

i )

u

(I)

U)

o

Q.-D

en

CO

CO

i>

o

c_

o

CNI

c

u

CO

t—

■

(-

>.

y

o

0)

o

r

CO

u.

01

en

J

(1)

E

o

CD

xJ

o

c

oo

o
F

<u

m

5

c

o

c_

03

E

CD

C

CO

CD

"1

c

CD

CO

I'll

b

o

b

c

>

m

E8

8|

S^
5 =)

cr oo

CO

-5 ^

to ^5

CO CL
CD i?

4= O

Z a?

"3

o10

~ u)

£ S

« to
Q E

■= en 5

=3 CD CO

O _c O

O > Q.

w > a>

•a T3

■O c ni

CD CO <K

t3 CO k

0) c o

-*- n C/J

3= O

c»o
o S o

S i» w

O to S
— O CD

.E 5 «

« 00 g

co -5 oo
< Q. c

coa *
°>-8

E 8ff
CJ^r:

i; O)

CO CO
CO CO

a- 5

CD CO

y e

CD ^ CI

C C
CD C
CO CO

-Q sz
CO o
_>^ CD

to ffl
^ T3

□3 .

s °-

"-1 oo co

_ CD p

+ >• CD

SS SJ c?

O CO 03

oo v

CO

JZ) CO O

5 J 2

c — ^-~

CD j: Q

co 03 S

a; cd §

O.Q- —

i^ CD CD

«, OT CZ

I 8 §

03 E

5 oo

, CD

+ -n

CD *-

c « O

E an

™ si

SSa

S co^
o E

cc CD ofl

cu

r

>

CO

_>.

CD

^

CB

CD

m

Hi

( )

C

cu

D)

i

! 1

Ml

CU

CD

r

P m

JO C

c o

CD Q.

— CO

Cl CD

o 91

3 CZ
T3 CD

-CD

CO

SI ~o

3 2
C/) 00

E

Q <=

g> c

00 CD
T3 §

< CO

O) 5

SE

r: a)

<-> X5
O ™

Z co

00 CD Q-
CD 3Z CD
_l CO "O

O) - 2

CO 5 5

r. 2 *

2 ^ co

CDS c

CO ^£ =

c M s.

23 00 CD

< E Q.

n

to

C)

x

00

^

E

ea

o

cr

o

Lll

c

Q

_l

O

O) "■
SO.

oo -^
to 00
CD 3

in o

O CD

E °

CO O

c?2

J3 to

CD O

to
o_

<? CO

o -t

ill

) oo E
i ~° -o

) 5 CD

: O W

o .5?

3 CD

2 co

P -o

II

g)fo

Aquatic Physical Features

609

section can be inventoried for particle size by per-
cent of the total cross section, and the transect site
marked by steel stakes for future monitoring (Duff
and Cooper 1978). The amount of the various parti-
cle sizes is estimated by ocular estimates. Estimates
are not reproducible to a high degree of accuracy
even by the same stream surveyor. This variable
does not have a high degree of precision from one
biologist to another or from one year to the next.
However, it is valuable to describe existing habitat
conditions on the date of the survey, which can be
used for general trend comparisons on a continuing
inventory basis. Hamilton and Bergersen (1984) de-
scribe substrate sampling in detail.

Surface Acreage

Surface acreage can be determined from a map
or aerial photograph using an acreage dot grid or
a polar planimeter. An average of three readings
from the map or photograph should be used for
acreage computation. The 256-dots-per-square-inch
dot grid should be adequate for lake and reservoir
measurement. Increased precision can be achieved
by using the microdot grid with 1,024 dots per
square inch. Hamilton and Bergersen (1984) describe
this technique and others for calculating acreage.

Shoreline Miles

The number of shoreline miles can also be de-
termined using a planimeter. This measurement is
used to calculate an index of shoreline development
described below.

Shore Development

Shore (or shoreline) development is a measure
of the convolution of the shoreline. An irregular
shoreline provides more littoral zones for the pro-
duction of aquatic fish food organisms and more
spawning and nursing areas for fish. An index to
shoreline development can be calculated using the
ratio of actual shoreline length to shoreline length of
a circular lake of the same acreage, as follows:

shore development =

2 A

where S equals the length of shoreline and A equals
the surface area of the lake in acres.

Water Temperature

Water temperature can be accurately recorded
by a pocket thermometer, maximum-minimum ther-
mometer, or thermograph. The temperature variable

A surface scatter turbidimeter being used to monitor stream turbidity- during a controlled
reservoir release designed to flush fine sediments from the stream bed.

610

Aquatic Physical Features

Table 6. Average turbidities found to be fatal to fish (from Federal Water Pollution Control
Administration 1968).

Length of

Turbidity

Species

Exposure (days)

(mg/l)

Largemouth bass (Micropterus salmoides)

7.6

101,000

Pumpkin seed sunfish (Lepomis gibliosus)

13.0

69,000

Channel catfish (Ictalurus punctatus)

9.3

85,000

Black bullhead (/. melas)

17.0

222,000

Golden shiner (Notemigonus crysoleucas)

7.1

166,000

is reproducible with a high degree of accuracy when
instruments are accurately calibrated. Hamilton and
Bergersen (1984) describe temperature measure-
ments in greater detail.

Turbidity

Turbidity is an expression of the optical prop-
erty of a sample of water which causes light to be
scattered and absorbed rather than transmitted in
straight lines through a sample (Federal Water Pollu-
tion Control Administration 1968). Turbidity is
caused by the presence of suspended material such
as clay, silt, finely divided organic matter, plankton,
and other microscopic organisms. Turbidity is mea-
sured by turbidimeters in Jackson Turbidity Units
(JTUs) or Nephelometric Turbidity Units (NTUs).

Turbidity of not more than 25 JTUs (25 NTUs)
is recommended for cold-water streams and not
more than 50 JTUs (50 NTUs) for warm -water
streams. High turbidities can cause fish mortality as
shown in Table 6. The examples given are unusually
high for extended periods of time. Instream activity
such as dredging and placer mining can cause high
turbidities that may not kill fish directly but cause

other detrimental effects such as streambed siltation.
Mining can cause high turbidities that may not kill
fish but may cover the streambed and suffocate fish
eggs and macroinvertebrates.

Hamilton and Bergersen (1984) discuss turbidity
in detail.

SUMMARY

Inventory and monitoring should be directed to
the variables that have the greatest possibility of
repeated results by various workers. The variables
that can be measured with the greatest accuracy
(i.e., water temperature, stream width, stream depth,
lake depth, surface acreage, stream gradient, and
turbidity) should be used in monitoring change in
habitat conditions related to specific land-use ac-
tions. Inventory and monitoring should measure
those variables that limit distribution and abundance
of aquatic organisms. Measured variables should
serve as indicators of effectiveness of management
practice. The ultimate measure of the effectiveness
or impact of management should be productivity of
the fish or other aquatic species.

Aquatic Physical Features

611

LITERATURE CITED

ARMOUR, C.L., K.P. BURNHAM, and W.S. PLATTS. 1983.
Field methods and statistical analysis for monitoring
small salmonid streams. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-83/33. 200pp.

BOVEE, K.D. and R. MILHOUS. 1978. Hydraulic simulation
in instream flow studies: Theory and techniques.
IFIP 5, Cooperative Instream Flow Serv. Group, U.S.
Dep. Inter., Fish and Wildl. Serv., Ft. Collins, CO.
130pp.

CUPLIN, P. 1978. The use of large scale color infrared
photography for stream habitat inventory. U.S. Dep.
Inter., Bur. Land Manage. Tech. Note TN-325. Denver,
CO. 11pp.

, R. BOROVICKA, J. ERDMANN, B. VAN HAVEREN,

L. LEE, R. MCQUISTEN, and M. WHITTINGTON. 1979.
Instream flow guidelines. U.S. Dep. Inter., Bur. Land
Manage., Service Center. Denver, CO. 57pp.

DUFF, DA. and J.L. COOPER. 1978. Techniques for con-
ducting stream habitat surveys on National Resource
Land. U.S. Dep. Inter., Bur. Land Manage., Tech. Note
TN-283. 72pp.

FEDERAL WATER POLLUTION CONTROL ADMINIS-
TRATION. 1968. Water quality criteria. Washington,
DC. 234pp.

HAMILTON, K. and E.P. BERGERSEN. 1984. Methods to
estimate aquatic habitat variables. Colorado Coop.
Fishery Res. Unit. Colorado State Univ., Fort Collins.

HAYS, R.L., C. SUMMERS, and W. SEITZ. 1981. Estimating
wildlife habitat variables. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-81-47. 111pp.

PARSONS, S.C. and S. HUDSON. 1985. Stream channel
cross-section surveys and data analysis. U.S. Dep.
Inter., Bur. Land Manage. TR-4341-1. Denver, CO.
48pp.

PFANKUCH, D.J. 1975. Stream reach inventory and chan-
nel stability evaluation. U.S. Dep. Agric, For. Serv.
Northern Region. R 1-75-002. 26pp.

PLATTS, W.S. 1981. Stream inventory: Acquisition and
utilization of aquatic habitat inventory information.
Pages 75-84 in Proc. Symp. Western Div. Am. Fish.
Soc, Portland, Oregon, October 1981.

, W.F. MEGAHAN, and G.W. MINSHALL 1983- Meth-
ods for evaluating stream, riparian, and biotic condi-
tions. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. INT-
138. 70pp.

ROBINS, C.R. and R.W. CRAWFORD. 1954. A short accu-
rate method for estimating the volume of streamflow.
J. Wildl. Manage. 18:366-369.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF
LAND MANAGEMENT. 1979. Wetland-riparian area
protection and management. Manual 6740. Washing-
ton, DC. 16pp.

612

Aquatic Physical Features

29

HYDROLOGIC
PROPERTIES

Bruce Van Haveren

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

Editor's Note: Biologists know that animals do not
exist in isolation and that the habitat in which
they live — the ecosystem of which they are a part —
must be understood. This interrelationship is fre-
quently the first principle ignored by students as
they eagerly pursue a narrow disciplinary ap-
proach. Water is the single most important life
requirement for all plant and animal life. Thus, the
movement and properties of water in an animal's
habitat should be of concern to the biologists con-
ducting a habitat inventory and monitoring effort,
particularly in the arid western United States. Yet
hydrology, the study of water in the natural envi-
ronment, is virtually ignored by many biologists.
This chapter describes some of the common meas-
urements of hydrologic properties useful to the
biologist.

INTRODUCTION

Water, one of the vital wildlife habitat elements
on publie lands, is the lifeblood of all ecosystems,
terrestrial as well as aquatic. Its natural occurrence
and circulation are especially vital to the functioning
and vigor of wetland and riparian systems. Biologists
must remember that all terrestrial and aquatic eco-
systems are tied to hydrologic and geomorphic sys-
tems. For example, to manage riparian habitat, one
must consider the surface and groundwater systems
that are hydrologically related.

Fishery and wildlife biologists should understand
that subtle changes in land use can produce not-so-
subtle alterations in hydrologic conditions, which
can directly influence fish and wildife habitat. Fur-
thermore, hydrologic changes may occur several
miles from the habitat site that is or could be
affected.

This chapter focuses on the measurement and
use of certain hydrologic properties important to
wildlife habitat. The hydrologic properties of both
lentic (lakes, ponds, bogs, meadows) and lotic (riv-
ers, streams, springs) habitats are discussed in terms
of measurement techniques and data applications.

Hydrologic Terminology

Biologists should be familiar with common
terms used in land-use hydrology' to facilitate com-
munication with their hydrologist colleagues. Precip-
itation, evaporation, runoff, soil water storage, and
sediment transport are all examples of hydrologic
processes. Hydrologic variables, such as instanta-
neous stream discharge, monthly groundwater levels,
or daily lake evaporation, quantitatively describe
hydrologic processes. Hydrologic variables are nearly
always continuous in time and, therefore, must be
observed or measured at given intervals to convert
them to discrete variables. When discrete variables

Hydrologic Properties

613

are described statistically, they are called hydrologic
parameters. Mean annual precipitation, the 100-year
peak river discharge, and the coefficient of variation
of monthly suspended sediment are all examples
of hydrologic parameters.

The era of the Environmental Impact Statement
(EIS) has brought with it a new vocabulary. In the
field of hydrology, terms such as hydrologic changes,
effects, and impacts are all used in environmental
reports, often interchangeably. However, used cor-
rectly, hydrologic change refers to the increase or
decrease (statistically at a given level of probability)
in a hydrologic variable or parameter. For example, a
100-mg/L increase (significant at the 0.80 level of
probability) in mean daily suspended sediment rep-
resents a hydrologic change.

The term hydrologic effect invokes a cause and
effect relationship. A significant hydrologic effect
would be road construction causing the 100-mg/L
increase in mean daily suspended sediment.

Finally, a distinction must be made between
effects and impacts. From a hydrologic or water re-
sources standpoint, an impact occurs only when the
value of that resource diminishes (negative impact)
or increases (positive impact). Referring back to our
suspended sediment example, the 100-mg/L increase
due to road construction cannot be called an impact
unless the effect has resulted in a change in a related
resource value.

Watershed Stability: The Concept

Watershed stability has an inherent relationship
to the stability and productivity of wildlife habitat.
Watershed stability includes the following elements:
soil productivity, hydrologic response to precipita-
tion, quality of surface and shallow groundwater,
movement of nutrients by water, and structure and
function of wetland-riparian systems.

These may be viewed as components as well as
partial indexes of watershed stability. They are
highly interrelated and directly influence habitat
variables.

If a simple equation could be used for measur-
ing the present watershed stability, it would look
like this:

Watershed stability = watershed constants +
climate variables + present watershed condition +
land-use variables.

The watershed constants that influence stability
include geology, soils, latitude and elevation, topog-
raphy (relief, slope, and aspect), and vegetation type,

both present and potential. These factors are static
over time and easily determined.

Watershed condition and trend are determined
semi-quantitatively by evaluating soil erodibility (see
Soils chapter), mass wasting potential (potential for
mass slope failures), stream channel stability (see
section on Stream Channel Stability and Sediment
Transport, this chapter), and watershed infiltration-
storage capacity.

The climate factors important to watershed sta-
bility are limited to the frequency and intensity of
runoff-producing precipitation, seasonality of precipi-
tation, and the frequency and duration of drought.

Land use refers to past and present land-use
factors that influence watershed stability, such as
road construction, timber harvesting, excessive live-
stock trampling, overutilization of forage, mining,
and mineral or energy exploration activities. Also in-
cluded would be natural events that may affect wa-
tershed stability such as storms, wildfires, floods, and
insect or disease attacks on vegetation.

All of these factors are evaluated in the present,
assigned an appropriate weight, and then combined
into an overall watershed stability rating. Trend in
watershed stability can be inferred by comparing
past and present watershed stability ratings and by
considering present and near-future land-use activi-
ties that influence watershed stability.

Watershed stability ratings are subjective. Rating
systems are nearly always tailored to fit specific lo-
cales and applications. For this reason, information
on watershed stability is generally not available from
other sources and must be obtained directly from
field surveys. Before field work, the area in question
must be stratified on the basis of hydrologic re-
sponse units (homogeneous areas that respond simi-
larly to a given hydrologic event).

Because land management activities can affect
stability, the specialist must determine and present
to management the acceptable versus unacceptable
levels of watershed stability. The consequences of
unacceptable watershed stability must be thoroughly
explained in terms of impact magnitude and proba-
bility of occurrence.

PRECIPITATION, SNOW, AND
EVAPORATION

Wildlife biologists may on occasion have need
for hydrometeorological data such as precipitation,
evaporation, and snow depth. Precipitation and evap-
oration have an indirect influence on habitat through
the fluctuations of lake and pond levels and the
water balance of wetland ecosystems. Snow, in terms

614

Hydrologic Properties

of depth and areal distribution, affects animal move-
ment and food availability, and thus the health of
many wildlife species.

Measurement of Precipitation

Precipitation includes rain, snow, and all other
forms of precipitated atmospheric water. The biolo-
gist should first ascertain the type and quality of data
needed. Are data needed on a year-round basis or
just seasonally? What is the desired frequency of the
data: monthly, daily, hourly? Will snowfall data be
needed as well as rainfall data? These questions must
first be answered before a gage and sample schedule
are selected. For most wildlife applications, monthly
precipitation totals collected year-round would suf-
fice, particularly for U.S. Bureau of Land Management
(BLM) special wildlife management areas where
water supply is a concern. Many national wildlife
refuges, for example, routinely collect this type of
precipitation data.

Cumulative rain is measured in a storage can
having an orifice diameter of at least 20.3 cm (8 in.).
Many rain gages are available that will store up to a
year's accumulation of precipitation. Lightweight
oil is added to retard evaporation.

Recording gages are necessary if one needs to
know rainfall intensity or the time of occurrence of
rainfall. There are several commercial types available.
Agricultural Handbook 224 (Brakensiek et al. 1979)
includes a good discussion on precipitation record-
ing and characteristics of various gage types.

Measurement of precipitation falling as snow is
much more difficult. Most precipitation gages used

currently in the U.S. underestimate snowfall. Effi-
ciency of catch depends on the size of orifice, gage
siting, and whether or not the gage has a windshield
(Figure 1). The efficiency of catch of snowfall at
this site has been at least 90% . Catch efficiency is
checked by measuring the water content of snow on
the ground at the gage and comparing this to the
gage catch for the same period of time.

Measurement of Evaporation

Evaporation is also a difficult variable to meas-
ure. Lake or pond evaporation is usually inferred
from measurements of pan evaporation. The National
Weather Service's 1.2-m (4-ft) diameter class A evap-
oration pan (Figure 2) is the standard instrument
for evaporation measurements in the U.S. The U.S.
Agriculture Handbook 224 contains a good discus-
sion on siting and operation of a class A pan.

Figure 1. A precipitation storage gage with an
"alter shield" at an ideal gage site in a forest
clearing.

Figure 2. Class A evaporation pan.

Measurement of Snow on the Ground

Depth of snow on the ground is important be-
cause it may limit access to food and make move-
ment difficult for big game animals. Depth of snow is
measured with a yardstick or graduated rod. Snow
depth, because of its extreme variability, must be
sampled carefully and systematically. Stratified sam-
pling, considering elevation, vegetation type, and
slope aspect strata, is recommended.

Hydrologic Properties

615

Measurement of the water content of snow is of
interest for calculating water balances; these data
may also supplement or verify winter precipitation
gage catch. The U.S. Soil Conservation Service (SCS),
for example, operates a Westwide network of snow
surveys used for forecasting spring and summer run-
off for irrigators, reservoir operators, and hydro-
power companies. There is no reason why the same
techniques could not be used to forecast wildlife
water supplies in certain situations.

The standard federal snow sampler is used for
measuring both depth and water content in moun-
tain snowpacks (Figure 3). Agriculture Handbook
224 gives a brief description of snow surveying. The
SCS State snow survey supervisor (located in the
SCS State Office of each western State except Califor-
nia) can be contacted for more information and
assistance.

Snow density, the ratio of snow water content
to snow depth, can be computed directly from the
measurements data taken with the federal snow sam-
pler. Snow density is a good indicator of the matu-
rity or ripeness of a seasonal snowpack. New snow
densities vary between 0.07 and 0.25 g/cm . Late
season snowpacks range between 0.30 and 0.45 g/
cm3. As snowpack density approaches 0.50 g/cm ,
the snowpack is said to be "ripe" and is ready for
snowmelt release.

Figure 3. A standard federal snow sampler kit for
measuring snow depth and water content.

ators, U.S. Forest Service, and Agricultural Research
Service experiment stations. Snow data are available
for selected NWS stations and from the SCS.

Hydrometeorological data are scarce for public
lands in general. The State climatologist is generally
a good source of information about data availability
for specific areas within the State. The State climatol-
ogist can also assist with data extrapolation and
interpretation.

STREAMFLOW MEASUREMENT AND
CHARACTERISTICS

Wildlife biologists, and particularly fishery biolo-
gists, occasionally need streamflow data. The antici-
pated use of these data will determine the data type
and desired accuracy. For example, for some pur-
poses such as water quality grab-sampling, an instan-
taneous discharge measurement will suffice (and
should always accompany a water quality sample).
For long-term characterization of streamflow, dis-
charge data continuously recorded over several years
are necessary. To characterize flow and channel hy-
draulic conditions for a given channel reach, one
would need only a few discharge measurements. The
BLM Instream Flow Guidelines (Cuplin and Van Hav-
eren 1979) should be consulted for advice on the
measurement of stream discharge in instream flow
studies.

A decision on the desired accuracy of these data
is important because it directly relates to the mea-
surement method chosen and corresponding cost.
Eight streamflow measurement methods can be used
for typical small stream applications:

( 1 ) Float method

(2) Volumetric measurements

(3) Current-meter and velocity-head measure-
ments

(4) Portable weirs and flumes

(5) Dye-dilution method

(6) Continuous streamflow recording — natural
control

(7) Continuous streamflow recording — artificial
control

(8) Indirect methods

Float Method

Sources of Published Data

Published precipitation, evaporation, and snow
data are available for the more populated areas in
the West. Sources of precipitation and evaporation
data include the National Weather Service (NWS),
State agricultural experiment stations, reservoir oper-

The float method is the least expensive, but
least accurate, of the available methods. A partially
submerged float (water-logged stick, orange, piece of
ice, or half-full beer can) is timed through a meas-
ured reach to estimate average water velocity (the
average of several runs is computed ). A cross section
of the wetted portion of the channel is then meas-
ured. The stream discharge is therefore the product

616

Hydrologic Properties

of the average velocity and the cross section. If a
good estimate of velocity is obtained, the method is
accurate within ± 10%. Under conditions where
velocity cannot be estimated adequately, the error
could be ±25%.

Volumetric Measurements

Volumetric measurements are taken whenever
the flow is sufficiently small and is concentrated
or can be concentrated so that all of it may be
caught in a container. Time required to fill a known
volume is measured with a stopwatch. The method
is fairly accurate, with the error averaging ± 2 to 3%
or less. Volumetric measurements can often be made
at culverts by using a 19-L or 38-L (5-gal. or 10-gal.)
bucket for small flows or a 208-L (55-gal.) drum
for larger flows. I have often carried a piece of thin
sheet metal in my vehicle for forming temporary
flumes to facilitate volumetric measurements.

At culvert sites where the flow is too large for a
volumetric measurement, the float method may be
used with reasonable accuracy.

Current-Meter and Velocity-Head
Measurements

Although several types of current meters are in
use at this time, the Price AA and pygmy vertical-axis
meters (Figure 4) are used most commonly by the
U.S. Geological Survey, Water Resources Division.
Because current meters measure velocity at a point,
several velocity measurements are taken at a stream
cross section. The number of measurements in the
vertical direction is chosen according to the accu-
racy desired. The one-point procedure, where a sin-
gle velocity measurement is taken (for each
horizontal point on the cross section) at 18.3-cm
(0.6-ft) depth below the surface, is most commonly
used for small streams. In larger streams and where
more accuracy is desired, velocity measurements are
taken at the 6.1- and 24.4-cm (0.2- and 0.8-ft) depths
below the surface. The gain in accuracy is about one
percentage point. Additional accuracy may be ob-
tained with the three-point method at 6-, 18-, and
24-cm (0.2-, 0.6-, and 0.8-ft) depths where stream
depths exceed 76.2 cm (2.5 ft). Spacing of measure-
ment points across the channel depends on the
width of the stream. Normal spacing is within 0.6 or

Figure 4. Price AA and pygmy current meters for measuring stream velocity.

Hydrologic Properties

617

1.5 m (2 or 5 ft); however, the points should be
spaced so that no subsection has more than 10% of
the total discharge.

Selection of a suitable site for current meter
measurements should follow these criteria (see
Figure 5):

( 1 ) Cross section lies within a straight reach and
flow lines are parallel.

(2) Velocities exceed 15 cm/sec (0.5 ft/see) and
maximum depths exceed 15.2 cm (0.5 ft).

(3) Streambed is relatively uniform, free of large
boulders, organic debris, and aquatic growth.

(4) Flow is uniform, free of eddies, slack water,
and excessive turbulence.

(5) The stream can be waded from bank to bank.

If a stream is too deep or velocities too high
for wading, a suitable bridge site must be located. In
this situation, a suspension cable is substituted for
the wading rod. Procedures for taking current meter
measurements from bridges are covered in detail
by Rantz (1982).

A standard discharge measurement form used by
the U.S. Geological Survey for all stream discharge
measurements is shown in Figure 6. The "midsection
method" (Buchanan and Somers 1969; Rantz 1982)
is used to compute stream discharge from this form.

Accuracy of the current meter method depends
on flow and channel conditions, number of velocity
measurements taken in the vertical direction, spacing
of measurements across the channel, condition and
calibration of current meter, and operator error. U.S.
Geological Survey provides the following accuracy
limits:

Overall Rating of
Measurement

Excellent
Good
Fair
Poor

Error

±2%
±5%
±8%
>8%

In spite of their lower cost and ease of opera-
tion, velocity-head rods (Figure 7) have not achieved
the same popularity as current meters for gaging
streams. The rod consists of a broad edge at least 2.5

Figure 5. A good location for a current meter measurement.
618 Hyiirologic Properties

cm (1 in.) wide and a sharp edge opposed 180° to
the broad edge. Constructed of hardwood or alumi-
num, the rod is scaled in English or metric units.
To measure the velocity head, the sharp edge of the
rod is pointed upstream and the water elevation read
from the rod. The rod is then revolved 180° so the
broad edge is perpendicular to the flow lines. The
water elevation is again read, and the difference be-
tween readings computed. The resulting number
is the velocity head expressed in feet or meters. Ve-
locity m/sec (ft/sec) is found by the equation —

V = 2gh

where g = acceleration due to gravity (ft/sec/sec)
and h the velocity head (ft).

Under very good channel conditions, the veloc-
ity-head rod will give discharge readings within a
few percentage points of the true value (Wilm and
Storey 1944). It is comparable in accuracy to the
one-point method using current meters. Heede
(1974) recommended use of the velocity-head rod
only if channel bottoms are fairly smooth and chan-
nels are approximately prismatic. This precludes
their use in boulder-strewn mountain streams with
highly turbulent flows, unless the gaged section can
be modified to fit the above criteria.

Because of the surging nature of flow in natural
streams, the precision of head measurements is
around ± 0.01 m ( ± 0.05 ft). Heede (1974) recom-
mended reading the water elevation that occurs
most frequently, rather than the average elevation.

Portable Weirs and Flumes

Portable weir plates and pre-calibrated flumes
can be used to accurately measure small flows. Both
rectangular and triangular (30°, 60°, 90° or 120°
notches ) weirs are used; the specific design is based
on the most probable range of expected flows (Van
Haveren 1986).

Parshall flumes are the most common pre-cali-
brated, portable flumes in use (Figure 8). They are
commercially available in a variety of sizes and are
easily installed at most locations (see Rantz 1982).

Dye-Dilution Method

The dye-dilution method is fairly sophisticated
and requires moderately expensive sampling equip-
ment. A known quantity of tracer dye is injected
at the upstream end of the reach being measured,
allowed to mix thoroughly through natural stream
turbulence, and the resulting concentration meas-
ured at a downstream point. The technique has defi-
nite advantages in very turbulent mountain streams
where more traditional stream gaging techniques are

not possible. Intensive training is required to effec-
tively use the technique. Refer to Rantz ( 1982) for a
complete discussion of this technique.

Continuous Streamflow Measurement —
Natural Control

Continuous measurement of streamflow gener-
ally involves more expense and expertise than the
methods discussed above. The method involves di-
rect sampling of water level (stage) and the estab-
lishment of a stage-discharge relationship from which
streamflow is computed. This method is expensive
in terms of labor and equipment. As in all stream
gaging methods, site selection is critical. Where an
artificial control (e.g., bridge, culvert, flume, bank
revetment) is not used, a suitable natural control
must be found. Rantz ( 1982) listed criteria to be
used in selecting a gage site with a natural control.
Water level is recorded continuously with a stage
recorder. Stream discharge is related to stage height
by periodically making discharge measurements at
known stages. A stage-discharge rating curve is estab-
lished from these measurements, taken at a wide
range of stage heights. The rating curve is checked
and modified, usually annually. The stage-discharge
curve will remain stable at sites having good natural
channel control and no net sediment accumulation
or channel erosion.

Continuous Streamflow Recording —
Artificial Control

Artificial controls are installed where a good
natural control cannot be found, or when a very
high degree of accuracy is desired. Artificial controls
include weirs; flumes; and concrete, rock, or treated
wood control structures that maintain a stable chan-
nel geometry. Further discussion of the types of arti-
ficial controls available, including their advantages
and disadvantages, is beyond the scope of this chap-
ter. Brakensiek et al. (1979) and Rantz ( 1982) pro-
vided good discussions of weirs and flumes. Some
weirs and flumes are available with known calibra-
tions, precluding the need for field development
of the stage-discharge rating curve.

Discharge measurements can also be made in
reservoirs where presumably all inflow is contained
in the structure with negligible leakage. A water
level recorder is used to measure water depth,
which is related to inflow volume through a reser-
voir stage-capacity curve.

Indirect Methods

Indirect methods of streamflow measurement
include channel geometry relationships and the Man-
nings equation.

Hydrologic Properties

619

9-275-G (Slnffle 5x8)

(Rev. 5-63)

Sta. No.

UNITED STATES

DEPARTMENT OF THE INTERIOR

GEOLOGICAL SURVEY

WATER RESOURCES DIVISION

DISCHARGE MEASUREMENT NOTES

Meu. No.

Comp. by -.
Checked by

Date , 19 Party

Width Area Vel G. H.

Method No. sees. G. H. change

Disch.

in hrs. Susp.

Method coef. .. _ Hor. angle coef. Susp. coef Meter No.

GAGE READINGS

Date rated . Used rating

Time

Recorder

Inside

Outside

for rod susp. Meter . _. ft.

above bottom or wt. lags checked

Spin before meas after .

Meas. plots % diff. from .. rating

Wading, cable, ice, boat, upstr., downstr., side

bridge feet, mile, above, below

gage, and

Check-bar, chain found

Weighted M. G. H..._

changed to at

Correct

Correct M.

G. H

Levels obtained

Measurement rated excellent (2 %), good (5%), fair (8%), poor (over 8%), based on following

conditions: Cross section

Flow Weather

Other Air °F@

Gage Water

-°F@

Record removed Intake flushed fr.

Observer

Control

Remarks

G. H. of zero flow ft.

■fr GPO : 1963— O-688360

Figure 6. Standard discharge measurement form used by the U.S. Geological Survey.

620

Hydrologic Properties

.0 .10 .20 .30 .40 .50 .60 .70 .75

River at —

— 'o
<

Dist.
from
initial
point

Width

Depth

> &
0-2

Rev-
olu-
tions

Time

in
sec-
onds

VELOCITY

Adjusted
for hor.
angle or

Area

.80
Discharge

At
point

Mean
in ver-
tical

.85

.90

.92

.94

.96

.97

.98

.99

o

1.00

.99

.98

.97

.96

.94

.92

.90

.85

.80

.0 .10 .20 .30 .40 .50 .60 .70 .75

Figure 6. Standard discharge measurement form used by the U.S. Geological Survey
(concluded).

Hydrologic Properties

621

The Manning equation can be used to estimate
flows in natural channels:

1.486AR23S,/2

Figure 7. A velocity-head rod for measuring stream
velocity.

Q =

n

where A is the cross section of the channel, R the
hydraulic radius, S the slope of the energy line (usu-
ally approximated by the water-surface slope ), and
n the roughness coefficient. All the independent
variables can be measured in the field except n,
which must be estimated. This technique is often
used for estimating flood peaks if sufficient high-
water marks can be found. A minimum error of 10%
can be expected even if the roughness coefficient
chosen is within ± 0.005 of the true value.

Channel geometry can be used to estimate mean
annual streamflow and flood peaks of various fre-
quencies. Using only active-channel width, Hedman
and Osterkamp (1982) constructed equations for
predicting the 2-, 5-, 10-, 25-, 50-, and 100-yr flood
discharges and the mean annual runoff for four re-
gions in the western U.S. State-specific channel ge-
ometry relations have been established for most
States by the respective Water Resources Division
District Office of the U.S. Geological Survey.

Field measurement of active-channel width re-
quires very little equipment and effort. Training is
required in order to properly identify the active
channel elevation.

KqU

Figure 8. A Parshall flume for the continuous measurement of streamflow.
622 Hydrologic Properties

Also, annual, monthly, or even daily streamflow
can be estimated by using index techniques. A gaged
stream with characteristics similar to the ungaged
stream of interest can be used to index or predict
some streamflow variables. This is accomplished by
first expressing the streamflow data on a "per square
mile" basis. Adjustments in certain streamflow vari-
ables should then be made, based on differences
in elevation, exposure, geology, or other appropriate
hydrophysiographic factors. This procedure is highly
subjective and should be done in consultation with
an experienced hydrologist.

Sources of Streamflow Data

Published streamflow data are available from the
U.S. Geological Survey, Water Resources Division,
on many larger streams and rivers in the nation. Data
for small watersheds in remote areas in the West
are more difficult to acquire. Figure 9 is taken from
Water Resources Data for Colorado, 1985 (U.S. Geo-
logical Survey 1985). This is a standardized stream-
flow data format used in every State. Maps, available
from the Water Resource Division Offices, show
locations of streamflow gaging stations.

HYDROLOGY OF LAKES, PONDS, AND
RESERVOIRS

The hydrologic factors that relate to wildlife
habitat characteristics of standing water bodies in-
clude inflow and outflow rates, volume (or capacity),
mean and maximum depths, basin shape, precipita-
tion-evaporation balance, drainage basin area, and
groundwater interaction (both physical and
chemical ).

Before undertaking a field inventory of a lake,
pond, or reservoir, an attempt should be made to
characterize the water body on the basis of origin,
geology, morphometric and physiographic properties
(which can be obtained from maps), drainage basin
characteristics, trophic state, and hydrologic setting.
Winter (1977) discussed hydrologic classification
and characterization of lakes.

Inflow and outflow characteristics (including
quantity and quality of the water) directly influence
lakes, ponds, and reservoirs. The measurement of
inflow or outflow usually involves stream gaging. Re-
fer to "Streamflow Measurement and Characteristics."

The State Engineer's Office is another potential
source of streamflow data. The local U.S Forest Ser-
vice hydrologist should also be consulted if national
forest lands are involved or are close to the area in
question.

Interpretation of Streamflow Data

Analysis and interpretation of streamflow data
depend on the intended use of these data. Discrete
values of streamflow (hourly, daily, monthly) can be
plotted as a "hydrograph" for visual analysis pur-
poses. Figure 10 is an example of an annual hydro-
graph based on monthly streamflow data. Low-flow
statistics, such as the minimum 7-day mean flow, can
be computed easily from daily flow data. For U.S.
Geological Survey continuous streamflow data, a
number of different streamflow statistics are available
through the WATSTORE data bank, accessible
through any U.S. Geological Survey Water Resources
Division.

Streamflow data for other than general interest
should be interpreted by an experienced hydrologist.
The wildlife biologist must tell the hydrologist the
specific objectives or intended uses of these data.

Flow data are usually expressed in units of vol-
ume per unit time, for example cubic feet per sec-
ond or cfs. Table 1 contains a set of factors for
converting different streamflow measurement units.

For water balance and certain limnological stud-
ies, lake volume often needs to be determined. As a
rough estimate, the lake surface area can be meas-
ured from a map, the basin shape assumed, and the
volume computed from mensuration formulas. A few
common lake basin shapes and their volume formu-
las are given in Figure 1 1 .

More accurate volume determinations are made
from detailed surveys and soundings of lake depths.
Depths are best measured in winter when a standard
land survey can be conducted on the frozen surface.
Depths are easily measured through holes augered
in the ice.

Water level often directly relates to habitat qual-
ity. Periodic, visual readings of water level can be
taken inexpensively with staff gages (Figure 12).
Continuous recording of water level requires the in-
stallation of a water level recorder, as in continuous
streamflow measurement.

The precipitation-evaporation ratio is important
in determining lake characteristics (Winter 1977).
The ratio should be calculated on an average annual
basis from reliable data. Precipitation data are usually
obtained from NWS. Evaporation data may also be
available from NWS or from a nearby reservoir oper-
ator. Pan evaporation data should be adjusted to
represent free-water evaporation.

Hydrologic Properties

623

GREEN

RIVER

3A3IN

09241000 ELK RIVER

AT CLARK

, CO

LOCATION

.— Lat 40°

43'03",

long 106

°54'55'

, in NWJNWJ

sec. 27

, T.9 N. ,

R.85 W. ,

Routt County, Hyd

rologic

Unit

14050001 ,

on left

bank 30

ft downstream from

bridge

on State

Highway

129, 0.8

mi north

of Clark,

and

2.0 mi upstream

from Co

ttonwood

Gulch.

DRAINAGE

AREA.--206 mi2.

PERIOD OF RECORD. -

-May 1910 to Sep

tember

192^

(publi:

hed as

"near CI

ark"), Ap

ril 1930

to current year.

Monthly

disch

arge only

for some

periods

, publishec

1 in WSP

1313.

REVISED

RECORDS. --

WSP 1733

: 1956.

GAGE. — Hater-stage

recorder. Datum of ga

ge is 7,267

75 ft,

(State H

ighway Dc

partment

bench mar

k ) . May

191C

to

September 1922,

nonrecording ga

ge at site

30 ft up

stream

at datum

0.15 ft

lower. A

pr. 23, 1930, to S

ept

27,

1934,

water-sta

ge recor

der at p

resent

si te

at datum 0.15

ft lower

REMARKS.

--Estimate

d daily

discharg

es : Nov

. 2C

to Apr.

7. Records goo

i except

for estimated daily dischar

ges

which

are poor.

Diversions abo

ve stat

ion

for im

gation

of abou

l 230 acres above

and about

460 acres below

station. Natural flow

of stream affected

by store

ge in

Lester Creek Reservoir (known also

as Pearl

Lake) ,

capacity, 5,660

acre-f t

since 1963 and

Steamboat L

ake ,

capacity ,

23,060 acre-ft since 1968.

Several

observations of

specific conduc

tance and water ten

perature were

obtained

and are p

ublished

elsewhere

in

this

repor

t.

AVERAGE

DISCHARGE.

--67 years, 339

ft3 /s;

245,

600 acre

-ft/yr

EXTREMES

FOR PERIOD OF RECORD. --Ma

ximum c

isch

arge, 4,

910 ft

'/s, May

>3, 1984,

gage hei

ght, 6.12

ft; minimum

daily

determine

d, 22 ft

/s, Dec

. 12, 1963,

but a 1

esser

discharge

may have

occurred

during p

eriods of

no

gage-

height record prior to 1939-

EXTREMES

FOR CURRENT YEAR.

— Peak d

ischarg

es e

ibove base of 1

,900 ft3/

5, and maximum (*]

D

ischarge

Gag

e height

Discharge

Gage

height

Date

Time

(ft3/s)

(ft)

Date

Time

(ft3/s)

(ft)

May 10

1900

2,000

4.44

June 8

2300

*3,

180

»5.

36

May 28

2200

2,300

4.68

Minimum daily d

ischarge

, 40 ft3

/s, Feb

. 2C

to Mar.

2.

DISCHARGE,

IN CUBIC FEET

PER

SECOND,

WATER

YEAR OCTOBER 1984

TO SEPTEMBER 1985

MEAI\

VALUES

DAY

OCT

NOV

DEC

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

1

121

104

72

100

60

40

226

1030

1200

527

226

72

2

126

98

66

100

60

40

226

1210

1180

501

221

75

3

118

107

70

98

60

42

225

1380

1220

491

211

97

14

115

95

72

100

60

45

223

1500

1290

483

194

86

5

124

92

74

100

60

47

223

1630

1330

463

185

78

6

131

104

74

98

60

50

223

1500

1350

437

179

72

7

123

97

76

98

58

56

223

11)10

1710

398

170

70

8

113

97

78

98

56

60

224

1580

2280

381

161

75

9

108

93

80

96

54

64

262

1640

2410

377

167

71

10

104

104

80

96

50

66

258

1720

2080

343

156

68

11

106

109

80

96

45

68

258

1620

1630

329

149

67

12

104

100

82

96

45

68

329

1300

1410

358

181

86

13

108

102

82

96

42

70

41 1

1120

1440

343

1it8

72

14

111

105

84

96

42

70

500

1000

1470

298

135

67

15

100

92

86

94

42

70

675

1030

1530

266

126

64

16

97

100

88

94

42

70

817

1 110

1570

251

119

65

17

96

106

90

94

42

70

89 3

1 160

1520

233

111

62

18

1 1 1

87

92

94

42

72

1090

1 150

1370

390

1 12

62

19

110

89

92

94

42

74

1240

1140

1280

715

115

73

20

1 12

90

92

90

40

76

1 100

1 170

1220

560

107

68

21

111

90

90

86

40

78

1060

1280

1210

409

101

70

22

103

80

88

84

40

80

987

1310

1090

374

99

73

23

108

80

86

82

40

80

964

1340

968

364

94

77

24

108

80

88

82

40

80

953

1500

892

408

89

72

25

97

80

94

80

40

82

789

1700

1120

322

85

71

26

105

80

94

74

40

84

574

i860

1000

302

83

71

27

106

80

92

70

40

86

569

1930

738

288

82

79

28

103

78

92

66

40

88

708

2000

631

251

83

92

29

108

76

94

62

90

832

1760

583

301

81

84

30

105

74

96

60

100

890

1590

565

273

77

68

31

106

—

98

60

—

105

---

1320

248

73

TOTAL

3398

2769

2622

2734

1322

2171

17952

43990

39287

11684

4123

2207

MEAN

110

92.3

84.6

88.2

47.2

70.0

598

1419

1310

377

133

73.6

MAX

131

109

98

100

60

105

1240

2000

2410

715

226

97

MIN

96

74

66

60

40

40

223

1000

565

233

73

62

AC-FT

6740

5490

5200

5420

2620

4310

35610

87250

77930

23180

8180

4380

CAL YR 1984 TOTAL

183267

MEAN

501

MAX

4090

MIN

53

AC-FT 363500

WTR YR 1985 TOTAL

134259

MEAN

368

MAX

2410

MIN

40

AC-FT 266300

Figure 9- An annual summary of dailv streamflow data collected and published by the U.S. Geological Survey
(1985).

624

Hydrologic Properties

Oct Nov Dec J»n Feb Mar Apr May Jun Jul Aug Sep

Figure 10. An example of an annual hydrograph
based on mean monthly streamfiow data, Elk
River at Clark, CO, Station 09241000 (U.S. Geo-
logical Survey 1985).

r^flH &

*

' :*

-"*-

*V

«3J0^

! 3.20 —

fir

i i

*'<'

i ...

3.10^—

3.00-^
2.90-^-

1

n

2.80 —
2.70^-

it?

pi

** ' ZKi

**N|r

2.60 ~ 9

jMPm*

*>■' "

m

■■

.2.50 ^-M

0 At\ " • ■

"'Jr 1

'.<:'. '-£L--&st&&!?

.

1

2.30- —
2.20^—

?.10^-
.'.00-^-

1

ll

■<■/.*■

I

1.90 -^ 1
1.80-^

^

#

•

~^»J*3>.

pi /

,1.70 ^-J
1.60-^
1.50^ — 1

1.40 -^rl

li

Figure 12. A staff gage for measuring water
level.

Elliptic depression

Cone frustrum

Volume : 3 (B^ B2+\JB,B2 )

where B, = iir,2 , B, = iir,2

Volume = 9 n abh

Sphere segment

r T=VJ

Volume = 6 (3a2+ 3b2+hJ)

Elliptic sinusoid

T"

Volume = 4 abD(l-fr)

Figure 11. Common lake basin shapes and their corresponding volume formulas.

Hydrologic Properties

625

Table 1. Streamflow unit conversions.

To convert from

Multiply by

To find

acre-ft/day

0.504167

cfs

acre-ft/day

0.123349

ha-m/day

acre-ft/day

0.325851

million gal./day

acre-ft/mo (28-day)

0.018006

cfs

acre-ft/mo (29-day)

0.017385

cfs

acre-ft/mo (30-day)

0.016806

cfs ^^^^^

acre-ft/mo (31 -day)

0.016263

cfs

acre-ft/yr
acre-in./hr

0.001370
2

acre-in./hr
acre-ft/day

acre-in./hr

1 .008333

cfs

acre-in./hr

1 .027906

ha-cm/hr

cfs

1 .98347

acre-ft/day

cfs

55.53719

acre-ft/mo (28-day)

cfs

57.52066

acre-ft/mo (29-day)

cfs

59.50413

acre-ft/mo (30-day)

cfs

61 .48760

acre-ft/mo (31 -day)

cfs

723.96694

acre-ft/yr

cfs

725.95041

acre-ft/leap year

cfs

0.99174

acre-in./hr

cfs

60.45620/area (a.)

cm/day

cfs
cfs

2.51 901 /area (a.)

cm/hr

1.69901

m3/min

cfs

0.02832

m3/sec

cfs

0.00021424

m'3/yr „«______«__«__.

cfs
cfs

6.66667
448.83117

yd3/min
gal./min

cfs

538168.7

Imp. gal./day

cfs

373.73

Imp. gal./min

cfs

23.801 65/area (a.)

in./day

cfs

0.99174/area(a.)

in./hr

cfs

28.31685

L/sec

cfs

0.64632

million gal./day

cfs

50

miner's inches (ID, KS, NE, SD,
ND, NM, UT, WA, S. CA)

cfs

40

miner's inches (AZ, MT, OR, NV,
N. CA)

cfs

38

miner's inches (CO)

cfs

2696.082

tons of water (50°F)/day
cm/day

m3/sec

864/area (ha)

m3/sec

315360/area (ha)

cm/yr

m3/sec

35.31467

ft3/sec (cfs)

m3/sec

8.640

ha-m/day

m3/sec

3153.60

ha-m/yr

m3/sec

3162.24

ha-m/leap yr ^^^^^^

m3/sec

86400010

Uday

m3/sec

60000

Umin

m3/sec
cm/day

1000.00012

Usee

0.01654 x area (a.)

cfs

cm/day

0.00116 x area (ha)

m3/sec

cm/hr

0.39698 x area (a.)

cfs

cm/yr

3.171E-06 x area (ha)

m3/sec

626

Hydrologic Properties

Table 1. Streamflow unit conversions (concluded).

To convert from

Multiply by

To find

gal./day
gal./min

0.00112
0.00442
1.61301

acre-ft/yr

acre-ft/day

acre-ft/yr

gal./min
gal./min
gal./min

0.00223

0.00006309

0.06309

cfs

m3/sec

L/sec

Imp. gal./min
million gal./day
million gal./day

0.00268
3.06888
1.54723

cfs
acre-ft/day

ha-m/day
ha-m/day
ha-m/day

8.10710

0.11574

115.74075

acre-ft/day

m3/sec

L/sec

ha-m/yr
ha-m/yr
ha-m/yr

8.10710

3.17E-04

0.31710

acre-ft/yr
m3/sec

in./day

in./hr

in./yr

0.04201 x area (a.)
1 .00833 x area (a.)
0.000115 x area (a.)

cfs
cfs
cfs

L/sec
L/sec
L/sec

0.864/area (ha)
0.036/area (ha)
315.36/area (ha)

cm/day

cm/hr

cm/yr

Usee
L/sec
L/sec

0.03531
0.001
15.85032

cfs

m3/sec

gal./min.

miner's inches (ID, KS, NE, SD,
ND, NM, UT, WA, S. CA)
miner's inches (AZ, MT, OR, NV,
N. CA)

0.020
0.025

cfs
cfs

miner's inches (CO)

acre-ft

acre-ft/mi.2

0.026
504.17
0.04763

cfs

Kft3/sec-days

cm over watershed

acre-ft/mi.2
acre-ft/mi.2
cm over watershed

0.04763

0.01875

20.99738

ha-m/km2

in. over watershed

acre-ft/mi2

cfs/mi2
cfs/mi2
cfs/mi2

0.01093
1.04132
1.07851

m3/sec/km2
in./mo (28 days)
in./mo (29 days)

cfs/mi2
cfs/mi2
cfs/mi2

1.11570

1.15289

13.57438

in./mo (30 days)
in./mo (31 days)
in./yr (365 days)

cfs/mi2
cfs/mi2
cfs/mi2

13.61157

0.03719

10.93320

in./yr (366 days)

in./day

L/sec/km2

m3/sec/km2

ha-m/km2

in. over watershed

91 .46457
20.99729
53.33333

ft3/sec/mi2
acre-ft/mi2
acre-ft/mi2

in. over watershed
Kft3/sec-days/mi2
Kft3/sec-days
L/sec/km2

0.02689

37.19008

0.00198

0.09146

Kft3/sec-days/mi2
in. over watershed
acre-ft
ft3/sec/mi2

Hydrologic Properties

627

GROUNDWATER FEATURES AND THEIR
RELATIONSHIP TO WILDLIFE HABITAT

I have taken the liberty in this section to discuss
springs, seeps, and related water features in an over-
all groundwater context, even though many classify
springs as surface water features. Springs and seeps
are usually an expression of the local groundwater
system.

Wildlife biologists may be concerned about the
reliability of wildlife water supplies from springs and
seep areas. Spatial distribution as well as discharge
of individual springs or seeps would be of interest
for inventory purposes. Measurement of spring dis-
charge is done by the same basic techniques covered
in "Streamflow Measurement and Characteristics"
and will depend on size and physical setting of the
spring.

Groundwater levels may be of interest to the
wildlife biologist, particularly where wildlife water
supplies are obtained from pumping wells or artesian
wells. The Water Resources Division of the U.S. Geo-
logical Survey operates a network of observation
wells around the U.S. That agency should be con-
tacted for data on water levels and water quality.
Nearly all completed water wells may be used for
measuring water levels. However, the well log
should be consulted to determine the depth of the
well, whether or not packers were included during
well construction, and any other construction
details.

STREAM CHANNEL STABILITY AND
SEDIMENT TRANSPORT

A very good overview of stream dynamics, from
a hydrologic and geomorphic point of view, has
been prepared by Dr. Burchard Heede of the U.S.
Forest Service (Heede 1980). This is highly recom-
mended reading for the biologist planning to engage
in stream channel stability or sediment transport
investigations, or riparian improvement projects.

Stream Channel Stability

The channel stability evaluation procedure used
most often by land management agencies is that
presented by Pfankuch (1978). An evaluation form
for this method is presented in Figure 1 3. Field train-
ing is required before using this procedure because
of its subjectivity. Furthermore, the summary numer-
ical rating system may need adjusting to fit the gen-
eral region of interest.

The interpretation of the results depends some-
what on the use of the ratings; it is not an aquatic
habitat evaluation. It can be used as an index to sedi-
ment delivery because poor stability usually indi-
cates high levels of sediment transport and good to
excellent stability correlates with low sediment
transport.

Standard statistical sampling guidelines should
be used to determine sample stratification and sam-
ple size (Platts et al. 1983).

Suspended Sediment and Bedload

Sediment transport and deposition in streams
directly influence riparian systems. For example,
aquatic habitat condition deteriorates with increasing
suspended sediment.

Field sampling of suspended sediment is a rela-
tively easy exercise in wadable streams. Williams and
Thomas (1984) discuss the collection and analysis of
sediment data from a land management perspective.
Grab samples should be taken with a DH-48 sampler
(Figure 14). Streamflow must be measured at the
time of sediment sampling to facilitate meaningful
interpretation of these data.

Continuous sediment sampling is both difficult
and expensive and should be left to experienced
hydrologic technicians.

Total sediment in a stream is composed of both
suspended and bedload fractions. Accurate bedload
sampling is a difficult procedure. Fortunately, the
wildlife biologist seldom has need for such data. Guy
and Norman (1970) discuss bedload sampling
procedures.

MONITORING HYDROLOGIC VARIABLES

Water in nature is highly transient in time and
space. Monitoring of hydrologic variables thus re-
quires some special considerations. Ponce (1980a,
1980b) and Jackson et al. (1985) provide excellent
guidance for hydrologic monitoring in land manage-
ment situations. Those references should be con-
sulted prior to engaging in any monitoring involving
water quantity or quality.

628

Hydrologic Properties

B-l STREAM REACH INVENTORY and CHANNEL STABILITY EVALUATION
BEACH LOCATION i Survey Date Time Obs.

Forest

Streaa

Rgr. Dist.
"P. W.I.
W/S No. _■

Reabh Description 4
Other Identification

Key j Stability Indicators by Classes (Fair and Poor on reverse BJdTJ

EXCELLENT

Bank slope gradient < 30%-

No evidence of past or any
potential for future aass
wasting into channel.

Essentially absent froa
imaediate channel area.

10

9056+ plant density. Vigor
and variety suggests a
deep, dense, soil binding,
root mass .

Anple for present plus soae
increases. Peak flows con-
ta ined . . W/b ratio <7.

6^36+ with large, angular
boulders 12"+ numerous.

Rocks and old logs firmly
embedded. Flow pattern with
out cutting or deposition.
Pools and riffles stable.

HI

(3)

(2)

(3;

(1)

(2)

Little or none evident.
Infrequent raw banks less
than 6" high generally. _

Little or no enlargement
of channel or point bars.

Sharp edges and corners,
plane surfaces roughened.

, , Surfaces dull, darkened, or
stained. Gen, not "bright'*.

12

13

Assorted Bizes tightly
packed and/or overlapping.

No change in sizes evident,

Stable materials 80-100%,

14

Less than 5% of the bottoa
affected by scouring and
deposi tion.

Abundant . Growth largely
15 moss-like, dark green, per-
I ennial . In swift water too.

(2)

CO

CO

(1)

(1)

(2)

CO

(6)

(1)

EXCELLENT COLUMN TOTAL •*■[ [ GOOD COLUMN TOTAL
Add values in each column* and record in spaces below. Add column scores,

GO0P

Bank slope gradient 30-40%.

Infrequent and/or very small,
Mostly healed over. Low
future potential.

Present but ssostly small
twigs and liabs.

70-90% density. Fewer plant
species or lower vigor
suggests a less dense or
deep root mass,

Adequate. Overbank flows
rare. Width to Depth (w/PJ
ratio 8 to 15.

40 to 65%, mostly small
boulders to cobbles 6-12".

Some present, causing erosive
cross currents and minor pool
filling. Obstructions and
deflectors newer and less
fim.

Some, intermittently at
outcurves and constrictions.
Raw bankB «tay be up to 12".

Soae new increase in bar
formation, mostly from
coarse gravels.

Rounded corners and edges,
surfaces smooth and flat.

Mostly dull, but may nave up
to 35% bright surfaces.

Moderately packed with
some ove r 1 apping.

Distrioution shift slight.
Stable materials 50-60%.

5-30^ affected. Scour at
constrictions and wnere
grades steepen. Some
deposition in pools.

Comnon, Algal forms in low
velocity & pool areas. Moss
here too and swifter waters.

HE

(6;

CO

(6)

(2)

(4)

CO

(8)

(8;

(2)

(2)

CO

(8j

(12;

(2;

+ F,

+ P,

Total Reach Score
Adjective ratings :<38=Excellent , 39-76=Good, 77-114-=Fair, 11

■Poor*

•(Scores above may be locally adjusted Dy Forest Hydrologist)

Rl-Form 25OO-5A Rev. 1-75 Side 1

Figure 13. Stream channel stability evaluation form used with the evaluation procedure
presented by Pfankuch (1978).

Hydrologic Properties

629

INVENTORY DATA: (observed or measured on this date)

Side 2

Stream Width ft.X Ave. Depth ft.X Ave. Velocity f/s= Flow cfs

Reach Stream rurbidity Stream Sinuosity

Gradient %, Order , Level , Stage , Ratio ,

Temperature A^r

°F or C of: Water , Others

Key

10

11

12

13

14

15

Stability Indicators by Glasses

FAIR

Bank slope gradient 40-60%.

Moderate frequency & size,
with some raw spots eroded
by water during; high flows

Present, volume and size
are both increasing

30-7 OJb density . Lower vigor
and still fewer species
form a somewhat shallow and
discontinuous root mass.

Barely contains present
peaks. Occassional overbank
floods. W/P ratio 15 to 25.

5. V/l

20 to 40%, with most in
the 3-6" diameter class.

Moderately frequent, moder-
ately unstable obstructions
& deflectors move with high
water causing bank cutting
and filling of pools.

Significant. Guts 12"-2^"
high. Root mat overhangs
and sloughing evident,

Moderate deposition of new
gravel & coarse sand on
old and some new bars.

Corners & edges well round'
ed in two dimensions.

Mixture, 50-50% dull and
bright. ±15% ie. 35-65%.

Mostly a loose assortment
with no apparent overlap.

Moderate change in sizes.
Stable materials 20-50%.

30-50% affected. Deposits
4 scour at obstructions,
constrictions, and bends.
Some filling of pools.

Present but spotty, mostly
in backwater areas. Season-
al blooms make rocks slick.

FAD* COLUMN TOTAL'

m

(9)

(6)

(9)

(3)

[6)

(6j

(12;

(12)

(3)

(3)

(6)

(12)

(18)

(3.

POOR

Bank slope gradient 6oJfr

+ .

Frequent or large, causing
sediment nearly yearlong OR
imminent danger of same.

Moderate to heavy amounts,
predominantly larger sizes,

<50% density plus fewer
species & less vigor indi-
cate poor, discontinuous,
and shallow root mass.

Inadequate. Overbank flows
common. W/D ratio > 25.

< 20% rock fragments of
gravel sizes, 1-3" or less.

Frequent obstructions and
deflectors cause bank ero-
sion yearlong. Sediment
traps full, channel
migration occurring.

Almost continuous cuts,
some over 24" high. Fail-
ure of overhangs frequent.

Extensive deposits of pre-
dominantly fine particles.
Accelerated bar development.

Well rounded in all dimen-
sions. surfaces smooth.

Predominantely bright, 65%+,
exposed or scourea surfaces.

No packing evident. Loose
assortment, easily moved.

Marked distribution change,
Stable materials 0-20%.

More than 50% of the bottom
in a state of flux or
change nearly yearlong.

Perennial types scares or
absent. Yellow-green, short
term bloom may be present.

HI

(12;

(8)

(12)

CO

(8)

(6)

(16)

(16;

CO

CO

(8)

:i6j

(24;

C+;

POOR COLUMN TOTAL-

1.
2.

3.
4.

Size Composition of Bottom Materials (Total to 100%)

Exposed bedrock,
Large boulders, 3'+ Dia,
Small boulders, 1-3'...,
Large rubble , 6" -12"

Small rubble, 3"-6

6. Coarse gravel, l"-3"..

7. Fine gravel, 0.1-1"
6. Sand, silt, clay, muck

Figure 13- Stream channel stability evaluation form used with the evaluation procedure
presented by Pfankuch (1978) (concluded).

630

Hydrologic Properties

Figure 14. Model DH-48 suspended-sediment sampler.

Hydrologic Properties

631

LITERATURE CITED

BRAKENSIEK, D.L., H.B. OSBORN, and WJ. RAWLS,
coords. 1979- Field manual for research in agricul-
tural hydrology. Agriculture Handbook 224. U.S. Dep.
Agric. 550pp.

BUCHANAN, T.J. and W.P. SOMERS. 1969. Discharge
measurements at gaging stations. Techniques of
Water-Resources Investigations, Book 3, Chapter A8.
U.S. Dep. Inter., Geol. Surv. 65pp.

CUPLIN, P. and B.P. VAN HAVEREN, eds. 1979. Instream
flow guidelines. U.S. Dep. Inter., Bur. Land Manage.
57pp.

GUY, HP. and V.W. NORMAN. 1970. Field methods for
measurement of fluvial sediment. Techniques of
Water-Resources Investigations, Book 3, Chapter C2.
U.S. Dep. Inter., Geol. Surv. 59pp.

HEDMAN, E.R. and W.R. OSTERKAMP. 1982. Streamflow
characteristics related to channel geometry of streams
in western United States. U.S. Dep. Inter., Geol. Surv.
Water-Supply Pap. 2193. 17pp.

HEEDE, B.H. 1974. Velocity-head rod and current meter
use in boulder-strewn mountain streams. U.S. Dep.
Agric, For. Serv. Res. Note RM-271. 4pp.

. 1980. Stream dynamics: An overview for land

managers. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
RM-72. 26pp.

JACKSON, W.L., S. HUDSON, and K. GEBHARDT. 1985.
Consideration in rangeland watershed monitoring.
U.S. Dep. Inter., Bur. Land Manage. Tech. Note 369.
Denver, CO. 25pp.

PFANKUCH, DJ. 1978. Stream reach inventory and chan-
nel stability evaluation. U.S. Dep. Agric, For. Serv.,
Northern Region, Missoula, MT. 26pp.

PLATTS, W.S., W.F. MEGAHAN, and G.W. MINSHALL
1983 Methods for evaluating stream, riparian, and
biotic conditions. U.S. Dep. Agric, For. Serv. Gen.
Tech. Rep. INT- 138. 70pp.

PONCE, S.L. 1980a. Statistical methods commonly used in
water quality data analysis. U.S. Dep. Agric, For. Serv.,
Watershed Systems Development Group, Tech. Pap.
WSDG-TP-001. 144pp.

. 1980b. Water quality monitoring programs. U.S.

Dep. Agric, For. Serv., Watershed Systems Develop-
ment Group, Tech. Pap. WSDG-TP-002. 68pp.

RANTZ, S.E. 1982. Measurement and computation of

streamflow: Volumes 1 and 2. U.S. Dep. Inter., Geol.
Surv. Water-Supply Pap. 2175. 631pp.

U.S. GEOLOGICAL SURVEY. 1985. Water resources data
for Colorado, Part I — Colorado River Basin. U.S. Dep.
Inter., Geol. Surv.

VAN HAVEREN, B.P. 1986. Water resource measurements:
A handbook for hydrologists and engineers. American
Water Works Assoc, Denver, CO. 132pp.

WILLIAMS, O. and R.B. THOMAS. 1984. Guidelines for

collection and analysis of sediment data. Draft report,
U.S. Dep. Agric, For. Serv., Watershed Systems Devel-
opment Group, Fort Collins, CO. 108pp.

WILM, H.G. and H.C. STOREY. 1944. Velocity-head rod
calibrated for measuring stream flow. Civil Engineer-
ing. l4(ll):475-476.

WLNTER, T.C. 1977. Classification of the hydrologic set-
tings of lakes in the north-central United States. Water
Res. 13(4):753-767.

632

Hydrologic Properties

30

WATER
QUALITY

Paul Cuplin

U.S. Bureau of land Management
Service Center
Denver, CO 80225

Editor's Note: Water quality, the purity of water, is
a concern from many perspectives — public health,
aesthetics, agriculture, and as habitat for aquatic
life. This chapter focuses on water quality require-
ments for fishes and other aquatic vertebrates, and
field methods to measure or monitor these
requirements.

INTRODUCTION

Water availability is critical to almost every re-
newable natural resource including wildlife and veg-
etation. Therefore, information on all of the water
that occurs on desert, forest, stream, lake, and wet-
land on public lands is of importance to land man-
agement. Water found on public lands represents all
of the major water systems, i.e., marine, estuarine,
riverine, lacustrine, palustrine, perennial ice or snow,
and terrestrial systems. The chemical and. physical
properties of water, as they affect "water" as a habi-
tat for aquatic organisms and as a substance for con-
sumption by wildlife and domestic livestock, are
the concerns of , this chapter.

LAND-USE ACTIVITIES THAT AFFECT
WATER QUALITY

Human activities can alter baseline water chem-
istry, resulting in detrimental effects on aquatic and
terrestrial organisms that drink or dwell in water.
Land-use activities such as grazing, mining, timber
harvest, and recreation may increase sediment pro-
duction and therefore affect water quality. Mining
can also produce "acid mine waste," making the pH
of water too acidic for any organism to survive. Her-
bicide or pesticide spraying or the dumping of toxic
materials can significantly alter water quality and
cause high mortality rates in sensitive aquatic life. To
protect aquatic habitats and water quality, the U.S.
Environmental Protection Agency (EPA) established
standards for the introduction of toxic materials into
water, which were published in "Quality Criteria
for Water" (U.S. Environmental Protection Agency
1976).

The Water Quality Act of 1965 required States
to set standards for releasing materials into streams
and lakes. The Federal Water Pollution Control Act
and Amendments of 1972 (Public Law 92-500; 86
Stat. 816), as amended by the Clean Water Act of
1977 (Public Law 95-217; 91 Stat. 1566), are also perti-
nent. See 33 U.S.C., Chapter 23:

[Section 208 directs that the Administrator of
the Environmental Protection Agency shall
promulgate comprehensive regulations in-
tended to control water pollution sources, in-
cluding nonpoint sources. Under the
regulations which have been promulgated, the

Water Quality

633

use of permits is required for any continuing
pollution sources. The Federal land manage-
ment practices are designed to prevent degra-
dation of streams and lakes on the lands
under their jurisdiction. ]

(author's emphasis)
[Section 404 authorizes the Army Corps of
Engineers to issue permits for activities which
result in earth disturbances contributing to
stream sedimentation or which will result in
decreased stream flows because of water
impoundment.]

UNIQUE PROPERTIES OF WATER

Water, a universal solvent, dissolves many sub-
stances, resulting in the alteration of water quality.
Water is the only known substance which allows the
solid state (ice) to float on the liquid state. If this
process were not possible, lakes would freeze from
the bottom up, eliminating aquatic life annually.
Some deep lakes would remain frozen forever. The
greatest water density is reached at 4°C (39.6°F),
allowing aquatic life to survive in an ice-covered lake
unless light penetration is prevented for long pe-
riods. Then the available oxygen may be completely
consumed, resulting in a "winter kill" of fish.

WATER QUALITY DATA SOURCES

Many sources provide water quality information.
The U.S. Geological Survey (USGS) annually pub-
lishes surface water reports, by State, that contain
streamflow and chemical analysis data for major
streams where sampling stations have been estab-
lished. The U.S. Environmental Protection Agency
maintains two computerized data bases: (STORET)
for chemical water information and (BIOSTORET)
for biological information on aquatic life. Data are
stored geographically by latitude and longitude.

State water quality control commissions and
State fish and wildlife agencies are also sources of
chemical and biological information on water
quality.

WATER SAMPLING

If animals become distressed from being in or
drinking from polluted waters on public lands, then
water sampling and analysis become urgent matters
in determining the cause and solution to the prob-
lem. Reduced oxygen or introduced toxic materials
are common causes. Dissolved oxygen (DO) can
be readily analyzed in the field with a field DO kit. If
the dissolved oxygen is 5 mg/1 or higher, the prob-
lem is more likely to be a toxic material problem
than a lack of oxygen.

At least a 1 -gallon water sample is collected and
carefully labeled as to time, date, location, and site
conditions. Water chemical analysis is then done by
a qualified chemist. The chemist must be informed of
the most likely substance causing the problem, as
the chemist cannot conduct a thorough analysis on
all elements which might be present in the water
sample.

Preferably, water quality should be monitored
on site since many parameters, e.g. temperature, DO,
and pH, will be altered during transportation to the
laboratory. If monitoring is done in the field, samples
need only be collected for the more detailed anal-
yses such as trace metals.

Beyond the crisis situation described above,
what chemical and physical characteristics of water
are useful for land management? The list is short: pH,
DO, temperature, alkalinity, conductivity, and nu-
trients. These variables aid in describing habitat con-
ditions suitable for specific aquatic organisms. In
addition to these variables, others may be of interest
or importance. The most important are listed with
references where more detailed information can be
found.

Water samples to be analyzed by a chemist (or
placed in storage ) should contain identifying infor-
mation such as sample or station number; name of
water, stream, lake or reservoir; time and date of
collection; collector's name and agency; and a report
about the chemical analysis done at the collection
site. Because storage of water samples in certain
containers for extended periods may alter water
chemistry, inquiries as to sampling procedures and
containers should be made from local and regional
water quality analysts or by consulting "Standard
Methods for the Examination of Water and Wastewa-
ter" (American Public Health Association 1976) and
"Methods for Chemical Analysis of Water and
Wastes" (U.S. Environmental Protection Agency
1979). The Federal Register, Vol. 44, No. 244, pp.
75028-75052, December 18, 1979, describes ap-
proved test procedures, containers, and preservation
and holding times.

Sampling Procedures

Water sampling analysis describes the water
qualities for a particular location on a special date, at
a designated time. This is sometimes called a "grab
sample" which may be more meaningful if the same
site is sampled over a period of time.

Water can be analyzed with a field chemical kit
at the sampling site or transported to a chemical
laboratory for more detailed analyses. The field
chemical kit can be used to analyze temperature,
turbidity, pH, hardness, alkalinity, sulfate, ammonia,
phosphate, nitrate, some metals, and dissolved gases.

634

Water Quality

In fact, dissolved gases, such as oxygen, carbon diox-
ide, and hydrogen sulfide, must be analyzed at the
site. (Special care is required to avoid altering the
levels of dissolved gases by aeration. ) All measure-
ments should be recorded carefully in a field note-
book or on appropriate forms.

Water Quality Criteria

Water quality criteria specify concentrations of
water constituents for various uses, e.g., fish, wildlife,
recreation, domestic water sources, and agriculture
(irrigation and livestock). Typically, fish and wildlife
criteria are the most sensitive. Criteria, derived from
scientific research and observation, are established
by each State and may vary regionally as well as
among States.

Water quality criteria are listed alphabetically in
"Quality Criteria for Water" (U.S. Environmental
Protection Agency 1976 ) with comments by the
American Fisheries Society (Thurston et al. 1979).
The issue of water quality criteria is unsettled. In
most cases, establishing water quality criteria is not
desirable or feasible at the local level. The State
water quality criteria should therefore be used
where applicable. Additional guidance can be ob-
tained from the literature.

Water Measurement Terms

Some common terms used in water quality anal-
ysis are described below.

mg/1 — milligrams of a chemical per liter of water; a
way to express the concentration of chemi-
cals; equivalent to parts per million (ppm) or
micrograms per gram (u-g/g) or milligrams
per kilogram (mg/kg).

u.g/1 = micrograms of chemical per liter of water;

equivalent to parts per billion (ppb) or nano-
grams per gram (ng/g) or micrograms per
kilogram ( u-g/kg).

LC = lethal concentration; usually expressed as
a concentration of a toxicant that is lethal
( fatal ) to a fixed percentage of the organisms
in a specified time. Thus a 96-hour LCs„ will
be lethal to 50% of the organisms within
96 hours.

Water quality criteria are generally listed as
dissolved (d) or total (t) depending on the applica-
tion. This is particularly important for metals where
the dissolved (ionic) forms can be quite toxic,
whereas the precipitated, bound forms are relatively
nontoxic.

Quality Criteria for Selected Substances

The following are criteria and wildlife and fish-
ery considerations for some of the more important
substances, considering that water quality changes
among sites, seasons, and even days. "EPA criteria"
refer to criteria listed in "Quality Criteria for Water"
(U.S. Environmental Protection Agency 1976).

Alkalinity. Alkalinity is the sum total of the
components in the water that tend to elevate the pH
of water above a value of about 4.5. It is measured
by titration with standardized acid to a pH value
of about 4.5 and expressed as mg/1 of calcium
carbonate. Alkalinity is therefore a measure of the
buffering capacity of water. The EPA criteria for
alkalinity is 20 mg/1 or more as CaCO^ for fresh-
water aquatic life except where natural concen-
trations are less. Thurston et al. (1979:3) say the EPA
criteria of 20 mg/1 are not valid. Controversy exists
in all stated criteria because of the great variability
in regional water quality and the aquatic species that
inhabit the waters.

Cadmium. TTie EPA criteria for cadmium are —

Soft Water
0.4 u.g/1 for cladocerans and salmonid fishes
4.0 (jig/1 for other, less sensitive aquatic life

Hard Water
2.0 u.g/1 for cladocerans and salmonid fishes
12.0 u.g/1 for other, less sensitive aquatic life

Thurston et al. (1979:51) indicate that "Toxicological
data would suggest criteria for five levels of hard-
ness: 0-35, 35-75, 75-150, 150-300, and 300 mg/1
CaC03."

Chlorine. The EPA criteria for total residual
chlorine are —

2.0 u.g/1 for salmonid fish
10.0 |xg/l for other fresh-water and marine organisms

Thurston et al. (1979:67-72) feel these criteria are
not acceptable and state, "A single criterion of 3-
5 mg/1 (measured by amperometric titration in con-
junction with polarograph) is more appropriate for
the protection of fresh-water organisms."

Chromium. The EPA criterion is 100 |JLg/l for fresh-
water aquatic organisms.

Thurston et al. (1979) recommend that in no case
should the criterion be higher than 50 u.g/1 for any
aquatic organisms.

Conductivity. Conductivity is a numerical
expression of the ability of an aqueous solution to
carry electric current, expressed in u.mhos/cm

Water Quality

635

(micromhos/cm) at 25° C. Conductivity can be used
to approximate total dissolved solids in water by
using the formula: Specific conductance x (0.65 ±
0.05) = mg/1 dissolved solids. The formula should
be verified by comparing specific conductance in
u,mhos with the total dissolved solids determined
empirically by filter-evaporation procedures.

Copper. The EPA criteria for fresh-water and
marine aquatic life, 0.1 X a 96-hr LCs„ value, is
determined through nonaerated bioassay using
sensitive aquatic (resident) species.

Thurston et al. (1979:100) do not agree with this
criterion (which is 10 times higher — 0.01 to 0.1 —
than the previous criterion); however, they do not
suggest another criterion.

Cyanide. The EPA criterion is 5 u.g/1 for fresh-water
and marine aquatic life and wildlife.

Thurston et al. (1979:100) suggest that —

"The criterion for cyanide, presumably an up-
per limit of total cyanide concentration at any
time or place, is unacceptable. A limit of the
concentration of free cyanide ( HCN and CN_ )
or of molecular hydrogen cyanide (HCN), with
the same numerical value ( 5 u.g/1 ) can be ac-
cepted as a reasonable water quality criterion
for general application with some reservations.
However, the . . . criterion evidently is a limit
for total cyanide . . . Such a criterion can leave
no sound toxicological basis, because only free
and not complexed cyanide has been shown
to be toxic at concentrations less than 1 mg/1
(1,000 u.g/1). Moreover, very low levels of mo-
lecular HCN can now be reliably determined
by any of several proven methods . . ."

Hardness. Hardness is reported as an equivalent
concentration of calcium carbonate (CaCOs).

Concentration of CaCO,
(mg/1)

Description

0-75
75-150

soft water
moderately hard water

150-300
300 and up

hard water
very hard water

The EPA (1976) suggests that the effects of hardness
on fresh-water fish and other aquatic life are related
to the ions causing the hardness rather than hardness
itself.

Iron. The EPA criterion is 1 mg/1 for fresh-water
aquatic life.

636 Water Quality

Thurston et al. (1979) suggest the criterion for iron
is too high. Lethality for aquatic insects has been
observed at 320 |xg/l.

Lead. The EPA criterion is 0.01 times the 96-hour
LC50 value, using the receiving or comparable water
as the diluent and soluble lead measurement (non-
filterable lead using 0.45 micron filter), for sensitive
fresh-water resident fish.

Mercury. The EPA criterion is 0.05 u.g/1 for fresh-
water aquatic and marine aquatic organisms.

Nickel. The EPA criterion is 0.01 of the 96-hour
LC50 value, for fresh-water marine aquatic organisms.

Nitrates and Nitrites. The nitrogen compounds in
natural waters are derived from the fixation of
atmospheric nitrogen or from pollution sources.
These compounds are available for absorption by
bacteria, which produce ammonia, nitrite, and
nitrate. Inorganic forms of nitrogen are rapidly
absorbed and concentrated by phytoplankton. Little
plant growth occurs if nitrate nitrogen is below 0.3
mg/1.

The EPA (1976) states —

"It is concluded that: (1) levels of nitrate nitro-
gen at or below 90 mg/1 would have no ad-
verse effects on warm-water fish ... (2) nitrite
nitrogen at or below 5 mg/1 should be protec-
tive of most warm-water fish . . . and ( 3 ) nitrite
nitrogen at or below 0.06 mg/1 should be pro-
tective of salmonid fishes . . . These levels either
are not known to occur or would be likely to
occur in natural surface waters. Recognizing
that concentrations of nitrate or nitrite that
would exhibit toxic effects on warm or cold
water fish could rarely occur in nature, restric-
tive criteria are not recommended."

Oil and Grease. The EPA (1976) suggests, for
aquatic life, that levels of individual petrochemicals
in the water should not exceed 0.01 of the lowest
continuous flow 96-hour LC50 value, since each
species has demonstrated high susceptibility to oils
and petrochemicals.

Oxygen, Dissolved. The EPA criteria for dissolved
oxygen (DO) for fresh-water aquatic life is 5 mg/1.
The amount of oxygen that will dissolve in water is
affected by temperature, elevation, and total
dissolved solids, although certain desert fishes such
as pupfish can survive low oxygen tension, e.g., 0.1 to
0.4 mg/1.

pH. The abbreviation "pH" is a measure of the
hydrogen-ion activity in a water sample. It is
mathematically related to hydrogen-ion activity
according to the expression: pH = - logi()[H + ],

where [H + ] is the hydrogen-ion activity. Stream
water in areas not influenced by pollution generally
has a pH range between 6.5 and 8.5, which is
acceptable for fish. The EPA criterion is 6.5 to 90
for fresh-water aquatic life.

Acid rain, the product of fossil fuel combustion,
produces sulfur dioxide and other oxides which
combine with water to form acids. Water bodies
with low buffering capacity (i.e., low alkalinity) can-
not counteract acid rain, and the pH level lowers
to a lethal level of less than 4.

Low pH caused by acid precipitation has elimi-
nated fish populations in many lakes in the Adiron-
dack Mountains of New York, the LaCloche
Mountains of Ontario, and the lakes and streams of
southern Norway and western Sweden ( Fritz 1980 ).

Phosphates. Phosphates occur in water as a result
of leaching from minerals or as one of the stabilized
products of decomposition of organic matter.
Phosphorus also occurs as orthophosphate from land
agricultural application and is essential to the growth
of organisms. It can be the nutrient that limits
productivity of a water body. The EPA (1976)
discusses effects of phosphorus concentrations on
aquatic life but does not set any criteria for fresh-
water fish or aquatic life.

Polychlorinated Biphenyls (PCB). The EPA

criterion is 0.001 u.g/1 for fresh-water and marine life
and for the consumers thereof.

Selenium. The EPA criterion is 0.01 mg/1 of the 96-
hour LCS() value, as determined by bioassay using
sensitive resident species for marine and fresh-water
aquatic life.

Silver. The EPA criterion is 0.01 mg/1 of the 96-hour
LCso value, as determined by bioassay using sensitive
resident species for marine and fresh-water aquatic
life.

Solids (suspended, settleable) and Turbidity.

Suspended or settleable solids include both inorganic
and organic materials. Turbidity is an expression of
the optical property that causes light to be scattered
and absorbed rather than transmitted in straight lines
through water. The EPA criterion for both solids
(suspended, settleable) and turbidity for fresh-water
fish and other aquatic life is that "settleable and
suspended solids should not reduce the depth of the
compensation point for photosynthetic activity by

more than 10 percent from the seasonally
established norm for aquatic life."

Thurston et al. (1979) disagree with this criterion
because "the criterion treats solids and turbidity as
synonymous, which they are not." They recommend
two sets of criteria: residue (solids) measured in
mg/1 and turbidity measured in Nephelometric Tur-
bidity Units (NTU's). Both sets of measurements
are to be done in accordance with "Standard Meth-
ods for the Examination of Water and Wastewater
(American Public Health Association et al. 1976).

Commercial turbidimeters measure light transmitted
in NTU's. Turbidity criteria vary regionally and may
depend on the type of fishery of concern, i.e. warm-
or cold-water.

Sulfide (Hydrogen Sulfide). The EPA criterion is
2 u.g/1 undissociated H2S for fish and aquatic life.

Tainting Substances. Tainting substances can
produce undesirable flavor in the edible portions of
fish or other aquatic vertebrates. Diesel oil in very
small amounts will taint fish flesh.

WATER QUALITY BASELINE DATA
COLLECTION— PROJECT PLANNING

Land-use projects having major impacts on
water quality require special attention in establishing
baseline pre-project water quality. An example of
gathering water quality data before surface mining is
displayed in Table 1 (Harris et al. 1983). The field and
laboratory analyses are divided logically. Baseline
physical and chemical data will provide a basis for
water quality monitoring during the lifetime of the
land-use project.

WATER QUALITY MONITORING PROGRAMS

Water quality monitoring can be related to man-
agement objectives on a long-term basis. Methods
for testing the statistical significance of change in
water data collected over a period of time are out-
lined in several publications. Two publications pro-
viding statistical analysis procedures for field use are
"Water Quality Monitoring Programs" (Ponce 1980)
and "Stream Monitoring Techniques" (Armour et
al. 1983).

Water Quality

637

Table 1. Water quality and bottom-sediment parameters measured in the Tyro Creek watershed, September
1981 through September 1982 (from Harris et al. 1983).

Field Determinations1

Laboratory Analyses1

Bottom Sediments

Discharge (cfs)

Suspended sediment

Sieve analyses

Temperature (°C)

Calcium (Ca)

Trace metals and

pH (as units)

Chloride (CI)

nutrients (mg/kg)

Alkalinity (as HC03)

Fluoride (F)

Arsenic (As)

Specific conductance

Potassium (K)

Cadmium (Cd)

(|jirmhos/cm)

Magnesium (Mg)

Chromium (Cr)

Color (Pt-Co)

Sodium (Na)

Cobalt (Co)

Turbidity (JTU)

Nitrate (N) (dissolved)

Copper (Cu)

Dissolved oxygen

Nitrate (N03)

Iron (Fe)

Silica (Si02)

Lead (Pb)

Sulfate (S04)

Mercury (Hg)
Selenium (Se)

Trace metals (u.g/1)

Strontium (Sr)

Arsenic (As)

Vanadium (V)

Cadmium (Cd)

Zinc (Zn)

Chromium (Cr)

Chloride (CI)

Cobalt (Co)

Phosphate (P04)

Copper (Cu)

Nitrate (N03)

Iron (Fe)

Lead (Pb)

pH (as units)

Vanadium (V)

Manganese (Mg)

Mercury (Hg)

Selenium (Se)

Strontium (Sr)

Zinc (Zn)

All measurements should be taken in mg/l unless otherwise indicated

LITERATURE CITED

AMERICAN PUBLIC HEALTH ASSOCIATION. 1976.

Standard methods for the examination of water and

wastewater, 15th edition. Am. Public Health Assoc.

New York, NY. 1134pp.
ARMOUR, C.L., K.P. BURNHAN, and W.S. PLATTS. 1983.

Field methods and statistical analysis for monitoring

small salmonid streams. U.S. Dep. Inter., Fish and

Wildl. Serv. FWS/OBS-83/33. 200pp.
FRITZ, E.S. 1980. Potential impacts of low pH on fish and

fish populations. U.S. Dep. Inter., Fish and Wildl. Serv.

Biological Services Program, National Power Plant

Team, FWS/OSS-80/40.2. I4pp.
HARRIS, C.H., P.E. O'NEIL, R.V. CHANDLER, M.F. METEE,

and E.J. MCCULLOUGH. 1983. Biological and hydro-
logical impacts of surface mining for federal minerals
on the Tyro Creek Watershed, Alabama, Phase 1.
Premining-aquatic baseline information. Contract per-
formed for U.S. Dep. Inter., Bur. Land Manage. 98pp.

PONCE, S.L. 1980. Water quality monitoring programs. U.S.
Dep. Agric, For. Serv., 3825 E. Mulberry St., Ft. Col-
lins, CO 80524. 66pp.

THURSTON, R.V., R.C. RUSSON, CM. FETTEROLF, Jr., T.A.
EDSALL, and Y.M. BARBER, Jr. 1979. A review of the
EPA Red Book: Quality Criteria for Water. Am. Fish.
Soc. Bethesda, MD. 313pp.

U.S. ENVIRONMENTAL PROTECTION AGENCY. 1976.
Quality criteria for water. U.S. Environmental Protec-
tion Agency, Washington, DC 20460. 501pp.

. 1979. Methods for chemical analysis of water and

wastes. EPA-600/4-79-020. Washington, DC.

638

Water Quality

31
VEGETATION

Bertin W. Anderson and Robert D. Ohmart

Center for Environmental Studies
Arizona State University
Tempe, AZ 85287

Editor's Note: Although vegetation is widely recog-
nized as one of the most important determinants of
wildlife abundance and distribution, most systems
for measuring and classifying vegetation have been
developed for other purposes. Development of vege-
tation measurements for wildlife purposes has
mostly focused on measuring forage supplies for
big game and other herbivores. Such techniques are
well-described in the literature and are discussed
briefly in Chapter 25, Ungulates.

This chapter focuses on vegetation measurement
technicpies that have been found useful for ( 1 )
evaluating habitat for vertebrate communities, (2)
predicting presence or abundance of individual
species for a wide variety of vertebrates, and (3)
predicting species richness of habitat areas. Because
of the diversity of possible approaches, the authors
have not attempted to provide a thorough review of
all techniques and their relative strengths. Rather,
they have described one system that is reasonably
efficient and has worked well for them. Biologists
will undoubtedly need to modify systems to meet
the needs of their inventory or monitoring pro-
grams. Nevertheless, the attributes described in this
chapter should be useful predictors of wildlife habi-
tat quality in a wide variety of habitats, and the
measurement techniques described should be usable
or modifiable for a diversity of problems.

INTRODUCTION

This chapter describes attributes of plant com-
munities that we found useful in evaluating habitat
for vertebrate wildlife. We describe methods for
measuring these attributes and show how these
measurements may be useful in predicting species
presence or absence, species abundance, and wildlife
species richness in various communities. We devel-
oped these primarily for riparian habitat, but the
general procedures should be applicable to other
habitats. This is not intended to be an exhaustive
catalogue of all or even a majority of the methods
available, nor a critical review of the numerous
methods for quantifying vegetation attributes. Re-
views of methods for measuring vegetation can be
found in U.S. Department of the Interior, Bureau of
Land Management ( 1985); Daubenmire (1968);
Mueller-Dombois and Ellenberg (1974); and other
standard texts.

One of the fundamental reasons for measuring
attributes of a plant community is to classify the
stand under consideration. Classification facilitates
communication by reducing the verbage required to
provide someone with a clear and, hopefully, unam-
biguous mental picture of the community being con-
sidered. We think classifying and describing plant
communities are virtually synonymous.

Vegetation

639

In our experience, time and personnel for meas-
uring vegetation are always limited. Therefore, we
have developed a reasonably small number of field
methods that can be done relatively quickly and,
from which, other plant community descriptors can
be generated. For example, we measure foliage den-
sity in the field and, from this, calculate patchiness in
the horizontal dimension, foliage height diversity,
and foliage density at various vertical layers (i.e.,
ground, shrub, canopy). We describe these methods
in this chapter. Other sets of more traditional meth-
ods can lead to the accumulation of data equally
useful in describing a vegetation community, but
their use will probably be more costly in terms of
time and personnel.

In classifying (describing) vegetation communi-
ties, a two- or three-dimensional approach should
be considered. In general, the physiognomy or struc-
ture of the vegetation represents two dimensions.
For example, a given stand of vegetation varies in
vertical and horizontal space. Variation in the verti-
cal dimension, whether single or multilayered, is
particularly useful in describing the stand. Similarly,
the floristics, i.e., species composition of a stand, is
often invaluable in describing that stand. If structure
and floristics can be relatively quickly and easily
quantified, limits of a vegetation type can be clearly
defined. Standard texts describing methods in syne-
cology (Daubenmire 1968) or in mapping vegetation
(Kuchler 1967) should be consulted for more details
on quantifying vegetation variables and classifying
vegetation.

The same characteristics used in quantitatively
describing a stand of vegetation can be used in de-
veloping predictive capabilities relative to the resi-
dent wildlife. We describe reasonably fast and
accurate field methods for quantitatively classifying
vegetation and correlating wildlife associations with
various attributes of vegetation communities.

VEGETATION ATTRIBUTES

Simple Basic Variables

We refer to foliage density, plant species com-
position, and fruit production as simple, basic varia-
bles because they are usually quantified in the field.

Foliage Density. Foliage density refers to the
amount of green foliage present or to the amount of
leaf-bearing stems and leaves per unit area. (Foliage
density should not be confused with "plant density"
or "density," which is the number of plants per unit
area.) Usually foliage density is measured at various
vertical levels (e.g., every yard or 0.1 yard [meter or
0.1 meter]). Foliage density measurements taken in
summer may be useful in describing the foliage
density in winter, in terms of the relative amounts of

leaf-bearing stems and leaves present, thus negating
the need for remeasurements in winter. This
procedure is most valid in areas dominated by trees
or annuals.

Species Composition. This can be easily
determined by counting individuals of each tree/
shrub species present. This is not as simple as it
sounds; size classes must be considered. Even then,
two trees of the same height and species can be
quite different. The general health of a tree or tree
density can affect general structure. Thus some
thought must be given to what will actually be
counted.

Fruit Production. In stands of vegetation that
produce fruit, especially fruits that are sought by
wildlife, it may be useful to obtain some idea of the
total fruit produced. We have done this when the
correlation between number of trees present and
fruit production is rather poor. For example,
mistletoe {Phoradendron californicum ) along the
Colorado River parasitizes honey mesquite (Prosopis
glandulosa) more frequently than other tree
species. However, the proportion of trees parasitized
varies widely from stand to stand. We therefore
obtained estimates of the number of mistletoe
clumps in a given stand. Similarly, pod production by
individual honey mesquite trees varies widely
between stands; thus, there is a poor correlation
between number of trees present and production of
pods. This is also true of pod production by
screwbean mesquite (P. pubescens). At higher
elevations, some shrub species may be important for
berry or seed production for wildlife. These shrub
species should be noted and given important
management concern.

Composite or Derived Variables

From these foliage density and species composi-
tion measurements, a host of other variables can be
derived. Foliage height diversity (FHD) is calculated
from the foliage density measurements taken at var-
ious vertical planes. Horizontal foliage diversity
( patchiness ) can be determined by considering foli-
age density variation in the horizontal plane. In verti-
cally complex (i.e., multilayered) communities, it
may be desirable to calculate patchiness for each of
the vertical layers, with total patchiness equal to the
sum of patchiness in the vertical layers. The number
of vertical layers recognized is an arbitrary matter,
as is the size of the area being called a patch. The
concept of FHD is depicted in Figure 1 and patchi-
ness in Figure 2.

Tree density estimates can be expressed as the
number of each species per unit area or as the pro-
portion of the total tree species present. When meas-
uring foliage density, one may record the plant

640

Vegetation

Figure 1. Diagram of foliage diversity in the vertical plane. The stand depicts an area of at least 25 a.
(10 ha).

C?

Q2 G>

O

I -J \K \w

L, L^Ok^ JWfi'a*

gggSj^ i ^

/

O

Figure 2. Diagram of foliage diversity (patchiness) in the horizontal plane at each of three vertical layers.
The blocks represent patches of roughly 5 a. (2 ha).

Vegetation 641

species contributing to the foliage density at each
point where measurements are taken. The density of
each tree/shrub species can then be given for each
vertical layer or in terms of the total foliage density
for the entire area.

Usefulness of the Variables

We found that the counts of tree/shrub species
were particularly useful in predicting the presence
and densities of many rodent and bird species (An-
derson and Ohmart 1984a, in press-b; Rice et al.
1983, 1984). We found that patchiness and FHD
have useful predictive value, but are not as good a
predictor as tree/shrub counts (Anderson and
Ohmart 1984a, in press-b; Rice et al. 1983, 1984).
Mistletoe counts were associated with the presence
of frugivorous birds (Anderson and Ohmart 1978).
In general, foliage density at low levels had a high
association with lizards (Anderson and Ohmart
1982) and also some rodents and birds (Anderson
and Ohmart, in press-b ). At various times and places,
desert mule deer (Odocoileus hemionus) were asso-
ciated with screwbean and honey mesquite, but foli-
age density and diversity were also important in
explaining their habitat selection (Haywood et al.
1984).

From this brief discussion, it seems that all of
the vegetation variables given may be important to
at least some species of wildlife. We found consider-
able seasonal variation in habitat selection within a
group, i.e., birds (Anderson and Ohmart 1984a).
Rodents used the vegetation differently in any given
season than the majority of bird species. The point
is, there is no a priori means of selecting a single
attribute, or even a few vegetation attributes, that
will be adequate for predicting wildlife use of a
habitat.

Findings from our research lead us to offer a
precautionary note to others planning habitat stud-
ies, particularly studies of avian habitat use. From
most literature sources, we would have been led to
de-emphasize tree species measurements completely.
We have found many habitat attributes to be crucial
to the prediction of habitat selection of at least some
bird species. Had only a small set of attributes been
collected, analyses would have been successful
enough that one would have been tempted to con-
clude erroneously that the important habitat attri-
butes and selection processes for the entire
community had been captured in the study.

Studies of habitat use should be designed to
sample the annual, seasonal, and spatial variation in
habitat use of the animal community. Within the
limits of biological reality, biologists should measure
as comprehensive a set of attributes as time and
money allow and should avoid measuring only a few
attributes based on knowledge of a few species, data

from other areas, or information from the literature.
Studies providing data for environmental manage-
ment plans, especially habitat modification projects
with impacts on substantial portions of the animal
communities, should be based on data sets adequate
for a thorough description of the system. If we had
studied selected subsets of our species, transects,
time periods, or habitat attributes, we could have
undoubtedly produced results supporting quite a
wide range of invalid proposals. Such actions are not
ecologically wise nor do they lead to effective
management.

MEASURING VEGETATION ATTRIBUTES

Establishing Study Sites

Lines or transects are established by cutting
swaths 3 ft ( 1 m ) wide through the middle of stands
encompassing at least 25 a. ( 10 ha), with dimensions
of at least 2,461 ft (750 m) long by 164 ft (50 m)
wide. Small patches (2.5 a. [1 ha]) of vegetation, dif-
fering in species composition or structure from the
major type in the stand, should be bisected by the
transect at right angles whenever possible. In no
case should a transect be situated so that vegetation
differing from the stand as a whole is paralleled by
the transect at a distance closer than 49 ft ( 1 5 m ).
When vegetation is very much unlike the stand as a
whole, it has an unduly large effect on the vege-
tation analysis and on estimates of bird and rodent
densities.

Semipermanent markers should be placed at the
beginning and end of each transect. A stake with
the distance from the beginning of the transect in-
scribed on it can be driven into the ground at inter-
vals every 492 ft ( 1 50 m ), which can represent a
patch size. The size of a patch will depend on the
species being studied. Patches 492 ft ( 150 m) long
x 656 ft ( 200 m ) wide represent average territory
size for bird species in southwestern riparian vegeta-
tion. A transect 2,461 ft (750 m ) long has five sub-
plots or patches, each 492 ft ( 1 50 m ) long on each
side, for a total of 10 subplots. Vegetation data
should be collected within each subplot. Each tran-
sect can be numbered, and the number and direc-
tional orientation recorded on a map. A typical
transect is illustrated in Figure 3-

Vegetation Measurements

In each subplot ( patch ), tree counts and foliage
density measurements are made. Tree counts are
made only once on each transect unless the area is
later affected by some major disturbance. Counting
is unaffected by amount of foliage present and can
be done any time. Individuals of each species of tree
or shrub within 49 ft ( 1 5 m ) of the transect are

642

Vegetation

/

Edge ot Community - Structural Type

Outer Boundary ofx Sample Area (130 m)

Subplot 1

Subplot 2

W Subplot 3

Subplot 4

Subplot 5

* 750

m >

600

m

450

m

300

m

150

m

*

s

0

m \

Subplot 6

Subplot 7

Subplot 8

Subplot 9

Subplot 10

Transect Path

Figure 3- Typical transect through a relatively
homogeneous stand of vegetation, showing
individual subplots (patches) and outer
boundaries.

counted in each 492 ft ( 1 50 m ) subplot. Each indi-
vidual tree is categorized by height (e.g., > or < 10
ft [> or < 3 m]), by presence or absence of mistle-
toe, and by its condition (alive or dead).

Sometimes shrubs or trees grow in densities so
great that it is not feasible or possible to count indi-
viduals. Often densely packed individuals provide
no more ground cover than trees in less dense areas.
For example, 20 trees in one area could equate, in
terms of ground cover, to 200 trees in another area.
To circumvent this problem, measure the height
and north-south crown diameter of 30 or more indi-
viduals of each tree species growing at various
heights in uncrowded conditions. From these meas-
urements, develop regression equations to determine
the ground cover by an individual tree ( shrub ) of a
given height ( Figure 4 ). Thus, when a dense patch of
trees or shrubs is encountered, measure the area of
the patch and obtain the average height of the trees
in it. Then, divide the area of the patch by the area
occupied by the average single tree of the same
height growing in uncrowded conditions to obtain
the equivalent number of trees or shrubs growing
under uncrowded conditions. This method may be
applied to all trees and shrubs to obtain a rough
estimate of the number of full-sized equivalents of a
given plant species in an area.

Figure 4. (A) Reference sample of quail bush. (B) Sample area with approximately 30% cover by quail bush
or 1,435 ft (135 m~). Since one shrub occupies 9 ft" (0.83 m ), the area has approximately 163 individ-
ual shrubs. The area occupied by each patch of quail bush was measured to determine the ground cover
by quail bush.

Vegetation

643

Foliage density estimates are made in all stands
annually between May and July; stands undergoing
succession (i.e., burned and regenerated areas) can
be measured again in September or October or even
more frequently if desired. Relative foliage density
estimates are made using a cover board technique
(MacArthur and MacArthur 1961 ). Sampling is done
on each side of a transect at three points ( 49 ft
[15 m], 246 ft [75 m], and 432 ft [135 m]) from the
beginning of each subplot (patch). Thus on a 2,461
ft (750 m) transect, there are 15 points per side
for a total of 30 points (Figure 5).

At each sample point, one observer paces one
step perpendicular to the transect. A second ob-
server holds a board (approximately 8 x 16 in. [20
X 40 cm]) at a given height behind the nearest
green, leafy vegetation on the appropriate side of the
transect. The first observer directs the second ob-
server to stop when green foliage covers half the
board. Distance from observer to board is measured
with a tape measure or range finder ( Figure 6 ). Foli-
age density measurements are recorded in feet be-
cause of the scaling of the equipment we used; also,
all vegetation calculations were based on the English
system of measures. All of this could just as easily
be done in metric units.

Round all measurements to the nearest foot
(0.3 m), except the first. In the first foot (0.3 m), a
distance of 2 in. ( 5 cm ) represents very dense vege-
tation, but zero means that foliage is absent. Thus,
any distance > 0 but < 1 ft (< 0.3 m) should be
regarded as 1 ft (0.3 m). Often it is difficult to obtain
agreement between two observers for distances
< 1 ft (< 0.3 m), yet the difference in, for example,

the foliage density estimate between 1 in. (2.5 cm)
and 2 in. (5 cm) is large. Rounding to 1 ft (0.3 m)
results in a conservative estimate of foliage density
in very dense places, but yields reproducible results.
Vertical foliage density determinations can be made
at, for example, 0.5 ft (0.15 m), 2 ft (0.6 m), 5 ft
(1.5 m), 10 ft (3 m), 15 ft (4.5 m), 20 ft (6 m),
25 ft (7.5 m), 30 ft (9 m), and every 10 ft (3 m)
thereafter, until no vegetation is present. Where
measurements are taken is up to the judgment of
those in charge of the project. Theoretically, one
should use a ladder to make measurements at higher
levels. This is impractical. We use a range finder to
locate a point, for example, at 26 ft (8 m), then esti-
mate as carefully as possible the distance to a second
point where leaves would cover half the board.

Distances are measured with a tape measure to
the nearest foot within the first 10 feet (3 meters),
because the foliage density index is more sensitive
to vegetation located nearby. Measurements beyond
this are estimated with the aid of a range finder. In
sparse areas, the distance to the few tall trees and
prominent shrubs with foliage should be measured
from only one sample point within a subplot — the
point to which they are closest. These data should
be recorded; many suitable ways are imaginable.

Frequency of Measurements. It is important to
take as many measurements as necessary to obtain a
reasonably accurate reflection of the true foliage
density and diversity of a stand. In general, the more
measurements that are taken, the greater the
precision; however, for most field workers, time and
personnel are restraining factors. Beyond some
number, additional measurements increase labor

75m ^

Figure 5. Sampling points for foliage density measurements within each subplot along the length of a
transect.

644

Vegetation

11 ft

M^U K^^

■H'V" mp&m2

5 ft

Mir^/

Annual

Atriplex

Willow

f Transect

Figure 6. Selection of vegetation for foliage density measurements.

requirements more than can be justified when the
increase in precision is very small.

In identifying the effort required to obtain rea-
sonably precise data, we made foliage density meas-
urements at 1,2, 3, 4, 5, 9, and 12 points within
each subplot on each side of 7 transects, which in-
volved 104 subplots. Foliage density calculations
were made using one measurement from each sub-
plot, then two measurements, and so on, until five
separate sets of calculations were made. This was
repeated using 9 and 1 2 measurements per subplot.
We assumed that 1 2 measurements per subplot
yielded results as close to reality as could possibly
be obtained with our methods. The foliage density
and diversity results obtained for each set of meas-
urements were expressed as percentage deviation
from the results obtained with 1 2 measurements per
subplot. The mean percentage difference and stan-
dard deviation decreases as the number of measure-
ment points increases (Figure 7), until five
measurements were included per subplot. Precision
did not increase with nine measurements per sub-
plot. We ultimately chose to make three measure-
ments per subplot. Making five measurements
decreases the error rate by only 1 % , but increases
the work by 67% . In our judgment, the additional
precision is not warranted by the effort to obtain it.

Repeatability. Work encompassing a large area
(100,000 a. [40,000 ha]) will probably require
several individuals to collect data. Thus, all
personnel must be carefully trained and the
similarity of data collected from the same stand by

25-

20-

cc

O 15-

cc
cc

LLI

1—

z

LU

o

CC '0-

LU

Q.

z
<

LU

5-

\

V

1

1 2 3 4 5 6 7 8 9 10 11 12

NUMBER OF POINTS PER STRIP

Figure 7. Mean percentage error associated with
foliage density and diversity measurements when
measurements were made at various numbers of
points in each subplot. The values obtained from
1 2 measurements per subplot were assumed to
accurately represent the true values. This standard
was used for comparing values obtained with
smaller numbers of measurements per subplot.

Vegetation

645

different observers, determined. In general, we found
least agreement in data from stands with tall
vegetation (> 33 ft [> 10 m]). Space precludes an
exhaustive presentation of our data; however, we
describe where discrepancies among observer teams
were greatest. For a more thorough presentation,
see Anderson and Hunter ( 1983)

Foliage measurements were taken by four teams
in a relatively dense stand encompassing about 25
a. (10 ha) and containing trees to 64 ft (20 m) tall.
It was assumed that some observer estimates would
be too high and some would be too low; thus the
average for the teams would approximate the "cor-
rect" estimate for each foliage variable. The devia-
tion from the average was then considered to be the
error rate for a team. The error divided by the "cor-
rect" value yielded the percentage of error. This
procedure indicated, for all variables, a mean error
of 8% . The greatest difficulty was in assessing vegeta-
tion density at the highest levels; the average error
at 26 ft ( 8 m) was 33% (Table 1 ). The mean error
was lowest for FHD (2% ) and total patchiness (4% ).
Since the error rates for foliage density estimates
were higher than hoped, we examined the results in
greater detail.

Although foliage density estimates had a mean
error rate of 15%, the profiles of vertical foliage
distribution derived from the estimates were nearly
identical in three of the four cases in the first test
and, in all four cases, the vegetation would be classi-
fied the same. In the second test, two of the three
profiles would have led to the same classification.
This is of major importance because the fact that all
four teams in the first set obtained results that led to

classifying the stand in the same way, should mini-
mize the importance of the observed variation in
data obtained by different teams. FHD and patchiness
estimates obtained by different teams for a given
stand were always similar.

Overall, we are satisfied that these methods
accurately and consistently differentiated areas of
high, intermediate, and low foliage density and diver-
sity in the vertical and horizontal planes. However,
stands differing slightly in foliage density probably
cannot be reliably separated.

Training Personnel. All personnel should be
thoroughly trained in field techniques before
collecting data on their own. More experienced
personnel should accompany less experienced
personnel in the field. Having people with
experience with the technique is not essential, but
having dedicated, conscientious personnel is.

Need for Preliminary Studies

In any study where discovery of relationships
between animals and environmental variables is a
goal, there is a need for preliminary studies (Green
1979; Platts 1981; Platts et al. 1983). Such studies
are essential because they allow investigators to test
to see if their methods satisfactorily produce the
desired kinds of information. Preliminary studies also
offer an opportunity to become thoroughly familiar
with the study area before beginning the study and
provide a training period for field personnel. In addi-
tion, preliminary studies allow the investigators to
determine differences between experimental and
control areas prior to treating the experimental area.

Table 1. Average error associated with various vegetation variables collected by four different teams from
the same stand. (The mean measurements by all four teams were the standards with which comparisons
were made.)

Foliage Density at Various Heights

Patchiness

0.0-0.6 m 0.6-4.5 m 4.5-7.5 m &7.5 m

0.0-0.6 m

0.6-4.5 m

4.5-7.5 m s=7.5 m

Team

(0-2 ft) (2-15 ft) (15-25 ft) (^-25 ft)

Total

FHD

(0-2 ft)

(2-15 ft)

(15-25 ft) (

^25 ft)

Total

TESTI

1

20.5 3.4 20.8 74.2

31.6

1.1

0.9

0.8

0.9

11.6

1.8

2

12.6 13.9 5.9 21.0

1.7

4.1

5.2

3.7

8.7

6.1

6.7

3

16.8 6.1 14.8 27.0

17.3

0.9

2.0

1.4

1.5

6.9

3.7

4

16.2 4.3 11.7 26.3

15.9

2.0

2.3

1.3

1.5

10.7

5.0

TEST II

1

8.5 3.2 1.0 13.5

2.4

2.8

1.5

0.2

2.4

0.0

2.8

2

4.6 7.3 10.2 42.0

1.3

2.8

1.3

0.9

0.3

2.0

2.8

3

12.7 10.5 11.0 28.0

1.6

0.1

1.6

2.1

1.2

0.6

0.1

Mean

13.3 7.0 10.8 33.1

10.3

2.0

2.1

1.5

2.4

5.4

3.3

SD

4.9 4.0 6.3 20.0

11.7

1.4

1.4

1.1

2.9

4.7

2.1

646

Vegetation

For example, if the objective is to determine the
impact of light grazing on an area in terms of
changes in wildlife densities and diversities, it is
critically important to establish plots where all varia-
bles, except the one being tested, are controlled.
This means that the vegetation and wildlife use must
be analyzed in the control and experimental areas
before grazing. Impacts can only be evaluated after
internal variation, before the impact, is known.

The argument is often heard that such prelimi-
nary studies are too time-consuming and expensive
to carry out. This is fallacy; preliminary studies can
save time and money and are the only way of obtain-
ing scientifically valid results (Green 1979). Without
them, results obtained after an impact cannot be
determined because the lack of knowledge of the
variation before the impact was unknown. In other
words, as Platts ( 1981 ) correctly pointed out, the
notion that collection of inventory "garbage" (i.e.,
data without preliminary studies ) leads to reliable
analyses is purely mythical. Unfortunately, it is a
pervasive myth. If preliminary studies cannot be
done, the study should not be undertaken or under-
taken with the realization that conclusions based
on the findings are not well-founded scientifically.

ANALYTICAL METHODS

Foliage Density

Each plant distance measurement is converted
to surface area per cubic unit of space (i.e., foliage
density ) according to the following formula:

K

log,2 0.693

These measurements will produce a table of
density measurements such as in Table 2, which can
be used for subsequent calculations of derived varia-
bles such as vertical and horizontal diversity.

Table 2. Sample foliage density estimates (ft "/ft )
used for calculating patchincss and foliage height
diversity.

0.15 m

0.6 m

1.5 m

3 m

4.6 m

Plot

(0.5 ft)

(2 ft)

(5 ft)

(10 ft)

(15 ft)

1

0.16

0.20

0.29

0.10

0.01

2

0.12

0.15

0.23

0.06

—

3

0.08

0.09

0.27

0.09

0.01

4

0.28

0.15

0.09

0.01

0.00

5

0.19

0.22

0.09

0.02

0.00

6

0.18

0.34

0.29

0.10

0.02

7

0.07

0.31

0.31

0.03

0.01

8

0.08

0.18

0.31

0.02

—

9

0.15

0.16

0.32

0.03

—

10

0.23

0.15

0.13

0.01

—

Vertical Diversity (Foliage Height Diversity)

Vertical or foliage height diversity (FHD) for
each transect is calculated according to the informa-
tion theory (Shannon and Weaver 1949):

D

D

FHD

-2

(p,)(l0gnPi)

Where: K= the foliage density and D
ured distance.

the meas-

Foliage density per subplot is the sum of the
average of the three measurements taken at each
vertical plane. For example, foliage density at 5 ft
(1.5 m) in one subplot, for which the distances were
9, 15, and 2 ft (2.7, 4.5, and 0.6 m), respectively,
would be calculated as follows:

0.693 0.693 , 0-693
9 15 2

■*■ 3 = 0.1556

Foliage density at 10 ft (3 m) for distances of 1, 2,
and 3 ft (0.3, 0.6, 0.9 m), respectively, would be —

0.693 0.693 0.693
1 2 3

3 = 0.4217

The density for the two vertical planes is 0.1556
+ 0.4217 = 0.5773. If no green foliage occurred at
a particular point, a zero is used in the calculations.

Where: p, is the proportion of total foliage density

contributed by the density at level i. (Sample
calculations are shown in Table 3- )

Many stands of vegetation include a shrubby
layer up to about 3 ft ( 1 m). We chose foliage den-
sity estimates from 0.5 ft (0.15 m) and 2 ft (0.6 m)
to represent this vertical layer. Foliage density in this
layer is the sum of the density at 0.5 ft (0.15 m)
and 2 ft ( 0.6 m ). Most stands will probably have an-
other layer extending, for example, from 2 to 15ft
(0.6 to 4.5 m). Foliage density estimates at 5 ft ( 1.5
m) and 10 ft (3 m) could be used to represent this
layer. Many stands have a third layer, usually poorly
developed, extending above the second layer for
an additional 7 to 10 ft (2 to 3 m) or more, depend-
ing on locality. We chose foliage density estimates
at 15 ft (4.5 m) and 20 ft (6 m) to represent this
layer. Foliage < 7.5 m (< 25 ft) could represent a
fourth layer. The concept of FHD with four vertical
layers is depicted in Figure 1.

Vegetation

647

Table 3. Example of calculation of foliage height diversity from data in Table 2.

Item

Height Class

Mean total density
Proportion (p,)
log-toPi
p,log10p.

0.15-0.6 m
(0.5-2 ft)

1 .5-3.0 m
(5-10 ft)

4.6-6.0 m
(15-20 ft)

> 7.5 m
(> 25 ft)

Total

0.35

0.55

-0.26

-0.14

0.28

0.44

-0.36

-0.16

0.00

0.01

-2.20

-0.01

0
0
0
0

0.63

FHD = - 2(p,)(log10p,) = -[(-0.14) + (0.16) + (0.01) + (0)]
= 0.31

Horizontal Diversity (Patchiness)

Horizontal diversity (or patchiness) is a struc-
tural feature of a habitat describing the regularity of
vegetation as it is distrihuted in the horizontal plane
(Figure 2). A citrus orchard with roughly equally
sized and evenly spaced trees has little horizontal di-
versity. Patchiness or diversity is greater in a honey
mesquite-quail bush (Atriplex lentiformis ) habitat
with irregularly spaced trees and shrubs of different
heights. Diversity in the horizontal plane can be
calculated for any vertical layer from which foliage
density estimates are made.

The variance associated with the mean total
foliage density for each vertical plane across all sub-
plots can be used as a measure of horizontal diver-
sity. For example, the summed and averaged foliage
densities for 0.5 ft (0.15 m) and 2 ft (0.6 m) in each
plot (Table 4) yields the mean foliage density for
the layer 0 to 2 ft (0.0 to 0.6 m), in this case, 0.35.
Horizontal diversity is the variance associated with
mean total foliage density, in this case, 0.01. Variance
or standard deviation squared (s") is defined as —

This variance is calculated for each vertical
layer. Total horizontal diversity is the sum of the
variances for all layers.

In calculating horizontal diversity, we are assess-
ing the variance between subplots. Therefore, if hori-
zontal diversity is thought of as patchiness, we are
defining a patch as a unit 492 ft ( 150 m) long and as
wide as the distance from the transect to the edge
of the stand under study, usually 420 ft (128 m).
Choice of this patch size was based on evidence that
many common birds in the area use patches of about
this size (Conine 1982). It is possible that an area
that rated very patchy on a smaller scale could be
rated homogeneous using this particular scale.

Since 0.00 and 0.69 represent minimum and
maximum foliage density values, respectively, maxi-
mum horizontal diversity or patchiness for a given
layer is 0.238. Since there are four layers, maximum
horizontal plane diversity is 4 x 0.238 or 0.952.
Since this is close to 1, the sum of the diversity for
the four layers closely represents the percentage
of maximum diversity possible for an area.

HDI

2

s =

2 (k, - K)2

n-1

It can be calculated more easily using the
formula —

HDI

2

s =

2 k,2 - (Sk,)2

i= 1 n

n-1

Where: kj = the foliage density of the ith sam-
pie
K = the mean foliage density for the

sample
n = sample size
HDI = Horizontal Diversity Index

Another method for calculating FHD and patchi-
ness is to simply record the presence or absence of
vegetation at various vertical positions. This could be
done with a long pole and/or a range finder. More
stops would have to be made, but FHD, relative den-
sity values, and patchiness estimates could be made
on the basis of the proportion of total points at
which foliage occurred. This method may be
quicker, could reduce the amount of calculation, and
might be equally accurate.

CLASSIFYING VEGETATION

If the goal of a field project is to determine hab-
itat associations for a wildlife group, such as birds,
over a relatively large area (e.g., 100,000 a. [40,000
ha]), the area must be sampled with sufficient inten-
sity so all habitats are represented by at least one

648

Vegetation

Table 4. Example of calculation of horizontal diversity index ( HDI ) from data in Table 2.

Item

Height Class

Sum of k.
Sum of squares

kj
Sum squared

(2 k?

n

0.15-0.6 m
(0.5-2 ft)

1 .5-3.0 m
(5-10 ft)

4.6-6.0 m
(15-20 ft)

> 7.5 m
(> 25 ft)

Total

3.49

1 .3073

12.1801
10

2.80

0.8967

7.8400
10

0.05

0.0007

0.0025
10

—

Ik,2 - (1
Hni ~2

k,)2
^ o m

+ 0.01

+ 0.00

+ —

= 0.02

HUI - o - n_1

— U.U I

sample plot ( transected area ). Of course, replication
is desirable. If this sampling is random, then the
number of transects per habitat will be proportional
to the abundance of that hahitat in the study area.
All transects should be about the same length and
should be within a relatively homogeneous stand. At
this point, some arbitrary decisions may have to be
made because of the ambiguity associated with the
term "relatively homogeneous."

In general, experience and common sense will
be immeasurably helpful. A field biologist relatively
familiar with an area will generally know when ( s )he
leaves one habitat type and enters another. It would
be poor judgment to establish a transect in a mixed
cottonwood (Populus fremontii) and willow (Salix
sp. ) habitat and have the transect cross 990 ft ( 300
m) of honey mesquite habitat. Having such a tran-
sect cross a small group of 10 or 15 honey mesquite
trees might not be inappropriate. A general familiar-
ity with what is available will usefully serve as a
guide. If the decision is made to include one or
more small patches of honey mesquite in the cotton-
wood-willow habitat, this will add heterogeneity to
the data. Vegetation often varies considerably over
small areas; thus, some heterogeneity is not necessar-
ily bad and is probably inevitable.

The next step is to cluster transects that are
similar in, for example, vertical diversity. This can be
done by calculating the overlap in vertical foliage
diversity. One way to do this is to use Horn's ( 1966 )
equation for overlap:

R«

2(x, + y,)log(x, + y~j) - SxjlogXj - Sy.logy,
(X + Y) log(X + Y) - X logY - Y logY

Where: x( and yt = proportion of total foliage
density occurring at vertical band i
for stand X and stand Y.

X,Y = the total foliage density for stand X and
stand Y. X and Y represent total foliage
density.

Although this equation appears formidable, it is ac-
tually very simple. The calculations are laborious but
in this age of computers, large numbers of overlaps
can be calculated in minutes. The biologist must,
of course, grasp the fundamental arithmetic process.
Biologically, high overlaps simply represent a mathe-
matical expression of a high degree of similarity in
vertical configuration. Once this is understood, the
computer can take care of the tedium.

From a matrix of overlap values, including all
possible two-way comparisons between stands (tran-
sects), a dendrogram, such as that in Figure 8, can
be constructed using the following relationship:

_ CA

CB

C, AB

Where: C, AB = the overlap of stand C with

stands A and B
CA = the overlap between stand

C and stand A
CB = the overlap between stand

C and stand B

Simply stated, the overlap in the vertical distribution
of foliage in stand C with stands A and B is equal to
the average overlap of C with A and C with B (Cody
1974). The dendrogram (Figure 8) was interpreted
as revealing six categories, based on vertical configu-
ration without consideration of dominant vegetation.
Each transected stand within a category (designated
I-VI ) has a vertical configuration more similar to
other intracategory stands than to any stand within

Vegetation

649

VEGETATION TYPE

I I! Ml

V VI

Figure 8. Dendrogram showing similarities in ver-
tical profile between transects, based on overlap
in vertical distribution of foliage.

any other category. This determination is based on
proportional distribution of the foliage in the vertical
dimension. These categories can now be subdivided
according to the dominant vegetation present
(honey mesquite, salt cedar [Tamarix chinensis],
screwbean-mesquite mix, quaking aspen (Populus
tremuloides), sycamore, etc.). Criteria used to arrive
at these divisions is up to the investigator. As a

guideline, consider that each habitat should be rep-
resented by at least two transected stands and that
the number of recognized habitats should not be
unwieldy. Stress similarities rather than differences in
attempting to reduce the overall number.

The categories could be subdivided again on the
basis of foliage density. For example, if we found a
stand dominated by willow to 20 ft ( 6 m ) containing
1,250 trees/a. (500/ha) and another stand with the
same vertical configuration but only half as many
trees, we might want to recognize these stands as
distinct habitats. Ordinarily, this situation does not
arise because the sparser ( more open ) area would
have a more highly developed layer of vegetation
(shrubs, annuals) and would, therefore, have a differ-
ent vertical configuration.

The vertical classification we arrived at for ripar-
ian vegetation is depicted in Figure 9. From this
information, a quick and easy-to-use key to the var-
ious habitats can be developed (Table 5).

Habitat Heterogeneity

Including several transects in one habitat can
increase within-habitat variation considerably. The
advantage of using the habitat concept is that habi-
tats can be mapped, data are less cumbersome to

Table 5. User's guide to classifying vegetation by dominant tree or shrub species present.

(1) A. Stand in which virtually 100% of the trees present are of one species or virtually 100%

arrowweed Go to 2

B. Trees within stand, clearly of mixed species. The different species may occur as mixed individuals or as
small clumps Go to 3

(2) A. Stand in which trees are composed of nearly 100% of some species (may be occasional, widely scat-

tered individuals of one or more species). Many large stands have arrowweed in patches encompassing
5 a. (2 ha) or more. Honey mesquite stands in addition to or instead of arrowweed may have quail bush,
four-winged salt bush, wolfberry, or mkweed Salt Cedar I-IV or Honey Mesquite III-IV

B. Stand composed of nearly 100% arrowweed; may be an occasional tree or widely scattered clump of
some other shrub Arrowweed

(3) A. Stand of vegetation is structural type I and trees are primarily salt cedar, cottonwood and/or willow with

occasional widely scattered screwbean or honey mesquite tree or clumps of trees. Arrowweed or some
other shrub may occur in relatively widely scattered clumps Salt Cedar-Cottonwood Willow mix

B. Vegetation not structural type I Go to 4

(4) A. Stand of vegetation is structural type II or III Go to 5

B Stand not structural type II or III Go to 6

650

Vegetation

Table 5. User's guide to classifying vegetation by dominant tree or shrub species present (concluded).

(5) A. Stand in which trees are salt cedar with large numbers of Cottonwood and/or willow present; may be

widely scattered individuals or clumps of screwbean or honey mesquite Salt Cedar-Cottonwood/

Willow mix

B. Stand in which trees are mainly salt cedar and screwbean mesquite; may be an occasional,
widely scattered clump or individual Cottonwood and/or willow or honey
mesquite Salt Cedar-Screwbean Mesquite mix

(6) A. Stand of vegetation in structural type IV Go to 7

B. Stand not structural type IV Go to 8

(7) A. Stand composed mainly of salt cedar but with significant numbers of cottonwood and/or willow present;

may be widely scattered individuals or clumps of screwbean or honey mesquite. Shrubs, mainly ar-
rowweed, abundant and occurring in moderate to relatively large patches, sometimes encompassing 5 a.
(2 ha) or more Salt Cedar-Cottonwood/Willow mix

B. Stand much as above but with screwbean mesquite or honey mesquite instead of cottonwood and/or
willow Salt Cedar-Screwbean Mesquite mix or Salt Cedar-Honey Mesquite mix

(8) A. Stand of vegetation is structural types V or VI Go to 9

B. Stand not structural type V or VI Go to 3

(9) A, Stand composed mainly of salt cedar, but with significant numbers of cottonwood and/or willow occurring

as scattered individuals or clumps. Arrowweed is usually abundant (occasionally some other shrub

species such as quail bush also present) and occurring in patches encompassing

several a. (ha) Salt Cedar-Cottonwood/WUlow mix

B. Stand composed primarily of salt cedar but with significant numbers of individuals or clumps of screw-
bean or honey mesquite. May be widely scattered individuals or clumps of screwbean or honey mes-
quite. Arrowweed present as in (9)A Salt Cedar-Screwbean Mesquite mix or Salt Cedar-Honey

Mesquite mix

NOTE: This key can be used to classify about 95% of the riparian vegetation found along the lower Colorado River By applying the
same general principles used to construct the key and a little imagination, rare vegetation types can also be classified

work with, and habitats can be communicated easier
than transects. Furthermore, habitat is usually man-
aged as a concept. However, if microhabitat variation
is marked, it will be a matter of concern to many
biologists. Use of the habitat concept should occur
after one is thoroughly familiar with the variation
that will be concealed and its meaning when inter-
preting subsequent data analyses.

Foliage Density. In separating transects into
structural types, having various structural types
different from one another at a statistically significant
(P < 0.05) level for at least one of the recognized
vertical layers could be desirable. Adjacent structural
types may overlap, but as long as some statistical

differences exist, one can legitimately maintain that
they represent different habitats. Data from our
studies (Anderson and Ohmart 1984b; Anderson and
Hunter 1983) are shown for foliage density in Figure
1 0 and the proportion of foliage in each of three
vertical layers in Figure 1 1 . These figures illustrate
the range of variation found among transects falling
into each category; they also show the mean and
two standard errors of the mean for each type. Note
that when using foliage density measurements, types
V and VI differed little from each other, but when
the proportion of the total foliage found in each of
three layers was considered, type VI had a
significantly greater proportion of its total foliage in
the lower layer and significantly less in the middle
layer ( Figure 1 1 ).

Vegetation

651

1 J v

AT vJ (r^X

4 I]

I J 1 ^JL

J*J<(

^^V*^,

IV

/>w£^.

Figure 9- Examples of vertical configurations for the vegetation structural types defined (Figure 1 1 ) in the
lower Colorado River Valley. Type I is open gallery forest; type II is closed gallery forest; type III is scrub
thicket; type IV is scrub thicket with patches of bushes; type V is mostly shrubs and bushes with scat-
tered scrubby thickets; type VI is bushes, shrubs, and grasses.

Heterogeneity in Tree Counts. The mean number
of trees of a particular species can also vary
considerably among patches. For example, the mean
number of salt cedar per salt cedar thicket with
patches of shrubs was 163 trees, with a very large
standard deviation (105; Table 6). This could be
expected in a very patchy habitat. Some patches will
have more trees than others, which makes them
patchy. However, habitats classified as honey
mesquite-woodland had very few tree species
present other than honey mesquite; salt cedar
thickets had only one species present other than salt
cedar (Table 6). Thus, while salt cedar is virtually
the only tree present in salt cedar habitats, the trees
within such habitats may be tall and relatively
homogeneously distributed (type I) or scrubby and
patchily distributed (type IV), with patches of shrubs
intermingled among the salt cedar.

Within-habitat counts can also vary through
highly localized edaphic features. For example, the
soil moisture level in an old oxbow that is inter-
sected by the transect may have a few individual
tree species not found elsewhere on the transect.

Local heterogeneity in soil layering and struc-
ture can cause heterogeneity in plant structure. The

distribution of soil types within a floodplain is typi-
cally heterogeneous. A highly localized, dense clay
soil type could cause a very local concentration of
soil electrolytes. Vegetation growing in such soil
often attains less stature and biomass (Anderson and
Ohmart 1982) and, therefore, vertical differentiation
is simpler than for adjacent vegetation. Such varia-
tion may be so frequent that it is not feasible or
desirable to delineate it within otherwise relatively
homogeneous stands.

Another source of variation includes scattered
individual trees that were once more widely distrib-
uted. In our study area, Cottonwood and willow
trees, often occurring as widely scattered individuals
or as small clumps (66 x 66 ft [20 x 20 m]) of
trees, are relicts of gradually disappearing habitat
(Ohmart et al. 1977).

Fire, another cause of within-stand heterogene-
ity, in varying degrees and at various times, has af-
fected nearly every stand of vegetation along the
lower Colorado River. When a stand is burned, not
all parts of it burn with equal intensity; some corners
or clumps remain intact. Parts of a stand may burn
more than once, so at any given time not all parts

652

Vegetation

CO

c>

z

LU
Q

LU
O
<

2.200
2.000 -
1.800
1.600 -
1.400 •

1.200 •

1.000

0.800

0.600-

0.400

0.200

0.000

TYPE TYPE

IV

TYPE TYPE

V VI

JUL

a 'B'cl a'b'c Ia'b'c a'b'c Ia'b'cIa'b'c

HEIGHT

Figure 10. Variation in foliage density between subplots within all structural types at each of three vertical
levels. The proportional distribution clearly differentiates the vegetation types, but foliage density does
not. Horizontal lines represent mean values; large rectangles represent ± 1 standard deviation; small rec-
tangles represent ±2 standard errors. A = 0-2 ft (0.0-0.6 m); B = 2-15 ft (0.6-4.5 m); C = > 15 ft
(> 4.5 m).

z

LU
Q

LU
O
<

LL

o

LU

o
<r

LU
0-

Figure 11. Proportional distribution of vegetation in three vertical layers among subplots within various
stands of vegetation, generally classified as one vertical structural type (I-VI). Horizontal lines represent
mean values; large rectangles represent ± 1 standard deviation; small rectangles represent ± 2 standard
errors. A = 0-2 ft (0.0-0.6 m); B = 2-15 ft (0.6-4.5 m); C > 15 ft (4.5 m).

Vegetation

653

are at precisely the same stage of post-fire recovery.
Even when burned evenly, not all parts of a stand
redevelop at precisely the same rate. Thus, at some
level of analysis, considerable heterogeneity could be
found within any fundamentally homogeneous stand.

Although such delineation of plant species het-
erogeneity may be important for understanding some
aspects of the distribution of vegetation or wildlife

using the vegetation, it may be beyond the scope of
study. Such delineation could require more time and
money than is available. Such small parcels typically
range from 2.5 a. (1 ha) to about 12.5 a. (5 ha).
Areas smaller than 25 a. (10 ha) could not be accu-
rately plotted on map scales of 1:9,449 cm. Availabil-
ity of funding and consideration of the desired scale
are factors that must be considered when deciding
how much edaphic variation should be delimited.

Table 6. Average number of trees (
the lower Colorado River.

1 SD) per subplot in each of 23 recognized riparian habitat types along

Vegetation
Type

N

Number of Trees Per Subplots
150 x 15 m (492 x 50 ft)

Percent of
Subplots
with No
Trees of
Dominant
Species

Salt
Cedar

Cottonwood

Willow

Screwbean
Mesquite

Honey
Mesquite

x SD

X

SD

X SD

x SD

x SD

Salt cedar

1

II
III
IV
V
VI

18
8
28
32
109
20

95 20
47 19
74 25
163 105
133 146
31 50

0
0
0
0
0
0

0
0
0
0
0
0

0 0
0 0
0 0
0 0
0 0
0 0

0 0
0 0
7 13

0 0

1 3
0 0

2 20
0 0
0 0
0 0
0 0
0 0

0
0
0
0
1
0

Salt cedar-
cotton wood/
willow

1

II
III
IV
V
VI

18
10
62
52
30
22

52 13

129 46

130 147
38 53
44 49
19 32

59

38

19

0

0

1

27
22

44
0
0
1

87 23
49 34
54 66
29 17
17 21
1 26

0 0
0 0
13 23
7 15
0 0
0 0

0 0
0 0
6 7
0 0
0 0
0 0

0-0
0-0
0-6
3-8
0-0
0-50

Salt cedar-
screwbean
mesquite

II
III
IV

V
VI

10
40
78
84
18

63 24
49 43
60 58
45 39
45 55

2
0
0
0
0

4
0
0
0
0

1 1
0 0
4 25
0 0
0 0

96 17
18 15
39 31
44 62
6 6

0 0
0 0
0 0
0 0
0 0

0-0
0-8
1-6
0-8
0-22

Salt cedar-
honey mes-
quite

IV

38

41 53

0

0

0 0

0 0

35 68

2-6

Honey mes-
quite

III
IV
V
VI

24

122

56

52

0 0
0 0
0 0
0 0

0
0
0
0

0
0
0
0

0 0
0 0
0 0
0 0

1*

0 0
0 0

1*

93 50

31 42

12 7

9 7

0
1
2
2

'Standard deviation not calculated where x < 1.

NOTE: N = number of subplots

654

Vegetation

Intrahabitat Heterogeneity among Patches. The

169 yd (150 m) subplots along transects traversing
stands of a given structural type will reveal much of
the heterogeneity within a habitat. This hetero-
geneity reflects differences between subplots
(patches) in vertical foliage distribution and foliage
density. If the manager decides that a patch is any 5
a. (2 ha) area, the classification system and any
subsequent wildlife use analyses will be entirely
unsuitable for any stand less than 5 a. (2 ha). The
system becomes more suitable as stand size
approaches 25 a. (10 ha; larger sample of patches).
For example, within vegetation classified as scrub
habitat with patches of shrubs, vertical configuration
in the various subplots more frequently resembles
this habitat type (Table 7) than any other habitat
type. However, in the strictest sense, it is not
appropriate to give structural type designations to
patches constituting a certain structural type.
Structural types were defined on the basis of
transects with similar average vertical foliage
distributions. The foliage distributions were
determined on the basis of measurements taken in
subplots. That it is wrong to classify subplots (or any
area 25 a. [10 ha]) can be seen from the following
analogy.

Suppose that all books in six private libraries
were measured and mean heights were found to be
8.7, 9.1, 9.4, 9.8, 10.2, and 10.6 in. (22, 23, 24, 25,
26, and 27 cm) for collections I through VI, respec-
tively. It would be improper to conclude from exam-
ining the books in collection IV that those books
< 94 in.(< 24 cm) in height came from collections
I, II, or III, or that all books > 9.8 in. (> 26 cm) in
height came from collections V and VI. Desirable
as it might be to know the origin of the books in
collection IV, such a determination simply cannot be
made from the evidence presented.

We present the data in Table 7 merely to em-
phasize that ( 1 ) there is heterogeneity between sub-
plots and (2) it is not valid to obtain foliage
measurements for a 5 a. (2 ha) plot and then be able

to determine its vegetation type. This would be anal-
ogous to having a book 10 in. (26.2 cm) in height
and concluding that it came from collection V. It
really could have come from any of the collections.
Desirable as it may be to classify a 5 a. ( 2 ha ) stand
and assess the wildlife use associated with it, such
a determination is not possible with the data pre-
sented. This must be compatible with the objectives
of the study. A classification at a smaller scale will
lead inexorably to a proliferation of vegetation types
recognized. This will be incompatible with the ob-
jective of emphasizing elements of similarity be-
tween stands rather than differences. More
importantly, we have learned that classification at a
smaller scale can lead to a cloudy or erroneous pic-
ture of how wildlife uses vegetation.

Cloudiness begins to appear at a scale of about
50 a. (20 ha; Anderson and Ohmart, unpubl. data),
and an opaqueness emerges from analyses at a scale
of about 5 a. (2 ha; Rosenberg 1980; Engel-Wilson
1982; Anderson and Ohmart, unpubl. ms), i.e., only
weak wildlife use patterns are discernible af a scale
of < 50 a. (< 20 ha), and either wrong impressions
or no impression emerges at a scale of < 10 a.
(< 2 ha). Investigators working in other habitats
have reported similar findings (Wiens 1981; Wiens
and Rotenberry 1981a, b).

Some of our conclusions about wildlife use of
riparian vegetation along the lower Colorado River
at the vegetation (habitat) type scale have been ex-
perimentally tested and confirmed (Meents et al.
1982; Anderson and Ohmart, in press-a). Other tests
are in progress. It is critical to clearly define the
purpose for conducting a study and to decide on the
appropriate scale. A classification made for one pur-
pose will probably fail when used for other than
the intended purpose.

Our discussion thus far has centered on hetero-
geneity within habitats. Next, we consider the differ-
ences among a group of habitats.

Table 7. Subplots of six recognized structural types.

Structural Type

Subplots

Number

Percent of Subplots of Structural Type

II

IV

VI

I

IV
V
VI

Open gallery

Closed gallery

Scrub thickets

Scrub thickets with patchy shrubs

Shrubs-scrub thickets

Shrubs, annuals, grasses with scattered trees

36
30
154
366
291
279

38

31

9

9

13

0

0

63

20

17

0

0

6

9

66

11

9

0

1

4

23

38

19

15

0

0

8

4

59

30

0

0

0

8

18

75

NOTE: Variation in vertical configuration among subplots within each structural type is indicated by the following data. See text for
discussion of tautology in this type of analysis.

Vegetation

655

ANALYZING HETEROGENEITY AMONG
HABITATS

Although many of the differences between two
habitats may be obvious to the observer (a patchy
salt cedar-scrub thicket is obviously different in
many ways from a cottonwood-willow gallery for-
est), the differences need to be quantified. Although
one may adequately describe the differences be-
tween two habitats, such a description may require
several pages and cannot be readily used in statistical
treatments. Therefore, differences must be expressed
quantitatively. Among the community attributes
measured, several may be intercorrelated, i.e., as the
value for one increases, the value for another also
increases. When colinearity exists between variables,
determining the extent to which either variable is
associated with wildlife will not be possible without
additional experimentation or without data from
some area where these variables are not confounded.
In such situations, a species or group of species
could be significantly associated with both variables.
In reality, one of the variables may be attracting
the species and the other one may not.

Such situations are probably inevitable with any
relatively large data set, and little can be done, ex-
cept to recognize that some constellations of com-
munity attributes are positively associated with
particular species or groups of species. Only care-
fully designed experiments will divulge which attri-
butes among the constellation are really attracting
wildlife. Analysis of a data set can be greatly simpli-
fied, however, if a constellation of intercorrelated
variables can be quickly recognized and treated as a
single variable.

Principal component analysis (PCA) is a statisti-
cal tool that combines intercorrelated variables into
new derived variables. We have used the technique
extensively; details are given elsewhere (Anderson
and Ohmart 1984b). The derived variables can usu-
ally be readily interpreted, and they can be treated
as independent variables in subsequent analyses.
Each habitat receives a score from roughly — 3 to
+ 3 on each derived variable. In our study, PCA
yielded four derived variables. The first included foli-
age density and diversity measures above the lowest
layer and FHD. Wildlife associated with such a de-
rived variable is most abundant in habitats with
dense foliage that is horizontally and vertically di-
verse. Managing wildlife associated with this derived
variable involves creating dense and diverse areas.

The second derived variable was bipolar, i.e.,
the number of honey mesquite per unit area was as-
sociated positively with this component, and the
number of salt cedar was associated negatively. Spe-
cies associated with this derived variable were posi-
tively associated with honey mesquite but negatively
with salt cedar. Negative associations indicated a

positive association with salt cedar, but no associa-
tion or a negative one with honey mesquite. Manage-
ment, in the first instance, might involve planting
honey mesquite trees. The other two principal com-
ponents were interpreted in similar ways.

PCA can be used to compress a large and com-
plex set of measurements (vegetation community
attributes) into a small set of derived variables that
can be used as independent variables. Associations
between wildlife and the attributes of the habitats
can be determined by using the wildlife populations
associated with various habitats in conjunction with
the score of that habitat for each of the derived vari-
ables. Techniques such as analysis of variance, simple
linear correlation, and multiple regression are appro-
priate for quantifying the extent of such associations.

Cluster analyses can be used to group transects
with similar vertical configurations. These clusters
can be further subdivided according to the numeri-
cally dominant vegetation present. By recognizing
relatively few vertical configurations (stress similari-
ties rather than differences, for example, 6) and rela-
tively few subdivisions by dominant vegetation
(again, 6), one can define up to 36 different habitats,
all of which differ from one another by vertical con-
figuration, dominant vegetation, or both. Recognizing
relatively few vertical configurations (3 to 9) and
having relatively few different dominant vegetation
types (3 to 9) are aids to the memory. Most people
can keep five or six definitions in mind. Thus, when
looking at a stand of vegetation, the manager needs
to answer only two questions to classify the stand:

( 1 ) What is the vertical configuration of the vege-
tation, i.e., is it four-layered, three-layered,
etc.?

(2) What plant species appear to be numerically
dominant in the stand?

Thus, in a short time, with a classification
scheme such as the one described here, the manager
can acquire enough general information about the
stand to describe it in detail in terms of the classifi-
cation system. This augments communication and
may be useful in making management decisions
when such decisions need to be made quickly.

DETERMINING WILDLIFE-HABITAT
ASSOCIATIONS

One can look for relationships between a given
species and the vegetation attributes. Conceivably,
every species in the habitats studied could be in-
volved in such analyses. This information is often
useful when the manager has collected information
on a single species that is of particular management

656

Vegetation

interest. For example, the Yuma clapper rail (Rallus
longirostris yumanensis), which is on the endan-
gered species list, is of great interest to biologists
along the Colorado River. We found (Anderson and
Ohmart 1985) that in spring, the Yuma clapper rail
was linearly associated with the first principal com-
ponent of an analysis of marsh habitat attributes. The
marsh habitats were analyzed according to the
method described above. The habitat breadth of
various species can be used to identify habitat spe-
cialists or those species with narrow habitat
breadths. The data in Tables 8 and 9 identify Bell's
vireo (Vireo bellii), summer tanager (Piranga
rubra), and yellow-billed cuckoo (Coccyzus ameri-
canus) as habitat specialists and cottonwood-willow
woodlands as their "preferred" habitats (Meents et
al. 1984).

Variables such as bird species diversity (BSD)
and species richness may also be related to the de-
rived variables. We found significant relationships
during most seasons for BSD and species richness
using various multivariate techniques (Rice et al.
1983). Densities of birds in various species groups,
such as permanent resident insectivores, may also be
related to the derived variables during one or more
seasons.

Here we add a cautionary note. Analyses should
be conducted over several seasons and years. We
have found that analyses over just one season or
over several seasons for only 1 or 2 years can lead to
egregious errors in assessing habitat importance to
wildlife (Rice et al. 1980, 1983, 1984).

Quite frequently, multiple linear regression is
used as an analytical tool. However, not all meaning-
ful relationships between wildlife and habitat attri-
butes would be linear, i.e., as the amount of the
habitat variable increased, the wildlife community
attribute would increase or decrease. For example,
we found that curvilinear relationships accounted for
a significant amount of the variance for 19 common
species of riparian birds.

We quantified deer use of habitats by finding
deer use areas (foraging, resting, fawning, etc.)
and analyzing the vegetation in plots 99 x 99 ft
(30 x 30 m). These were subdivided into four sub-
plots. Vegetation measurements and tree counts
were taken within these subplots in the manner de-
scribed above. In addition, measurements were taken
in a series of randomly selected plots. Attributes of
the vegetation in the deer use plots were then com-
pared statistically with the randomly selected control
plots (Haywood et al. 1984).

Pinyon- juniper/sagebrush association showing deer feeding area next to cover or resting areas.

Vegetation 657

>

u

o

"S

Ul

o

— ^

o
u

I

u
•S

"8

c

3

E

a

o
<J

c
a

a

T3

O
O

.C
O

c
u

T3

U

SP

U
3

X

1

3

rr

CD

>

^

>^

>

CD

c

o

X

—

CO
CD
O
Q)
Q.
CO

m

•<t ncoowo -t co co
oj *- cvj r-~ cm

o
cvj

^ ^ cm n- o ■>- t- cm

CM t- t-

CD

*3-

CO
CO

i- co co in o

T- ■*

ONCDlO

T- ■*

IflNOOOOr-
t- CO

CM CM CO CD CO

1- i- T- CM

CO ID CM
LO CM

LO Tt O

"fr CM

(COXO
O CM

CO h- t-
*- <tf CO

CM CO CM N- N-
CM CM

r*« ■* i- 1^. N.
CM CM

t- Tt tJ- ■■- |-~ |^

CM

CD
CO

CO
CO

O

CO

lO h- CM CM CM
CM CM

CM CM

>» t- i- CD CD
CO CO

-3- CD CM

CM i-

t- LT) O

CO tt-

CM ■* CD
CO

CO CM CM

LO CD in

CM r-

COO)*

O)

h» oo co ■«*• i—

CO

oso

t- ■* CM

CD

■T—

•* r^ •>- co f-

co co co

CM CM

CM t-

C0 CD CO CM O CDCMCO

CO

O CO

co coun-^-T-o m in o
CM t- ^ ^t

CO CO ^ -r- CM IflCDN

CO i— CO i—

in CM CO N- CD CO -^ r- co

cm t- wr-

en CMin-r-l^-CO LO h- CM
CO CM CO t-

CO

sconot

CM

CO O t-
r- 00 CM

CO

in

I"-- CD CD i— 00

T- CO T-

CM •'J CM

o

O CM t- CM CO

x- CO **

CO CO CO

CM

CO

o o cm i- in

CO

CO i- CD

CM CM

>- CD

CD ~ ^ "O i_ CD

dj g3-c ra o id c) >
2 S £- nuo ± -o.e

£o c « ~ c >-o o !=t5

| i£Msi*liii

t-CDOCOOO O t- O

T- CM T- ■*

■tf CO O CM CM O O Is- O

oiootco

O O O CO CO o

COCJO^-i-O O CD O
t- CM t- CM

TTWi-como m o co

CM "tf CM CO

osomioo

T~

o

CM

co co o m cm o

o

CO

CO t- O CO CO o

o

CO

o

CD CO O CO CM T-
CM r-

'r~

CD
CD

COCOOtlfiO

T- ■* T-

o

co

CM

CO CO O CO CD CO

^- 1-

CM

r^

**

c

CD

w

CD

c
o

CD
J» CD-o
- O ro CD "O

CD -c C ~ ,_ CD

O 2 ££ g>

cDcoocoino in cm o

t- t- CM

NCO r- CDCDO t- CM O

CMOOCDCOO CM CO

-r- t- y- t- ^

CONONNO CD in

T- T" t^

o-^-incocoo CO CO CM
cm t- t- ,- r»

co

cDcoommo ■>- in o

T- T- CM

CMCDOincDO ^r ■* t-

T- T- T- CM

CM CM CO N- O CM CO r)"
t- t- i- CM CO

r^ co co m cd co o m in

x- x- t- CM t- CO

t-tj-ooocdo co co in

t- t- <M i- t- y- CD

inTt-O-r-Oh- O CD CD

CO t- CM 'vf

m *- <D CD

2 V& o o

CO CD -S

• CD j

- CO

ro^tocD-e^^2f
ooo.^coEcd0

^50O5>5CDpc|:

< CQ CO _l ZC0 CO

O CD

■D

CD i ^

w < o "

"° "CD

>-

Tl

,

CD
0)

TO
"D

5

o

o

co

TO

w

II

II

rr

I

<

(Xi

658

Vegetation

Table 9. Habitat breadth of summer bird species in
riparian communities of the lower Colorado
River Valley.

Habitat

Percent of Maximum

Bird Species

Breadth*

Habitat Breadth

Resident

Abert towhee

1.28

0.94

Black-tailed

gnatcatcher

1.28

0.94

Crissal thrasher

1.30

0.96

Cactus wren

1.20

0.88

Gambel's quail

1.30

0.96

Gila woodpecker

0.95

0.70

Ladder-backed

woodpecker

1.20

0.88

Mourning dove

1.23

0.90

Verdin

1.32

0.97

Nonresident

Ash-throated

flycatcher

1.31

0.96

Blue grosbeak

1.33

0.98

Bell's vireo

0.62

0.46

Lucy warbler

1.26

0.93

Northern oriole

1.26

0.93

Summer tanager

0.66

0.49

Brown-crested

flycatcher

0.98

0.72

White-winged dove

1.12

0.82

Yellow-billed

cuckoo

0.86

0.63

"Habitat breadth calculated using thenequation:
HB = -Sp, log10 p,

where p, =

Habitat Breadth Mean =
Standard Deviation =

the proportion of a species'
population occurring in community.
1.137
0226

We studied lizard use of the landscape over a
heterogeneous 75 a. (30 ha) area by scattering plots
measuring 10 x 10ft(3 x 3m). These were subdi-
vided into 1 x 1 yd (m) subplots. Attributes of the
vegetation were then determined by the methods
described above. The nature of the substrate (sand,
hardpan, etc. ) was also noted. These plots were vis-
ited daily at the time of peak lizard activity; the num-
ber of each species detected was recorded. The
characteristics of the landscape could then be associ-
ated with greatest numbers of detections of various
lizard species (Anderson and Ohmart 1982). Data
obtained from can traps were used to corroborate or
refute observational data.

Rodent association with various habitats was
determined using multiple regression analysis where
relative densities of each rodent species were the
dependent variables and the vegetation factor scores
for each habitat (determined from PCA) were the
independent variables. Curvilinear relationships were
also considered (Anderson and Ohmart 1984a).

Finally, we determined the impact of clearing
vegetation from areas by first obtaining the vegeta-
tion attributes of control and experimental areas
(those to be cleared) before clearing. The impact of
clearing, according to different patterns and
amounts, was determined for each vegetation attrib-
ute separately (FHD, foliage density, etc.). The con-
trols were used to indicate the extent of change
when no clearing was done. We also noted the
change in principal component factor scores for
each of the habitats affected. Bird and rodent num-
bers were obtained before and after clearing in ex-
perimental and control areas in order to assess
impacts on wildlife in various habitats (Anderson and
Ohmart, in press-a).

Vegetation

659

LITERATURE CITED

ANDERSON, B.W. and R.D. OHMART. 1978. Phainopepla
utilization of honey mesquite forests in the Colorado
River Valley. Condor 80:334-338.

and . 1982. Revegetation for wildlife en-
hancement along the lower Colorado River. Final Rep.
to U.S. Dep. Inter., Bur Reclamation, Lower Colorado
Region, Boulder City, NV.

and . 1984a. A vegetation management study

for enhancement of wildlife along the lower Colorado
River. Final Rep. to U.S. Dep. Inter., Bur. Reclamation,
Lower Colorado Region, Boulder City, NV.

and . 1984b. Vegetation community type

maps: Lower Colorado River. U.S. Dep. Inter., Bur. Rec-
lamation, Lower Colorado Region, Boulder City, NV.

and . 1985. Habitat use by clapper rails in

the lower Colorado River Valley. Condor 87:1 16-126.

and . In press-a. Evaluation of the impact

of vegetation removal. Final Rep. to U.S. Dep. Inter.,
Bur. Reclamation, Lower Colorado Region, Boulder
City, NV.

and . In press-b. Riparian vegetation as a

mitigating process in stream and river restoration in
Gore, J.A., ed. The Restoration of Rivers and Streams.
Butterworth Publ., Stoneham, MA.

and W.C. HUNTER. 1983- Quantifying variables for

classifying desert riparian vegetation. Pages 34-44 in
Moir, W.H. and E.L. Hendzel, tech. coords. Proc.
Workshop on Southwestern Habitat Types. U.S. Dep.
Agric, For. Serv., Albuquerque, NM.

CODY, ML. 1974. Competition and the structure of bird
communities. Monogr. Population Biol. 7, Princeton
Univ. Press, Princeton, NJ.

CONINE, K. 1982. Avian use of honey mesquite, interior
and agricultural edge habitat in the lower Colorado
River. M.S. Thesis, Arizona State Univ. Tempe.

DAUBENMIRE, R. 1968. Plant communities: A textbook of
plant synecology. Harper and Row, Publishers. New
York, NY. 300pp.

ENGEL-WILSON, R.W. 1982. Comparison of bird commu-
nities in salt cedar and honey mesquite along the
lower Colorado, Verde, and Gila rivers. M.S. Thesis,
Arizona State Univ., Dep. Zoology, Tempe.

GREEN, R.H. 1979. Sampling design and statistical meth
ods for environmental biologists. John Wiley & Sons,
New York, NY.

HAYWOOD, R.J., B.W. ANDERSON, and R.D. OHMART.
1984. Habitat use by four radio-collared mule deer
along the lower Colorado River. U.S. Dep. Inter., Bur.
Reclamation, Lower Colorado Region, Boulder City,
NV.

HORN, H.S. 1966. Measurement of "overlap" in compara-
tive ecological studies. Am. Nat. 100:419-424.

KUCHLER, AW. 1967. Vegetation mapping. Ronald Press,
New York, NY.

MACARTHUR, R.H. and J.W. MACARTHUR 1961. On bird
species diversity. Ecology 42:594-598.

MEENTS, J.K., B.W. ANDERSON, and R.D. OHMART. 1982.
Vegetation relationships and food of sage sparrows
wintering in honey mesquite habitats. Wilson Bull.
94:129-138.

, , and . 1984. Sensitivity of riparian

birds to habitat loss. Pages 619-625 in Warner, RE.
and K.M. Hendrix, eds. California Riparian Systems:
Ecology, Conservation, and Productive Management.
Univ. California, Berkeley.

MUELLER-DOMBOIS, D. and H ELLENBERG. 1974. Aims
and methods of vegetation ecology. John Wiley and
Sons, New York, NY. 547pp.

OHMART, R.D., WO. DEASON, and C. BURKE. 1977. A

riparian case history: The Colorado River. Pages 35-47
in Johnson, R.R. and DA. Jones, tech. coords. Impor-
tance, Preservation, and Management of Riparian
Habitats: A Symposium. U.S. Dep. Agric, For. Serv.
Gen. Tech. Rep. RM-43-

PLATTS, W.S. 1981. Stream and inventory garbage in —
reliable analysis out: Only in fairy tales. Pages 75-85
in Armantrout, N.B., ed. Acquisition and Utilization of
Aquatic Inventory Information. Western Div., Amer.
Fish. Soc.

, W.F. MEGAHAM, and G.W. MARSHALL. 1983.

Methods for evaluating stream, riparian, and biotic
conditions. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
138.

RICE, J., B.W. ANDERSON, and R.D. OHMART. 1980. Sea-
sonal habitat selection by birds in the lower Colorado
River Valley. Ecology 61:1402-141 1.

, , and . 1983- Habitat selection attri-
butes of an avian community: A discriminant investi-
gation. Ecol. Monogr. 53263-290.

, , and . 1 984. Comparison of the

importance of different habitat attributes to avian
community organization. J. Wildl. Manage. 48:895-
911.

ROSENBERG, K.V. 1980. Breeding bird community orga-
nization in a desert riparian forest. M.S. Thesis, Ari-
zona State Univ., Dep. Zoology', Tempe.

SHANNON, C.E. and W. WEAVER. 1949. The mathematical
theory of communication. Univ. Illinois Press, Urbana.
125pp.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF

LAND MANAGEMENT. 1985. Rangeland Monitoring—
Trend Studies. U.S. Dep. Inter., Bur. of Land Manage.
Technical Reference 4400-4. 1 30pp.

WHITTAKER, R.H. 1975. Communities and systems, 2nd
edition. MacMillan and Co. New York, NY.

WIENS, J.A. 1981. Scale problems in avian censusing. Pages
513-521 in Ralph, C.J. andJ.M. Scott, eds. Estimating
Numbers of Terrestrial Birds. Studies in Avian
Biology.

and J.T. ROTENBERRY. 1981a. Censusing and the

evaluation of avian habitat occupancy. Pages 522-532
in Ralph, C.J. and J.M. Scott, eds. Estimating Numbers
of Terrestrial Birds. Studies in Avian Biology.

and . 1981b. Habitat associations and com-

munity structure of birds in shrub-steppe environ-
ments. Ecol. Monogr. 51:21-41.

660

Vegetation

32

MACRO-
INVERTEBRATES

Fred Mangum

U.S. Forest Service
Aquatic Laboratory
Brigham Young University
Provo, UT 84602

Editor's Note: Macroinvertebrate surveys are one of
the more specialized habitat analysis techniques.
Macroinvertebrate analyses could, in theory, be
used on any type of habitat; in practice, however,
the technique is used primarily for stream habitat.
In aquatic habitats, macroinvertebrate analyses
have proven to be good indicators of habitat condi-
tion and water quality. In the future, such analyses
may be used for other purposes, such as analyzing
suitability of marshes for waterfowl.

INTRODUCTION

Macroinvertebrates are those invertebrates that
can be detected with the unaided eye. Although
macroinvertebrates are an important habitat compo-
nent for many terrestrial vertebrates, such as insecti-
vorous birds, terrestrial macroinvertebrates are
rarely sampled as part of terrestrial habitat inventory
and monitoring programs.

On the other hand, macroinvertebrates in the
aquatic environment provide a link in the food chain
between microscopic, multicelled organisms and
fish populations. They are essential to the growth
and production of fish and, because of their strict
habitat requirements, are very useful indicators of
aquatic habitat changes. The number, size, and spe-
cies of aquatic invertebrates are important to fisher-
ies habitat, as they are the primary food source for
most salmonids and warm-water fish. Aquatic inverte-
brates include insects, such as mayflies, stonefiles,
caddisflies, and diptera (two-winged flies); crusta-
ceans, such as crayfish and shrimp; mollusks, such as
snails and clams; and fresh-water earthworms.

Macroinvertebrate data and many other types of
scientific information should be considered by land
managers in the decisionmaking process. Inventory
data covering physical, chemical, and biological
water quality and aquatic habitat factors can be ana-
lyzed and interpreted in a manner that can be under-
stood by interdisciplinary team members and
planners. Current environmental disturbances should
be identified as well as potential problems that could
develop if new management activities adversely af-
fect a stream. When linked to other factors, macroin-
vertebrate data can assist the decisionmaker in
assessing impacts to stream ecosystems from land
management activities.

Positive or negative effects from almost every-
thing that occurs in a drainage show up in streams
and affect the macroinvertebrate community compo-
sition. The community composition can reveal the
health of an ecosystem, provide a warning system, or
show the relative intensity of pollution problems.
Macroinvertebrate analysis often identifies types of
pollution, which is valuable information for consci-
entious land managers. To make use of all available

Macroinvertebrates

661

physical and biological water quality and aquatic
habitat data, there must be full coordination among
specialists and management.

Data about macroinvertebrates are site-specific.
Aquatic ecosystems cannot be managed on the basis
of average values over large diverse areas. Particu-
larly for grazing prescriptions, time and space con-
siderations within a watershed must be considered
in the planning process if high standards of aquatic
habitat and water quality are to be maintained.

The U.S. Bureau of Land Management (BLM) and
the U.S. Forest Service have used macroinvertebrate
data to assess habitat quality and nutrients available
for resident and anadromous fisheries. Effects from
livestock grazing, dam building, mining, timber har-
vest, rotenone projects, instream improvement struc-
tures and exclosures, and sanitary facilities in
campgrounds can be measured.

This chapter describes procedures for sampling
aquatic macroinvertebrates and analyzing resulting
data. These procedures are primarily used for stream
surveys; however, parts of the procedure may be
used for lake, pond, and reservoir surveys.

MACROINVERTEBRATE SURVEYS FOR
STREAMS

Procedure

Surveys of macroinvertebrates consist of distinct
phases: field sampling, laboratory analysis, and data
analysis. The field biologist usually does the field
sampling and sends samples to an aquatic laboratory
for analysis. Field sampling methods are described
here as well as data analysis and interpretation since
many of the latter techniques are specific to aquatic
macroinvertebrates.

Sampling Design. When designing a stream survey
for macroinvertebrates, consult the aquatic
specialists responsible for habitat management. This
will ensure that the data and information generated
will be useful in answering pertinent management
questions.

A minimum of three stratified, random aquatic
macroinvertebrate samples should be taken at each
station, as described in the techniques section. A
station is defined as a stream reach up to 30 m (100
ft) long. The samples generally provide desired com-
munity representation and data that can be statisti-
cally analyzed. Four samples may be needed for
some streams, depending on the statistical reliability
desired.

The macroinvertebrate community can reveal
the quality of habitat components essential to

aquatic fauna, such as water quality, substrate com-
position, riparian habitat quality, ecosystem stability,
and past history.

Criteria for sampling station selection, duration,
and frequency differ depending on the basis of in-
ventory or monitoring emphasis. For example, if an
action plan was developed to evaluate a new grazing
prescription and an exclosure, three stations should
be established — above, within, and below the pro-
posed exclosure area. The "above" station should be
within 50 yards of the exclosure; the "within" sta-
tion, on the lower end of the exclosure; and the
"below" station, with gradient and other natural
physical characteristics similar to the "within" sta-
tion, generally within 1.6 km (1 mi.) of the exclo-
sure. Spring, summer, and fall samples should be
taken before installing the exclosure; sampling on
those dates should be repeated for 3 to 5 years fol-
lowing exclosure installation.

When inventorying physical, chemical, and bio-
logical parameters, I recommend establishing a mini-
mum base period of 3 years to show effects of
natural extremes in physical and water chemistry
influences on an ecosystem, and a minimum fre-
quency of three dates annually — spring, summer, and
late fall. Additional samples may be taken to cover
projected heavy-use periods.

For monitoring or surveillance of the effects of a
management activity, a minimum of 1 year is neces-
sary, but up to 3 or 5 years can be required depend-
ing on the activity or pollution source being
monitored. The location and number of sampling
stations depend on the management activity being
monitored and its duration or projected commence-
ment. Monitoring requires a minimum of two sta-
tions— one above (a control) and one below the
activity monitored. A third station downstream helps
define downstream impacts or recovery.

To more specifically locate a source of stream
stress, stations can be added within the management
area initially or upon detection of stress conditions.
Point or nonpoint pollution sources may be moni-
tored. To show the quality of the habitat and the
water as it leaves public land, management plans for
some streams may be adequately served by a single
station. Control stations should be placed in a riffle,
if possible, between 45 and 180 m (50 and 200 yd)
above the upper boundaries of the suspected pollu-
tion sources.

Experimental stations should include at least
one station below the lower boundaries of the sus-
pected source. Stations may be placed in two cate-
gories depending on their projected use: ( 1 ) long-
term (permanent) stations established on the main
stream or major tributary in each drainage to moni-
tor long-term projects or continuing land-use activi-

662

Macroinvertebrates

ties that, by their nature, could cause on-site or
downstream, accumulated impacts or stress or (2)
short-term (temporary) stations established to moni-
tor projects such as timber harvest, prescribed
burns, insect spray projects, or others having impacts
over a definite time period. All such monitoring
should include predisturbance (before), operation
(during), and post-operation (after) sampling.

Substrate Sampled. Samples are taken in riffle areas
with streambed substrate of gravel or rubble. A
majority of the species occurring in the stream's
aquatic macroinvertebrate community will be found
there. Replicate samples should be taken within a
100-foot stretch of a stream, with each sample being
taken in locations where velocity and depth are as
similar as possible.

Equipment. Equipment needed for macroin-
vertebrate sampling is listed below:

• Modified Surber Net

• 250-micron sieve

• Three plastic bottles per station with strip of
masking tape attached for identification pur-
poses

• Preserving solution

• Hip boots

• Saturated saltwater solution

• Two aluminum bread pans

• Waterproof gloves

• Laundry pen (waterproof marker)

Field Sampling.

Modified Surber Net Samples. The modified
Surber square -foot sample net ( Figure 1 ) has proven
to have a correlation coefficient equal to or better
than other currently used sampling devices. I recom-
mend using a 280-micron mesh, nylon net when
sampling all macroinvertebrates. Three steps should
be followed when using this modification.

1. The foot-square Surber frame is placed over the
gravel-rubble substrate in the stream with the
net downstream. As the rocks within the frame
are hand scrubbed, macroinvertebrates are car-
ried into the net by the flowing water. The sub-
strate underlying the gravel-rubble is also stirred
by hand to a depth of 7 to 10 cm (3 to 4 in.),

if possible.

2. After allowing the water to drain from the net,
the net is inverted into an aluminum pan con-
taining a saturated saltwater solution. As the
saltwater is poured into a second pan, the or-
ganic materials thus floated are caught in a 250-
micron mesh sieve. The saltwater is then poured
back into the first pan, the contents are again
vigorously stirred, and the floating materials and
specimens are poured for a second time through
the sample sieve. The sample may require siev-
ing two or three times.

3. The sample in the sieve is then washed from the
sieve pan into the sample bottle with an alcohol
solution. The alcohol solution is mixed at a ratio

Nitex 280 micron mesh netting

Viny

1"x 3/16" aluminum frame

Figure 1. Modified Surber sampler.

Macroinvertebrates

663

of 1 cup of 10% formalin to 38 L ( 1 gal.) of 70%
ethanol. Enough alcohol should be added to
the sample bottle to cover the sample. Caddis-
flies in their cases must be handpicked from the
pan and added to the sample.

Artificial Substrate Basket Samples. The use

of artificial substrate sampling baskets is also recom-
mended by the U.S. Environmental Protection
Agency because of their comparatively low coeffi-
cient of variation in collection of macroinvertebrate
individuals and taxa (see Figure 2).

The following methods are recommended for
basket sampling:

1 . Placement. The baskets, filled with two layers of
baseball-sized rubble gathered from the immedi-
ate stream area, are worked down into the sub-
strate in a riffle area among rocks of nearly the
same size, if possible. The rocks may be col-
lected along the bank of the stream or in the
stream and placed in the basket. If the rocks are
gathered from the stream substrate, they should
be scrubbed clean to remove any existing in-
sects. For statistical purposes, three baskets are
placed at each station. Each station should not
exceed a 100-foot distance between the up-
stream and downstream baskets.

2. Retrieval. After a 1 -month (not to exceed 33
days) colonization period, samples are collected
from the baskets. The baskets should be ap-
proached from their downstream side, the Sur-
ber net placed over the basket, and the basket
pulled into the net. With one hand on the bas-
ket and the other holding the net, both are
pulled from the stream in an upstream motion.
The rocks from the baskets are placed in a
bucket half filled with water. The Surber net is
inverted and washed in the bucket of water, and
the rocks are individually scrubbed with a
brush. After scrubbing, the rocks are placed
back into the basket that is then repositioned
into the stream. The basket should be covered
with water and camouflaged, if possible, to avoid
vandalism. The water and sample in the bucket
are poured into a 250-micron mesh sieve to
remove the water. If a lot of debris and sand are
present, the sample can be placed in a saturated
saltwater solution that will float all organisms
out of the sand and gravel. The sample is
washed into the sample bottle from the sieve
with alcohol. Alcohol is added to the bottle
until it just covers the sample.

Supporting Data. Several types of supporting
data are commonly collected with macroinvertebrate
samples. Water chemistry and physical habitat data
are integrated into the elevation to bring the total

Figure 2. Artificial substrate basket.

biological analysis into perspective for the biologist
and land manager. In particular, data on alkalinity,
sulfates, gradient, and substrate composition are nec-
essary for calculating a Biotic Condition Index (BCI).
The following data are commonly collected with
macroinvertebrate data:

1 . Specific Conductivity. Specific conductivity,
measured as umhos/cm at 25°C (77°F), is partic-
ularly important for streams with high alkalinity
and hardness.

Alkalinity. Alkalinity, measured as HC03 con-
centration in mg/1, is closely linked to stream
productivity and is necessary to determine the
BCI. Alkalinity correlates with community den-
sity, diversity, and biomass, and also affects the
primary production in a stream.

3. Water Temperature. Temperature is a limiting
factor on certain sites.

4. Nitrate Nitrogen. Nitrate nitrogen, measured as
mg/1, may indicate organic enrichment.

5. Orthophosphate. Orthophosphate, measured as
mg/1, may indicate organic enrichment.

6. Sulfate. Sulfate, measured as mg/1, may indicate
sources of accretion flows. It is a factor affecting

664

Macroinvertebrates

community composition and is necessary to
determine the BCI. An increase in sulfate
generally indicates a deterioration of natural
water quality. Many poor quality waters in
the western U.S. have high sulfate levels.

7. pH. High or low pH also limits biological
productivity.

Gradient. Gradient, measured as a percent, is
related to a stream's ability to maintain sub-
strate quality. It affects community composi-
tion and is used in BCI analyses because of its
positive correlation with macroinvertebrate
community diversity.

Elevation. Elevation is used to define ranges
for aquatic species.

10. Substrate Composition. Substrate composi-
tion, measured as the relative composition
(ranking) of streambed boulder, rubble,
gravel, and sand-silt materials, is necessary to
determine the BCI. Rubble substrates are
preferred by the greatest number of species
because they provide a greater diversity of
microhabitat than smaller substrates.

11. Streamside Vegetation. Streamside vegetation,
measured as the relative dominance (rank)
of grass, shrubs, brush, deciduous trees, and
evergreen trees present on streambank ripar-
ian areas, affects stream stability and nutrient
sources.

12. Flow Rates. Mean summer and mean low-flow
rates, measured in ft /sec, significantly affect
macroinvertebrate community structure.
Methods for measuring each of these varia-
bles are described in Chapters 28 and 30.

Sample Analysis. Analyses are based on several
factors such as the —

• Diversity and taxa index (DAT);

• Biomass of the macroinvertebrates per sample;

• Numbers of organisms per species;

• Species and trophic balance in the community;

• Abundance and degree of dominance among
indicator species; and

• Tolerances of species in the community, as indi-
cated by the community tolerance quotient
(CTQ).

The BCI suggests a management strategy that
integrates biological, chemical, and physical factors
present in and adjacent to the stream's environment.

Analysis Elements (Major). The following
three major analysis elements will be found on data
sheets included in reports, along with scales indicat-
ing the significance of these data values:

1. BCI. The BCI measures a stream against its own
potential and not that of other streams. The
BCI is based on mean community tolerance,
composed of the tolerances of individual taxa or
species found in the community, which varies
in response to intensity of perturbations in the
ecosystem.

The BCI is calculated as follows:

a. Determine

( 1 ) Gradient Percentage

( 2 ) Relative Substrate Dominance given a
ranking from 1 through 4
Boulder > 30 cm (over 12 in.)

Rubble 7.5-30 cm (3-12 in.)
Gravel 0.3-7.5 cm (1/8-2.99 in.)

Sand-silt < 0.3 cm (< 1/8 in.)

Analysis

Laboratory Analysis. Normally samples are sent to
regional laboratories for processing and analysis.
Samples should be properly labeled with agency,
administrative unit location, stream, station, and date.

Laboratory personnel analyze these data. These
analyses are described below since the field biolo-
gists are probably familiar with how they are derived
and interpreted even if they do not analyze the raw
data.

(3) Total Alkalinity (mg/1)

(4) Sulfate (mg/1 )

b. Determine, using the above information and
the key in Table 1 , the predicted community
tolerance quotient (CTQP) for the reach on
the study stream.

c. List, using at least three or four macroinverte-
brate samples at a station, the taxa and their
associated tolerance quotients (TQ)

(Table 2).

Macroinvertebrates

665

d. Add the TQ values and divide by the number
of taxa to get the actual community tolerance
quotient

The BCI is

STQs

CTQa

CTO
e. Determine the BCI = ^p x

L* 1 V^a

100

The scale of BCI values listed below can be
used as indexes or values to determine the
health of a stream ecosystem for which a de-
fined management strategy can be based for
each stream reach.

Scale

BCI

Excellent

Above 90

Good

75-90

Fair

Below 75

Poor

Below 75

Sensitive to all types of environmental stress;

Applicable to various types of streams;

Capable of giving a linear assessment from un-
stressed to highly stressed conditions;

Independent of sample size, providing the sam-
ple contains a representative assemblage of
species;

Based on data readily available or easily ac-
quired; and

Capable of meshing with and supporting exist-
ing stream habitat or water quality management
programs.

Dominance and Taxa Diversity Index (DAT).
This index combines a measure of species domi-
nance in the community with the number of
species present. Many other indexes measure
one or the other of these aspects.

Table 1. A key giving predicted community tolerance quotients (CTQp's) for various combinations of

gradient (%), substrates, total alkalinity (mg/1 CaC03), and sulfate (mg/1 S04) for any given stream (from
U.S. Department of Agriculture, Forest Service 1981).

Combinations

Goto
Key No.

CTQp

1 . Stream gradient 0.1—1.2

2

51

1 .3-3.0

15

>3.0

2. Substrate mostly boulder & rubble

Gravel & rubble

28

3

7

Sand & boulder,
rubble, or gravel

3. Total alkalinity 0-1 99

11

4

200-300

>300

4. Sulfate 0-1 49

5

6

150-300

71

>300

90

5. Sulfate 0-1 49

53

150-300

71

>300

90

6. Sulfate 0-1 49

90

150-300

96

>300

108

7. Total alkalinity 0-1 99

8

53

200-300

>300

9

10

8. Sulfate 0-149

150-300

85

>300

103

666

Macroinvertebrates

Table 1. A key giving predicted community tolerance quotients (CTQps) for various combinations of
gradient (% ), substrates, total alkalinity (mg/1 CaC03), and sulfate (mg/1 S04) for any given stream
(continued).

Combinations

Goto
Key No.

CTQP

9. Sulfate 0-1 49

55

150-300

. 86

>300

..103

10. Sulfate 0-149

89

150-300

97

>300

..108

1 1 . Total alkalinity 0-1 99

12

200-300

13

>300

14

12. Sulfate 0-149

60

150-300

. 90

>300

..108

13. Sulfate 0-149

60

150-300

90

>300

..108

14. Sulfate 0-149

90

150-300

. 99

>300

..108

1 5. Substrate mostly boulder & rubble

16

50

Gravel & rubble

Sand & boulder,
rubble, or gravel

1 6. Total alkalinity 0-1 99

20

24

17

200-300

>300

17. Sulfate 0-149

18

19

150-300

65

>300

90

18. Sulfate 0-149

50

150-300

65

>300

. 90

19. Sulfate 0-199

. 90

150-300

. 96

>300

108

20. Total alkalinity 0-1 99

200-300

>300

21 . Sulfate 0-1 49

21

22

230

50

150-300

. 80

>300

103

22. Sulfate 0-1 49

55

150-300

80

>300

108

23. Sulfate 0-1 49

80

200-300

96

>300

108

24. Total alkalinity 0-1 99

200-300

>300

25

26

27

Macroinvertebrates

667

Table 1. A key giving predicted community tolerance quotients (CTQp's) for various combinations of
gradient (% ), substrates, total alkalinity (mg/1 CaC03), and sulfate (mg/1 S04) for any given stream
(concluded).

Combinations

Goto
Key No.

CTQP

25. Sulfate 0-1 49

66

150-300

88

>300

..108

26. Sulfate 0-1 49

65

150-300

. 88

>300

1 08

27. Sulfate 0-1 49

85

150-300

93

>300

108

28. Substrate mostly boulder & rubble

Gravel & rubble

Sand & boulder,

rubble, or gravel

29. Total alkalinity 0-1 99

200-300

>300

30. Sulfate 0-1 49

29

33

37

30

31

32

50

150-300

62

>300

100

31 . Sulfate 0-1 49

50

150-300

62

>300

108

32. Sulfate 0-1 49

85

150-300

90

>300

108

33. Total alkalinity 0-1 99

200-300

>300

34. Sulfate 0-1 49

34

35

36

50

150-300

77

>300

108

35. Sulfate 0-1 49

50

150-300

77

>300

108

36. Sulfate 0-1 49

90

150-300

99

>300

108

37. Total alkalinity 0-1 99

200-300

>300

38. Sulfate 0-1 49

25

26

27

80

150-300

100

>300

1 08

39. Sulfate 0-1 49

80

150-300

.... 1 00

>300

..108

40. Sulfate 0-1 49

100

150-300

108

>300

1 08

668

Macroinvertebmtes

Table 2. Tolerance quotients (TQs) of aquatic macroinvertebrates based on tolerance to alkalinity, sulfate,
and sedimentation including low stream gradients (from U.S. Department of Agriculture, Forest Service
1981).

Taxa

TQ

Family Tipulidae

72

Antocha monticola

24

Dicranota sp.

24

Hexatoma sp.

36

Holorusia grandis

72

Helobia sp.

36

Tipula sp.

36

Family Psychodidae

36

Maruina sp.

36

Psychoda sp.

36

Pericoma sp.

36

Family Blephariceridae

2

Bibiocephala grandis

2

Agathon sp.

2

Family Deuterophlebiidae

4

Deuterophlebia coloradensis

4

Family Culicidae

108

Aedes sp.

108

Culex sp.

108

Anopheles sp.

108

Mansonia sp.

108

Psorophora sp.

108

Culiseta sp.

108

Family Dixidae

108

Dixa sp.

108

Family Simuliidae

108

Family Chironomidae

108

Family Ceratopogonidae

108

Family Stratiomyidae

108

Euparyphus sp.

108

Family Tabanidae

108

Tabanus sp.

108

Family Rhagionidae

24

Atherix pachypus

24

Family Dolichopodidae

108

Family Empididae

108

Hemerodromia sp.

108

Family Ephydrinae

108

Ep hydra sp.

108

Family Muscidae

108

Limnophora sp.

108

Family Syrphidae

108

Chrysogaster sp.

108

Tubifera sp.

108

Helophilus sp.

108

Family Dytiscidae

72

Derovatellus sp.

72

Laccophilus sp.

72

Bidessus sp.

72

Agabus sp.

72

Hygrotus sp.

72

Hydroporous sp.

72

Oreodytes sp.

72

Illy bi us sp.

72

Rhantus sp.

72

Dytiscus sp.

72

Acilius sp.

72

Cybister sp.

72

Deronectes sp.

72

Taxa

TQ

Family Dytiscidae (continued)

Thermonectes sp.

72

Coptotomus sp.

72

Family Hydrophilidae

72

Helophorus sp.

72

Hydrochara sp.

72

Berosus sp.

72

Enochrus sp.

72

Hydrophilus sp.

72

Tropisternus sp.

72

Hydrobius sp.

72

Paracymus sp.

72

Crenitus sp.

72

Ametor sp.

72

Helochares sp.

72

Laccobius sp.

72

Enochrus sp.

72

Cymbiodyta sp.

72

Family Elmidae

108

Zaitzevia sp.

Narpus sp.

Stenelmis sp.

Dubiraphia sp.

Optioservus sp.

108

Heterlimnius sp.

£/m/s sp.

Simsonia sp.

Microcylloepus sp.

Z_ara sp.

Family Gyrinidae

108

Gyrinus sp.

108

Family Amphizoidae

24

Amphizoa sp.

24

Family Hydraenidae

72

Order Diptera

Family Hydroptilidae

Leucotrichia sp.

108

Alisotrichia sp.

108

Mayatrichia sp.

108

Family Limnephilidae

108

Limnephilus sp.

108

Dicosmoecus sp.

24

Hesperophylax sp.

108

Oligophlebodes sp.

24

Apatania sp.

18

Amphicosmoecus sp.

18

Neothremma sp.

8

Lenarchus sp.

18

Chyranda sp.

18

Psychoglypha sp.

24

Ecclisomyia sp.

24

Homophylax sp.

18

Allocosmoecus sp.

18

Asynarchus sp.

108

Clistoronia sp.

108

Grammotaulius sp.

108

Imania sp.

48

Neophylax sp.

24

Onocosmoecus sp.

18

Pycnopsyche sp.

72

Macroinvertebrates

669

Table 2. Tolerance quotients (TQs) of aquatic macroinvertebrates based on tolerance to alkalinity, sulfate,
and sedimentation including low stream gradients (continued).

Taxa

TQ

Family Leptoceridae

54

Oecetis sp.

54

Leptocella sp.

54

Triaenodes sp.

54

Mystacides sp.

54

Ceraclea sp.

54

Family Lepidostomatidae

18

Lepidostoma sp.

18

Family Brachycentridae

24

Brachycentrus sp.

24

Micrasema sp.

24

Oligoplectrum sp.

24

Amiocentrus sp.

24

Family Helicopsychidae

18

Helicopsyche borealis

18

Family Polycentropodidae

72

Polycentropus sp.

72

Nyctiophylax sp.

72

Family Sericostomatidae

72

Gumaga sp.

72

Order Coleoptera

Family Haliplidae

54

Brychius sp.

54

Haliplus sp.

54

Peltodytes sp.

54

Family Periodidae

Pictetiella expansa

18

Diura knowltoni

24

Isoperla sp.

48

/. ebria

24

1. fulva

48

1. mormona

48

1. quinquepunctata

48

Family Chloroperlidae

24

Family Perlidae

24

Acroneuria abnormis

6

Claassenia sabulosa

6

Hesperoperla pacifica

18

Perlesta placida

24

Doroneuria theodora

18

Order Trichoptera

Family Rhyacophilidae

18

Rhyacophila sp.

18

Atopsyche sp.

18

Himalopsyche sp.

18

Family Glossosomatidae

32

Glossosoma sp.

24

Anagapetus sp.

24

Protoptila sp.

32

Culoptila sp.

32

Family Philopotamidae

24

Chimarra sp.

24

Dolophilodes (sortosa)

24

Wormaldia sp.

24

Family Psychomyiidae

108

Polycentropus sp.

108

Nyctiophylax sp.

108

Psychomyia sp.

108

Ti nodes sp.

108

Taxa

TQ

Family Hydropsychidae

108

Hydropsyche sp.

108

Cheumatopsyche sp.

108

Arctopsyche sp.

18

Smicridea sp.

72

Diplectrona sp.

48

Macronema sp.

48

Parapsyche sp.

6

Family Hydroptilidae

108

Hydroptila sp.

108

Agralyea sp.

108

Ochrotrichia sp.

108

Stactobiella sp.

108

Neotrichia sp.

108

Ithytrichia sp.

108

Oxyethira sp.

108

Order Ephemeroptera

Family Gerridae

72

Gem's sp.

72

Rheumatobates sp.

72

Family Naucoridae

72

Ambrysus mormon

72

Pelocoris sp.

72

Family Notonectidae

108

Notonecta sp.

108

Buenoa sp.

108

Family Veliidae

72

Microvelia americana

72

Rhagovelia distincta

72

Family Mesoveliidae

72

Mesovelia sp.

72

Family Macroveliidae

72

Macrovelia sp.

72

Order Plecoptera

Family Nemouridae

36

Amphinemura

6

Malenka

36

Prostoia besametsa

24

Podmosta

12

Zapada

16

Nemoura

24

Family Capniidae

32

Capnia

32

Eucapnopsis

18

Isocapnia

24

Mesocapnia frisonl

32

Utacapnia

18

Family Taeniopterygidae

48

Taenionema

48

Doddsia

24

Oemopteryx

48

Family Leuctridae

18

Paraleuctra sp.

18

Perlomyia sp.

18

Family Pteronarcyidae

24

Pteronarcella badia

24

Pteronarcys californica

18

P. princeps

24

Family Periodidae

48

670

Macroinvertebrates

Table 2. Tolerance quotients (TQs) of aquatic macroinvertebrates based on tolerance to alkalinity, sulfate,
and sedimentation including low stream gradients (concluded).

Taxa

TQ

Family Perlodidae (continued)

Megarcys signata

24

Skwala parallela

18

Cultus aestivalis

12

Isogenoides sp.

24

/. elongatus

24

1. zionensis

24

Kogotus modestus

18

Order Megaloptera

72

Family Sialidae

72

Sialis sp.

72

Family Corydalidae

72

Corydalus cognata

72

Order Lepidoptera

72

Family Pyralidae

72

Paragyractis kearfottalis

72

Order Ephemeroptera

72

Family Siphlonuridae

72

Ameletus sp.

48

Siphlonurus occidentalis

72

Isonychia

48

Family Baetidae

72

Baetis sp.

72

Callibaetis sp.

72

Pseudocloeon sp.

72

Centroptilum sp.

36

Dactylobaetis sp.

36

Paracloeodes sp.

72

Family Oligoneuriidae

36

Lachlania saskatchewanensis

36

Homoeoneuria sp.

36

Family Heptageniidae

48

Heptagenia sp.

48

Stenonema sp.

48

Cinygmula sp.

21

Rhithrogena sp.

21

Epeorus sp.

21

Anepeorus sp.

48

Family Leptophlebiidae

36

Paraleptophlebia sp.

24

Leptophlebia sp.

24

Choroterpes sp.

36

Traverella sp.

36

Family Tricorythidae

108

Tricorythodes sp.

108

Leptohyphes sp.

72

Family Ephemerellidae

48

Ephemerella sp.

48

E. grandis

24

E. doddsi

48

E. coloradensis

18

E. tibialis

24

Taxa

TQ

Family Ephemerellidae (continued)

E. inermis

48

E. infrequens

48

E. spinifera

24

Family Ephemeridae

36

Ephemera simulans

36

Hexagenia limbata

36

Phylum Coelenterata

108

Class Hydrozoa

108

Phylum Aschelminthes

108

Class Nematoda

108

Phylum Mollusca

108

Class Gastropoda

108

Family Lymnaidae

108

Lymnaea sp.

108

Family Physidae

108

Physa sp.

108

Family Planorbidae

108

Phylum Annelida

108

Class Hirudinea

108

Class Oligochaeta

108

Family Tubificidae

108

Tubifex sp.

108

Family Lumbricidae

108

Lumbricus aquaticus

108

Phylum Platyhelminthes

Class Turbellaria

108

Order Tricladida

108

Phylum Arthropoda

Class Arachnida

Suborder Hydracarina

108

Class Crustacea

108

Order Isopoda

108

Family Asellidae

108

Asellus sp.

108

Order Amphipoda

108

Family Talitridae

108

Hyalella azteca

108

Family Gammaridae

108

Gammarus lacustris

108

Order Decapoda

108

Family Astacidae

108

Pacifastacus gambeli

108

Cambarus laevis

108

Order Cladocera

108

Daphnia sp.

108

Order Copepoda

108

Order Ostracoda

108

Class Insecta

108

Order Collembola

108

Family Poduridae

108

Podura aquatica

108

Family Entomobryidae

108

Macroinvertebrates

671

GASTROPODA

Lymnaea

AMPHIPODA

Gammarus

OLIGOCHAETA

Lumbricus

HYDRACARINA

Mideopis

LEPIDOPTERA

Paragyractis

TRICLADIDA

Planaria

MEGALOPTERA

Sialis

EPHEMEROPTERA

Stenonema

Figure 3. Common aquatic macroinvertebrate orders used in ecosystem analysis.
672 Macroinvertebrates

DIPTERA

Psychoda

TRICHOPTERA

Helicopsyche

PLECOPTERA

Isoperla

COLEOPTERA

Gyrinus

DIPTERA

Simulium

HEMIPTERA

Lethocercus

DIPTERA

Chironomus

Figure 3. Common aquatic macroinvertebrate orders used in ecosystem analysis (concluded).

Macroinvertebrates

673

The procedure for determining DAT is as
follows: the dominance part of the index is de-
termined by placing a sample of aquatic insects
in a petri dish with lines marked on the bottom.
Each species of insects viewed through a dis-
secting microscope can be identified if the dish
is moved in a snake-like fashion so the same
organisms are not viewed twice. Each organism
and series can be counted. A series is recorded
each time a sighted species is different from the
previous one. If the sample has a sufficient num-
ber of organisms, 300 to 400 should be ob-
served to determine the diversity index value.
The number of series (s) is divided by the num-
ber of organisms (o) to determine the domi-
nance value (Dom. = s/o), which will be a
fraction between 0 and 1 . The number of taxa
in the sample is multiplied by this dominance
value to obtain the DAT Diversity Index
(Dom. x Taxa = DAT).

DAT Scale

Stream Value

18-26

Excellent

11-17

Good

6-10

Fair

0-5

Poor

3. Standing Crop. The dry weight of the organisms
is determined for each sample. Small aluminum
pans are predried for 8 to 24 hours and are
weighed to the nearest tenth of a milligram on a
Mettler balance. The samples placed in these
pans are dried at a temperature of 65°C (149°F)
for 8 to 24 hours; then the pan and its contents
are weighed. This is important in indicating
whether a stream is reaching its potential and
shows the stream's potential for supporting a
resident fishery. Figure 3 shows some of the
more common aquatic insects used in aquatic
ecosystem analysis.

Analysis Elements (Minor). The following
summary analyses are helpful in evaluating aquatic
ecosystems:

1. Total Number of Species. Total number of spe-
cies is an indication of macroinvertebrate com-
munity diversity. It often gives an initial
indication of the stability of the environment or
of the community within the environment.

2. Mean Number of Organisms Per Taxa Per
Square Meter. This value is an indication of the
stability of the community, habitat, and water
quality. Taxa may be transient individuals or
resident populations that may indicate whether
or not conditions are favorable.

3. Standard Error of the Mean. This is a measure
of similarity in samples which indicates when
enough samples have been taken to obtain the
desired reliability. According to the U.S. Envi-
ronmental Protection Agency standards (Weber
1973; Green 1979), this value should not be
under 20% ; on most streams, this can be
achieved with three samples. Some streams re-
quire four or five samples. This value shows
when the mean is within 80% confidence limits.

4. Coefficient of Variation Coefficient of variation
is a value that indicates variability in samples,
independent of the number of samples. It evalu-
ates the sampling technique and effectiveness
of the sample equipment. The U.S. Environmen-
tal Protection Agency has set the standard at
under 50% for acceptable reliability.

5. Standard Deviation. Standard deviation meas-
ures sample variation and indicates whether
sampling techniques and number of samples are
sufficient to effectively show community struc-
ture.

DISCUSSION

Macroinvertebrate data, when combined with
data on aquatic features and water quality, can be
useful to land managers in managing aquatic re-
sources and surrounding areas. Positive or negative
effects from almost every activity that occurs in a
drainage are eventually reflected in the macroinver-
tebrate community composition of the streams that
drain it. This community composition can reveal the
health of the ecosystem, provide a warning system,
or reflect the relative intensity of pollution problems.
Macroinvertebrate analysis often can identify the
type(s) of pollution, making it a valuable tool for
conscientious land managers.

Macroinvertebrate data are site-specific and av-
erage values for large diverse areas should not be
used for decisions. Time and space considerations
are very important in maintaining high standards of
aquatic habitat and water quality, particularly in for-
mulating grazing prescriptions. Monitoring plans
for such prescriptions need to incorporate the same
considerations of time and space.

Finally, the biologist and manager should be
cautioned against relying on macroinvertebrate data
alone. Macroinvertebrate data should be used in
conjunction with other types of data to provide
sound information on the health of aquatic-riparian
ecosystems.

674

Macroinvertebrates

LITERATURE CITED u.s. department of agriculture, forest ser

VICE. 1981. Aquatic habitat surveys handbook FSH

2609. 2 3 for the General Aquatic Wildlife System. U.S.

Dep. Agric, For. Serv. Region 4. Ogden, UT. 257pp.

(Draft).
GREEN, R.H. 1979. Sampling design and statistical meth- WEBER, C. 1973- Biological field and laboratory methods

ods for environmental biologists. John Wiley & Sons. for measuring the quality of surface waters and ef-

251pp. fluents. EPA. 670/4-73-001.

Macroinvertebrates 675

V SPECIAL STUDIES

33 Radiotelemetry

34 Food Habits

35 Weather and Climate

33

RADIO-
TELEMETRY

Paul L. Hegdal1

U.S. Fish and Wildlife Service
Denver Wildlife Research Center
Building 16, Denver Federal Center
Denver, CO 80225

Bruce A. Colvin

Editor's Note: Within the past 25 years, wildlife
research has moved from subjective field observa-
tions to highly complex, objective measurements of
animal behavior. The objective measurements, in
many cases, are accomplished by use of radiotelem -
etry. This chapter provides an overview of current
state-of-the-art uses of radiotelemetry in locating
habitats used by subject animals.

While the urge to use new "gadgets" to learn
about animals is understandable, today's biologists
should conform to conventional responsibilities
toward experimental design and project planning.
"Does the end justify the means?" This chapter,
therefore, also identifies current processes and
equipment needed to plan and execute a telemetry
program useful in habitat management, but recog-
nizes the field is continuously expanding.

Department of Biological Sciences
Bowling Green State University
Bowling Green, OH 43403

INTRODUCTION

The development and implementation of radio-
telemetry in wildlife research has tremendously
broadened the opportunity to examine components
of species' natural history and ecology. Since its in-
troduction into wildlife research approximately 25
years ago (Marshall and Kupa 1963; Mech 1983),
radiotelemetry use has vastly increased, especially in
recent years with the increasing attractiveness of
"high-tech" approaches to research, their usefulness
and availability.

Today, examples of radiotelemetry research are
commonly found in literature. Radiotelemetry hard-
ware, techniques, and uses are continuously being
improved and evaluated. The number and diversity
of species being studied with radiotelemetry also are
continuing to grow and include mammals (Marshall
et ai. 1962; Mech 1977; Madison 1978; Barrett
1984), birds (Nicholls and Warner 1972; Cochran
1975), reptiles (Carr 1965; Schubauer 1981; Osgood
1970; Kenward et al. 1982), amphibians (Jansen
1982), fish (Winter et al. 1978), and even crabs
(Wolcott 1980) and crayfish (Covich 1977). Habitats
where these studies have been undertaken range
from the polar circle (Kolz et al. 1980) to temperate
regions (Verts 1963; Imboden 1975), tropical re-
gions (Bruggers et al. 1983), and oceans (Garshelis
and Garshelis 1984).

There are several possible reasons why radiote-
lemetry may be implemented as part of a wildlife
research or management scheme. Accessibility of the
species often is a principal reason. Physical charac-
teristics of a species' habitat, such as rough terrain or
dense vegetation, may limit the opportunity to seek

1 Current affiliation: Denver Wildlife Research Center, APHIS/
USDA Building 16, Denver Federal Center, Denver, CO 80225.

Radiotelemetry

679

and observe a species. Also, a species may be noctur-
nal, highly secretive, difficult to trap repeatedly,
capable of wide-ranging movements, or even subter-
ranean or aquatic in its habits.

A second advantage of radiotelemetry is that
data can take a continuous form rather than the dis-
crete form which occurs for example, through
trapping and marking. Once an operating radio trans-
mitter is attached to an animal, that particular animal
can potentially be located continuously either day
or night. Radiotelemetry also provides the opportu-
nity to remotely follow or census wildlife. Once
instrumented, the specific animal can be identified
and observed in a non-disruptive manner and, thus, a
more accurate depiction of the species' movements,
habitat use, and ecology may be acquired.

In contrast, when traps are used to locate an
animal, one must assume that a trap is available to
capture the animal at a particular location and time
and that the animal will enter the trap. Traps (or
observers) are typically placed at preconceived (sub-
jective) locations where animals might range, and
thus movement data generated from trapping (or di-
rect observation) may be strongly biased. Addition-
ally, trapping often takes on a day or night censusing
form, depending on the activity periods characteris-
tic of the species, thus adding to the discrete and
limited nature of observations that can be made.
Continuous observations also can be made by direct
observation; however, without radiotelemetry, locat-
ing target animals, distinguishing individuals, and
monitoring the activities of many isolated individuals
may be far more difficult.

The purpose of this chapter is to discuss radio-
telemetry techniques; types of available equipment;
and applications for determining movement, migra-
tion, and habitat use by wildlife. In addition, we de-
scribe potential problems, limitations, and costs.

RADIO-TRACKING EQUIPMENT

Radio tracking should be considered as no more
than a technique for extending the range of one's
observational powers. Even the simplest equipment
requires a significant financial investment. A simple
system for tracking about 10 animals with hand-
carried receivers and antennas will cost $3,000 to
S4,000 (1986 values).

Many improvements have been made since the
initial successful studies in the early 1960s, such
as more efficient transmitters; better encapsulating
materials; lighter, more energy-efficient batteries;
solar units; and more efficient, easy-to-use receivers.
In addition, all components and complete systems
now are commercially available from several sources
(Table 1).

Table 1. Possible sources of supply for radio-
tracking equipment.

Advanced Telemetry Systems, Inc.
23859 Northeast Highway 65
Bethel, MN 55005
(612)434-5040

AVM Instrument Company
6575 Trinity Court
Dublin, CA 94566
(415)829-5030

Cedar Creek Bioelectronics Laboratory
University of Minnesota
Bethel, MN 55005
(612) 434-7361

CompuCap

8437 Yates Avenue North

Brooklyn Park, MN 55443

(612)424-2373

Custom Electronics
2009 Silver Court West
Urbana, IL 61801
(217)344-3460

Custom Telemetry and Consulting
185 Longview Drive
Athens, GA 30605
(404)548-1024

L. L. Electronics
Box 247

Mahomet, IL 53405
(217) 586-2132

Ocean Applied Research Corp.
10447 Roselle Street
San Diego, CA 92121
(619)453-4013

Smith-Root, Inc.

14014 Northeast Salmon Creek Avenue

Vancouver, WA 98665

(206) 573-0202

Stuart Enterprises

Box 310, 124 Cornish Court

Grass Valley, CA 95945

(916)273-9188

Telemetry Systems, Inc.
Box 187

Mequon, Wl 53092
(414) 241-8335

Telonics

932 East Impala Avenue
Mesa, AZ 85204-6699
(602) 892-4444

680

Radiotelemetry

Table 1. Possible sources of supply for radio-
tracking equipment (concluded).

Wildlife Materials, Inc.
R.R. 1, Grant City Road
Carbondale, IL 62901
(618) 549-6330

'Use of companies on this list does not imply Federal
Government endorsement.

There are basically two components in a radio-
tracking system: a transmitting system and a receiv-
ing system. The transmitting system consists of the
transmitter which is attached to an animal by an
appropriate method such as a radio collar on big
game animals, patagial transmitters on birds, and im-
planted transmitters in fish or snakes. The receiving
system consists of a receiver, a receiving antenna,
and either an operator or recorder.

Patagial transmitter attached to a hen pheasant.

Transmitters

The specific type of transmitter chosen will
depend on size, morphology, and behavior of the
animal under study; possible attachment methods;
transmitter availability and cost; necessary transmis-
sion range; habitat where used; and the particular
data to be collected.

Almost any animal that weighs over 15 g (0.5
oz) can be equipped with a radio transmitter and
monitored for at least a short time. Obviously, the
larger the animal, the larger the transmitter package
can be. Transmitters are available commercially rang-
ing from slightly over a gram (0.03 oz) to several
kilograms. The actual transmitter per se does not

vary much in weight but the power source, packag-
ing material, and attachment material can add sub-
stantially to the weight. Weight considerations for
the animal usually become important only for the
smaller (<lkg[<2.2 1b]) species. This can be the
most critical consideration with very small animals
(< 20 g [< 0.7 oz]) and especially birds. Generally,
the transmitter should be no more than 5% of body
weight. The Banding Office of the U.S. Fish and Wild-
life Service recommends no more than 3% of body
weight for transmitters used on birds. Cochran
(1980) stated that many species seem to tolerate a
package that is 4% of their body weight and appear
to behave normally not too long after such a package
has been attached. He also added that there is noth-
ing "magic" about 4% . However, there may be some
species that cannot be radio-equipped with these
transmitter-weights, or even lighter transmitters,
without having significant behavioral or physical ef-
fects as a result of instrumentation.

Ordinarily, each transmitter used in a particular
study area is on a unique frequency (or channel);
therefore, to tune in each animal, the operator
merely turns a dial or activates switches on a radio
receiver to the appropriate frequency (or channel).

Radio frequency management in the U.S. is
based on the Communications Act of 1934. This
legislation divided frequency spectrum users into
two groups (federal and non-federal). The National
Telecommunications and Information Administration
(NTIA) controls the federal portions of the spec-
trum, while the Federal Communication Commission
(FCC) controls the non-federal portions of the spec-
trum for use by state and local governments as well
as the private sector. Since segments of the spectrum
are shared by federal and non-federal users, opera-
tions are mutually coordinated by the Interdepart-
mental Radio Advisory Committee (IRAC). Each
Federal Government agency and the FCC are repre-
sented on the IRAC to address spectrum manage-
ment issues. The IRAC, through NTIA, has authority
to approve, disapprove, and cancel any federal radio-
frequency assignment. In effect, IRAC is the licensing
authority for federal users while the FCC licenses
non-federal users.

Two portions of the spectrum have been set
aside for wildlife telemetry use (40.16 to 40.20 MHz
and 216 to 220 MHz); however, these frequencies
have not been widely used by wildlife researchers or
managers. Radio-frequency spectrum managers are
becoming increasingly concerned about unlicensed
operations and the violations of rules and regulations
pertaining to radio communications. By law, teleme-
try users must operate within the rules and regula-
tions because any interference occurring between
authorized and unauthorized uses will cause unau-
thorized use to shut down, resulting in costly losses
of time and effort expended on studies. One must

Radiotelemetry

681

have authorization to use a particular frequency be-
fore any transmitters can be placed on free-ranging
animals. For further guidance and information, see
Kolz (1983).

Most transmitters use mercury or lithium batter-
ies to power the transmitter; however, solar-powered
units, some with rechargeable batteries, also are
available (Cochran 1980). Battery choice must be
tempered by the amount of additional weight the an-
imal can carry and the length of time the animal
needs to be tracked. If the transmitter is retrievable
or the animal recapturable, transmitters can be
reused and their longevity extended by changing
batteries.

Most simply, activating a transmitter includes
soldering a final connection and then potting (cover-
ing) that connection with an appropriate quick-set-
ting potting material (i.e., dental acrylic, epoxy
patch, or silicone rubber); others are activated by
cutting a wire that allows completion of a circuit.
Additionally, with some transmitters, placing or re-
moving a magnet near an imbedded switch turns the
transmitter on or off. A magnet system is preferred
as it does not require any soldering or potting by the
user and is more likely to be completely sealed from
moisture. Normally transmitters are activated only
shortly before attachment to an animal. However,
some users routinely turn transmitters on for a few
days prior to attachment to assure proper operations.
The loss of a few days of battery life may be critical
only for small, short-lived transmitters.

Most transmitter antennas are some sort of
tuned whip. The most efficient antenna lengths are
in quarters of wave lengths of a particular frequency
(such as lA, V-z, or whole wave). Except for the high
frequencies, most wildlife transmitter antennas can-
not be that long (a full wave length at 164 MHz
= 1.82 m) and, therefore, are tuned with a coil to
whatever length can be tolerated on the animal.
Whip antennas are more efficient than loop antennas.
However, in some applications, the loop antenna
serves as an integral part of the attachment collar.

There are a variety of methods for attaching
transmitters to animals, and before any transmitters
are attached to any species, one should review avail-
able literature on the species, try the attachments,
and act on the results of those endeavors. Neck col-
lars are the most common radio-attachment method
for terrestrial mammals (Cochran 1980). Harness,
ear tag, and implanted transmitters have been used
on some mammals that are difficult to collar (Mech
1983). For birds, harness, tail-clip, poncho, glued-on,
leg-band, and patagial transmitters have been suc-
cessfully employed (Bray and Corner 1972; Fitzner
and Fitzner 1977; Amstrup 1980; Bruggers et al.
1981 ). Implantable or ingestible transmitters are use-
ful where external attachment is impractical, such

as with fish or snakes (Osgood 1970). Tethers have
proven useful for sea turtles and manatees (Timko
and Kolz 1982).

A number of factors need to be considered
when designing or specifying the type of collar, har-
ness, or other attachment to be used. Shape, width,
contouring, durability, flexibility, smoothness next to
the animal, size adjustability, compactness, cryptic
design, internal or external antenna, and ease of at-
tachment all can be important (Cochran 1980).
Sharp edges and points should be avoided or at least
placed where there will be minimal irritation from
contact or pressure. Without at least considering
these factors, routine behaviors such as running,
flying, feeding, mating, or even resting may be ad-
versely affected. If necessary, one may have to test
various attachments on some species where methods

Radio-collared opossum.

-%

"•»*

%

Radio-collared coyote.

682

Radiotelemetry

and designs have not been evaluated by others. Ide-
ally, the transmitter will be on the animal only as
long as needed to collect the desired data. Some
transmitters are designed to fall off after some prede-
termined time; however, there have been some
problems with premature transmitter loss. With a
well-designed attachment, many animals have not
had significant problems carrying the transmitter for
life. We have had transmitters on owls for over 5
years with no apparent problems.

Transmitters for remote locations generally cost
between $100 and $250 each. A variety of special-
purpose transmitters have been developed for de-
tecting movement or mortality (Kolz 1975); measur-
ing heart rate, respiration rate, temperature, and
blood pressure, as well as conducting electrocardi-
ograms and encephalograms (Amlaner 1978); firing
darts on command for recapture (Mech et al. 1984);
and some specific behaviors such as frequency of
urination. However, these transmitters and receivers
are more expensive than transmitters for remote
locations.

Receivers

A variety of receivers are available, ranging from
$700 for the simpler receivers to about $2,000 for
some of the newer, wider band models with pro-
grammable automatic switching functions, often re-
ferred to as "scanners." However, paying more
money for a receiver will not guarantee a more sen-
sitive receiver. Most can be tuned to equal "state
of the art" sensitivity ( ± 150 db) — even the $700
versions.

When planning purchases of telemetry equip-
ment, at least one extra receiver should be acquired
for the project as a backup. There is nothing more
frustrating than having radio-equipped animals and
no working receivers to track them. Extra batteries
should always be available for receivers and any
other equipment used in conjunction with them.

Most receivers have a —

( 1 ) power switch, often with a position for inter-
nal or external power;

(2) channel selector (for digital frequency display
models, this may be dials or switches for
changing the displayed frequency);

(3) fine-tuning frequency dial (on non-digital
models);

(4) gain (volume) dial;

(5) sweep switch that allows automatic sweep
within a channel or between preset exact
frequencies; and

(6) switches for setting frequencies to be
searched.

Most receivers also have a variety of input and
output plugs for antennas, headphones, external
power, meters, pulse-interval counters, recorders,
and other attachments.

Earphones are not necessary in many tracking
situations. The inexpensive models ( $20 or less) are
adequate in most cases for hand or vehicle tracking.
However, for tracking in aircraft in windy or noisy
conditions, earphones are necessary, and it is worth
the extra money to get the best — $100 to $200
each. (Long-continued use of earphones at high au-
dio levels has resulted in diminished hearing abilities
among some research biologists.)

Antennas

Almost any piece of metal connected to the
antenna plug of a receiver will enhance the signal,
but properly tuned and tested antennas should be
used for maximum efficiency. The yagi antenna is the
most commonly used antenna for radiotelemetry
studies. A good, single hand-held yagi antenna costs
$50 to $100. As frequency increases, wave length
decreases and the size of antennas also decreases.
Yagi antennas are too large to be very practical for
hand-held or vehicle-mounting at the lower (30 to
50 MHz) frequencies but can readily be used on
vehicles or hand-held at the higher (over 150 MHz)
frequencies.

Antennas can be mounted on towers, vehicles,
or aircraft. Two (or more) yagi antennas can be
combined in such a manner (called stacking) to in-
crease their range and directional ability. In dual
array systems, antennas can provide a bearing accu-
racy of 1 to 2°. Often, permanent tracking stations
consisting of a few strategically placed antenna tow-
ers or masts mounted on the highest hills in the
study area will yield adequate signals. These towers
can be similar to the large, automatically rotating
system described by Nicholls and Warner (1972) or
relatively simple structures with a coaxial cable left
dangling near the base (Merson et al. 1982).

Permanent or temporary tower antennas can
work well for sedentary animals; however, the accu-
racy of radio locations deteriorates near the base line
of the antennas (because the angle of intercept nears
zero). If additional towers or mobile tracking sta-
tions can be employed, this problem can be
overcome.

Radiotelemetry

683

Radio-tracking vehicles (mobile tracking sta-
tions) can be equipped with roof-mounted, dual yagi
antenna systems. Bray et al. (1975) described a re-
movable cartop antenna system. However, for most
studies using mobile tracking stations, we strongly
recommend vehicles equipped with a through-the-
roof antenna system similar to that described by
Hegdal and Gatz (1978). These antennas are rotated
from inside the vehicles, and radio bearings are indi-
cated on a 360°, 25-cm ( 10-in.) protractor by a
pointer attached to the antenna mast. The pointer is
aligned to the null of the dual yagi antennas. Coaxial
cables from the dual yagis are attached to a null-

peak switchbox (available from commercial sup-
pliers), which allows switching from in-phase (for
maximum signal strength) to out-of-phase (for pre-
cise bearings) operation. Additionally, radio commu-
nication, a plotting table, auxiliary batteries, and
extra lighting installed in the vehicles will expedite
tracking studies. Figure 1 shows an inexpensive
method for equipping vehicles with a roof-mounted,
dual-beam antenna system. If properly balanced,
antennas can be rotated easily and animals can be
tracked while the vehicle is moving as fast as 55
mph.

o o

o o

Jk

An

tJi"muiiT .

I

f

9

It

■tt ?

j ,,, [4

-t-

tr, i. ■-■■V- i v.-

I, I. inn

I

Figure 1. A roof-mounted, dual yagi, radio-tracking system.
684 Radiotelemetry

On aircraft, antennas are usually mounted on
wing struts (with certified antenna mounts) in a
"side-looking" fashion, perpendicular to the fuselage
(Gilmer et al. 1981). Care also must be taken to
ensure that coaxial cables are properly secured with
tape, not crimped, and that they and the rest of the
telemetry equipment do not interfere with the pilot's
operation of the aircraft.

Too often, new (and experienced) radio-track-
ers do not periodically test and check their antenna
system. They may have an extremely sensitive re-
ceiver matched to a poorly tuned antenna and there-
fore have (unknowingly) a very inefficient receiving
system.

Coaxial cables are used to connect antennas or
other equipment, such as recorders, to receivers.
This is a shielded cable (a center wire conductor
surrounded by the ground wire ). Care must be taken
to avoid sharp bends, twisting, and flattening or
pinching; distortion of the cable shape can drastically
affect or eliminate signal reception.

'^Rc~^^ ^

A radio-tracking vehicle at a marked tracking station.

Other Equipment

Pulse interval counters are available for about
$400 and can be used in conjunction with several
receivers. These devices measure the time between
consecutive, pulsed radio signals and usually display
that value in thousandths of a second. They are espe-
cially helpful for identifying pulse rates (animals)
when there is more than one transmitter operating
on the same frequency (or channel). Recorders can
be used in conjunction with automatic switching
(scanning) receivers to record presence or absence
of radio-equipped animals over time at sites such
as dens or nests (Gilmer et al. 1971; Harrington and
Mech 1982; Mech 1983). However, as with auto-
matic tracking systems, one must have a good strong
signal for most of the recorders and counters to
work properly. The human ear can detect signals
from the receiver better than any of the available
auxiliary equipment.

Satellite radiotelemetry also has been used suc-
cessfully on a few wildlife species (Kuechle et al.
1979; Timko and Kolz 1982). However, transmitters
and receiving equipment are more specialized and
considerably more expensive than those we have
described, and it is beyond the scope of this chapter
to fully discuss these aspects of radiotelemetry.

Advice

We strongly recommend that the telemetry user
depend on electronic specialists for constructing
and servicing transmitters, receivers, and other spe-
cialized equipment. However, the biologist must
have a general knowledge of radio signal propaga-
tion; factors that affect it; and some skills and equip-
ment for servicing, checking, and repairing parts of
the telemetry system. Users should be able to charge
power sources for receivers and other equipment;
change batteries in some transmitters (some must be
returned to the manufacturer); replace coaxial cables
and antenna elements; turn transmitters on and off;
check power sources, telemetry, and antenna sys-
tems for shorts and open circuits; and recognize that
such problems exist. Minimal support equipment
should include a volt-ohmmeter; battery tester; sup-
ply of miscellaneous wire connectors and terminals;
soldering gun; solder; wire strippers; and various-
sized screwdrivers, pliers, and adjustable wrenches.

DATA SAMPLING

The planning stages of a radiotelemetry study
must include decisions on specific use or need of the
technology, equipment, personnel availability, and
cost. Just as important, however, are decisions on
how data will be collected, analyzed, and inter-
preted. The recording of a single data point in the
field can be a simple procedure; however, deciding

Radiotelemetry

685

the value or use of hundreds or thousands of coordi-
nates may not be and should not be left until the
field work is complete.

Locating an animal either on a monthly, weekly,
daily, hourly, or minute-by-minute basis all may be
possible depending on the available hardware, per-
sonnel, and funding. However, while a large number
of radio-equipped animals can be located continu-
ously, rarely is that done. The specific form and fre-
quency of sampling should be particularly relevant
to the specific questions that are being addressed
with radiotelemetry use. It may be appropriate to
simply monitor general movements or continued
survivorship; thus, the sampling scheme may be ex-
tensive or random and the radio-tracking results may
be presented in a rather descriptive form. In con-
trast, when evaluating habitat selection or rates of
movement, for example, the sampling scheme may
have a highly intensive or systematic style with data
recorded in more of a quantitative manner (e.g.,
effort- or time-specific). In the latter case, animals
may be located at specific and independent time
intervals or continuously tracked within discrete and
independent time periods.

Sampling scheme decisions also may be affected
by the species being studied. Animals that are capa-
ble of wide-range movement or frequent habitat
change may have to be located more often to avoid
loss of contact than those that are largely sedentary.
The length and timing of daily activity cycles of ani-
mals, for example, nocturnal versus diurnal, also
will affect the length and timing of radio-tracking
efforts; the sampling scheme for any given individual
may have to be compromised considerably depend-
ing on how many other animals are simultaneously
radio-equipped. Additionally, more frequent sampling
may be needed if the hardware being used includes
short transmission-reception range or a limited bat-
tery or radio attachment life.

A sampling scheme to evaluate the importance
of specific habitat types might be based on the total
time an animal spends in each habitat, the number of
locations that are randomly or systematically re-
corded in each habitat type or, in contrast, the num-
ber of times an animal enters each habitat from
alternative habitats. The time that an animal spends
in a particular habitat may not be equated necessar-
ily to the value of that habitat to the species; fre-
quent, but short visits to other habitat types may
indicate a greater resource value. In addition, when
sampling for home ranges, Smith et al. ( 1981 ) stated
that it is important to define a large portion of the
area over a relatively short observation period. This
is because home ranges are really time-specific,
meaning they potentially can change with season,
habitat, or the animal's reproductive state. Thus,
intensive sampling efforts in a specific time period

may be necessary when accurately defining a home
range.

One additional concern related to sampling is
movements recorded immediately following animal
instrumentation and release. Such movements should
be considered biased (rather than exemplary) be-
cause of the disruption created by trapping and han-
dling. Therefore, although tracking an animal
immediately following release (e.g., 1 day) may be
appropriate, there should be a degree of indepen-
dence used in the initial observation of habitat or
home range analyses.

FIELD-TRACKING

Field-tracking first dictates that one must be-
come an expert in capturing live, uninjured animals
of the desired species, which can become a major
effort in itself for some species. Secondly, we
strongly recommend that those new to radio-track-
ing and those not using it for 2 to 3 years visit and
obtain training from a state-of-the-art ongoing proj-
ect. Novice radio-trackers, especially, will encounter
many pitfalls and these may be avoided by such help.
Also, there are new developments annually that may
be useful in any radiotelemetry project.

Triangulation is the basic principle in most ra-
dio-tracking. It can be accomplished by taking bear-
ings from two locations; the animal is assumed to be
near the point where the bearings cross (the loca-
tion or "fix"). Ideally, bearings should cross at about
a 90° angle and be taken simultaneously by two ra-
dio communication observers and as close to the
animal as possible. As the angle of interception be-
comes more acute and the distance from the animal
to the observers increases, the locations become less
accurate and the error polygon increases. For a more
thorough discussion of error polygon and means of
estimating radio-location errors, see Mech (1983),
Lee et al. (1985), and Saltz and Alkon (1985). Con-
siderable experience in the field is necessary to be-
come skilled at rapidly selecting tracking locations
for accurate bearings.

When learning to radio-track, one should first
practice locating an activated transmitter hidden by
someone else. Repetition of this "game" can most
readily acquaint anyone with radio-tracking tech-
niques and the sensitivity and capability of the
equipment.

Hand Tracking

In its simplest form, field radio-tracking is done
by carrying the receiver and hand-held antenna and
"homing in" (walking out) on the radio-equipped
animal. Proper "homing" procedures involve deter-
mining the approximate direction of the transmitter,

686

Radiotelemetry

tuning in the signal, reducing gain to minimal level,
again reducing gain to minimal level, and moving
in the direction of the signal. While moving toward
the signal, the antenna should be waved in an ap-
proximate 120° arc, back and forth, continuing to re-
duce gain to a minimal level. When gain cannot be
reduced further, it indicates the transmitter is
nearby. If the animal has not been observed or the
transmitter found, it can be located by removing the
antenna, turning the gain up so it barely can be
heard and, depending on signal strength, determining
the location by moving the receiver toward the
strongest signal. A small "locator" loop antenna also
is useful for radio-tracking at close range. These loca-
tor loops are especially useful when digging trans-
mitters from underground burrows.

"Homing" procedures have some obvious draw-
backs. For example, the animal may not be observed
by the biologist before it has been "pushed" or
flushed from roosting, foraging, or loafing areas, and
the location recorded may not reflect the habitat
utilized prior to the investigator's disturbance. Care
must be used to avoid these false locations and artifi-
cial or stimulated movements caused by the investi-
gator. Some animals (such as big game, raptors, or
animals that are nocturnal) are difficult to observe or
approach without disturbing them. This tracking
method may be relatively useless for determining
movements of nocturnal animals during activity pe-
riods, but may (with care) readily be employed to
determine specific daytime roosting, denning, or
bedding areas during periods of inactivity. When ra-
dio-tracking large numbers of animals, hand-held
equipment becomes inefficient as it may take several
hours to locate some animals by "walking them out."
This is especially true in inaccessible areas or in
areas of dense vegetation.

With some small, relatively immobile and fosso-
rial species (such as rodents, small birds, and rep-
tiles), virtually all radio-tracking may be done with
hand-held equipment. Hand-held tracking will always
be necessary to locate and recover dead animals,
animals that must be recaptured, or transmitters that
have fallen off the carrier animal.

Recovery of lost transmitters or dead animals
can be difficult, sometimes requiring considerable
digging, use of heavy excavating equipment, or ex-
plosives. For example, black-tailed prairie dogs were
recovered in South Dakota at depths of 2.5 to 3 m
(8.25 to 10 ft — deepest 4.3 m [14 ft]). Also, one
cannot assume that radio-equipped animals always
will be in normal habitat or locations expected for
that species. Finding an animal is often based on
what a species is "supposed to do" rather than what
it may actually do. Birds, rodents, or most any spe-
cies can be taken by predators as well as legally or
illegally by humans. Animals and their transmitters
have been recovered in raptor nests, burrows, under

buildings, in car trunks, freezers in homes, and many
other unusual locations where the animal being
tracked would not be expected to venture.

Biologist tracking using a hand-held vagi.

Vehicle Tracking

Mobile tracking stations (vehicles) allow one to
reduce the antenna-to-animal distance and take ad-
vantage of topography in the area for better tracking
points. Vehicle tracking can only be done if there are
reasonable access routes within the study area. Mo-
bile tracking stations (two or more), equipped with
the described antenna systems, radio communication
equipment, auxiliary batteries, extra lighting, and
plotting tables, allow investigators to quickly search
areas and follow fast-moving animals such as migrat-
ing birds in daylight or darkness. During activity, it
may be especially important to obtain simultaneous
bearings. Since many animals can move rapidly, false
interception points may be recorded if there is a
time lag between bearings. Most dual-beam antenna
systems mounted on a vehicle require a minimum
clearance of about 3 m ( 10 ft). If the study area
contains wooded areas with low-hanging branches, it
may be necessary to trim several kilometers of road
for adequate clearance of the tracking vehicle. In
areas where this is not practical, a single yagi should
be mounted in a horizontal plane. This will lower
clearance requirements and still give fairly reasona-
ble bearings, provided the distance from the tracking
vehicle to the animal is short when bearings are
taken.

Radiotelemetry

687

Turning directional yagi antenna from inside the tracking
van

Recording signals on plotter system.

Aerial Tracking

Usually, aerial tracking starts with searching the
area for the animal's last known location, increasing
altitude to 3,000 m (9,900 ft), and searching in an
enlarging circle or flying swaths (20- to 50-km [12.4-
to 3 1 -mi. ] wide ) for complete coverage of the areas
(Gilmer et al. 1981). In heavily populated areas near
cities, there may be too much radio interference to
fly at higher altitudes (over 500 to 1,000 m [310
to 620 ft]). Once a signal is detected, aerial tracking
simply can become "homing," similar to "walking
out" a radio-equipped animal (reducing gain and
lowering altitude as the source of the signal).

The use of aircraft is especially important in
large or inaccessible areas and with wide-ranging or

migrating species. Additionally, aerial searches for
missing or lost animals are usually much more effi-
cient than ground searches. The additional height
provided by aircraft greatly increases reception
range. For example, our vehicle tracking system has
a range of about 3 to 4 km ( 1.8 to 2.5 mi.) with a
7 g (0.24 oz) transmitter on a bird, while with air-
craft, a 35-km (21.7-mi.) receiving range can be
achieved.

The first step in aerial tracking is consulting
with a certified aircraft mechanic to make sure that
the antenna mounting system is certified by the Fed-
eral Aviation Administration (FAA). The biologist
should be familiar with areas to be searched and re-
view the flight, search area, and procedures with
the pilot before the flight. Exact tracking procedures

688

Radiotelemetry

depend on terrain and habitat, as well as the mobil-
ity of the species being followed. Experience of the
observer and pilot also can influence the efficiency
of aerial tracking (Hoskinson 1976).

Yagi antenna mounting system on aircraft.

Data Recording

Radio locations need to be plotted initially on
maps of an appropriate scale (U.S. Geological Survey
1:24,000) or aerial photographs (for many species,
a l:7,920-scale seems best [1 km = 12.7 cm,
1 mi = 8 in.]). A clearly defined grid coordinate
system should be superimposed on the maps or pho-
tos, using acetate (5 mil thickness works well) over-
lays. Permanent ink felt pens should be used for
marking the grid and any other permanent features,
while temporary markers are best for plotting on the
acetate. Our experience has indicated that black,
red, green, blue, and purple are excellent colors to
use. After the locations are recorded on data sheets,
the temporary markings can be wiped off the acetate
with a damp cloth. Our experience has shown that
the Universal Transverse Mercator System (UTM; U.S.
Department of the Army 1958) is ideal for most
radiotelemetry studies. UTM coordinates are noted
on U.S. Geological Survey maps and can be trans-
ferred to aerial photographs to create the grid coor-
dinate system. Data then can be digitally recorded to
the nearest km (0.62 mi.), 0.1 km (0.06 mi.), or
even 0.001 km (0.006 mi.) as appropriate for the
species. Additionally, the coordinates can be entered
directly into a computer terminal for data analysis.

Figure 2 illustrates tracking forms used for re-
cording radiotelemetry data. Observations are se-
quentially noted and include date, time, location, and
any pertinent behavioral or habitat descriptions.
Time is recorded as 0000 to 2400 hrs to prevent any
possibility of errors. UTM coordinates include nu-
merical and directional values, for example, 2760N,

824E, as explained by the U.S. Department of the
Army (1958).

Problems

Radio interference can, in some locations, create
difficulties. Bearings should be taken from natural
high points in the area. Power lines, fences, and large
buildings should be avoided since they can "bounce"
radio signals and readily produce false and highly
confusing bearings.

The number of biological studies using radiote-
lemetry has been increasing each year. For example,
in 1985, over 20,000 transmitters were placed on
wildlife species in the U.S. and a considerable por-
tion of these were not authorized. In addition, some
of the unauthorized frequencies being used are in
the range of police or other users who are emitting
very strong signals (compared to most wildlife trans-
mitters) that can drastically interfere with anyone
trying to track wildlife on these frequencies. These
facts stress the importance of having the proper
clearances for use of particular radio frequencies and
the need to coordinate activities with telemetry
users in the same area. This point becomes critical
when tracking migrating birds. While tracking owls
and hawks, problems have been encountered with
other researchers using unauthorized transmitters on
black bear. Obviously, one could be dangerously
surprised "walking out" a bear while thinking it was
an eastern screech owl.

Radiotelemetry

689

■o

c

12
u

■o
si

690

Radiotelemetry

No matter how well individuals or equipment
perform, contact with some transmitters could be
lost. To help locate these missing animals, aerial
tracking may be necessary. In addition, receiving
equipment should be left on at all times (tuned to
the appropriate missing frequencies ) while traveling
in vehicles; missing animals are often located when
least expected (frequently near their known home
range ). A bent transmitter antenna; the orientation of
an animal's body and, thus, antenna orientation; or
location of an animal in a low or structurally se-
cluded area (such as underground) may drastically
reduce reception range and thus cause the animal to
essentially "disappear." Some transmitters will come
off animals (or be taken off by them) and, depending
where they end up, can be difficult to locate. Unfor-
tunately, some lost transmitters are never located
and no biological conclusions can be made in most
of these situations.

DATA ANALYZING

Radio-tracking data may be reported in several
different ways. Initial description of results usually
includes the length of time that the animal carried
an operating transmitter, fate of the radio-equipped
individual (i.e., mortality, lost contact, dead battery),
the number of data points recorded, and possibly
the number of tracking periods or days. These data
are helpful in relating the kind and number of obser-
vations made to the amount of tracking time and
effort spent.

Enumeration data, reported in either descriptive
or tabular form, include the frequency at which var-
ious events occurred or when observations were
made — for example, the number of times that var-
ious types of mortality were documented or the
frequency that animals were located at each type of
roost or habitat type. Statistical comparisons then
can be made between the frequency of habitat use
and the availability of those same habitats within the
range of the animal, resulting in determining habitat
preferences (Johnson 1980).

Measurement data might include rates of move-
ment (e.g., km/h), rates of habitat interception (e.g.,
number of locations per habitat type per hour), time
spent in each habitat type (e.g., number of hours or
percentage of total tracking time ), distances between
nest sites and foraging areas, and home range size.
Home range size may be calculated several different
ways. The recommended method is reviewing home
range analysis techniques before deciding on any
one method (Jennrich and Turner 1969; Dunn and
Gipson 1977; Dixon and Chapman 1980; MacDonald
et al. 1980; Anderson 1982; Hackett and Trevor-
Deutsch 1982). However, most often radiotelemetry
home ranges have been described as some form of
a minimum area or convex polygon created by en-

compassing the recorded outermost points (see Bar-
rett 1984; Garshelis and Garshelis 1984). Also,
computer programs can be used to calculate home
range size (Ford and Krumme 1979; Anderson
1982).

Presentation of radiotelemetry data in graphic
form is common practice and may offer a reader
a perceptive view of movements. To prepare a figure
showing radiotelemetry results, tracing paper can
be placed over a grid pattern of UTM coordinates
marked to scale, and UTM coordinates recorded in
the field can then be plotted on tracing paper. Var-
ious landmarks or habitat types can be added appro-
priately to emphasize relationships between
locations, movement distances, and environmental
features. The completed art work then can be repro-
duced photographically. Computer-assisted graphics
also can be generated.

When diagramming radiotelemetry data, one
should consider showing only those data that illus-
trate a particular point, rather than trying to present
all or much of the data since that may not be neces-
sary and can result in a confusing illustration. Move-
ment data presented in graphic form may only
exemplify the movement observed during any one
tracking period.

Maps showing tracking data may show specific
routes or sites where events such as initial capture,
migration, nesting, or mortality occurred (Figure 3).
Other types of graphics can include "scatter" dia-
grams that show the range of locations and centers
of activity (Figure 4). Also, computer graphics can
be particularly effective in relating centers of activity
or utilization distributions (Tarter and Kronmal
1976; Anderson 1982), especially if the graphics are
in three-dimensional form (Figure 5). Range overlap
among conspecifics, or predator range versus that
of prey, can be diagrammed and home range sizes
presented clearly (Figure 6). Continuous movement
patterns can illustrate the rate and directionality of
movement (Figure 7), and the same radio-tracking
data imposed on habitat types can show distinctly
the relationships between movement patterns and
particular habitat types (Figure 8). Active video-
graphic techniques allow viewing distribution and
movements of many animals in a continual, spatial,
and temporal dispersion (White 1979).

Once locations are tabulated or graphically plot-
ted, a critical aspect of a radiotelemetry study must
follow, that is, interpreting the findings and answer-
ing why an animal was at a particular place at a par-
ticular time. Sanderson ( 1966) made an important
point when he stated that researchers must shift
their emphasis from concern over the movements of
animals to the reasons for the movements. Sanderson
(1966) continued by stating that movement patterns
are established and regulated by the density of the

Radiotelemetry

691

LEGEND
A Release site
▼ Aircraft location
• Satellite location
O Visual (Halo)

92°

~r

91°

90°

89°

Figure 3. Movement of a loggerhead turtle radio-tracked from October 16, 1979 to June 8, 1980 in the Gulf
of Mexico (from Timko and Kolz 1982).

N*

•\

\

I J\V. .

V * J>A

N

***

A

O

^

\

1 km

y

/

\

_./

y

/

6>\

D Experimental Nest
O Treated Farmstead

♦ Day Roost VTimes Observed

• Night Location

Figure 4. Range, night locations, and daytime roost sites recorded for a female common barn owl with
young, 3-10 weeks old, in southwest New Jersey. The owl was radio-tracked randomly from June 2 to
July 29, 1982. Dashed lines encompass outermost locations recorded (approximately 977 ha [2,250 a.];
from Colvin 1984).

692

Radiotelemetry

o

z

I-

0)
LU

cr

l-
z

UJ
O

<r
u
a.

NELSON BAY

Figure 5. Distribution of resting locations (N = 1,163) of radio -tagged sea otters in Nelson Bay, a male area
in Prince William Sound, Alaska, 1979-1981. Preferred rest area was near the center of the Bay (from
Garshelis and Garshelis 1984).

tN

/"

/

7

\

/

/>

s

/

/

/ ^

s /

-' /
/

i/

/

/

1 km

\/

/

□ Nest 90
— Female
(} Nest 108

Female

___ Male

Figure 6. Home range overlap of three adult, common barn owls, radio-tracked in southwest New Jersey,
June-July 1982 (from Colvin 1984).

Radiotelemetry

693

d Nest Site, Day Roost
• Night Location

1 km

Figure 7. Pattern showing rate of movement of an
adult, common male barn-owl between its nest
site (3 young, 7.5 weeks old) and foraging habi-
tats in southwest New Jersey. The owl was
radio-tracked continuously 2043-2253 h on
12 August and 2107-2330 h on 13 August 1982
(total tracking time = 4 h, 33 min) (from Col-
vin 1984).

species, food supply, reproductive activity, quality
and physiographic arrangement of the habitat and,
likely, many other factors. In other words, the total
life-history strategy of a species (Stearns 1976) and
both the biotic and abiotic components of its envi-
ronment may have to be considered when interpret-
ing the significance of a movement pattern or the
habitat use disclosed by radiotelemetry. Therefore,
background information on a species' natural history
and ecology become critical in accurately interpret-
ing movements. Additionally, detailed data on
weather; availability of selected habitats or food re-
sources; population density; and animal age, sex, and
reproductive state all may be essential in evaluating
movements and successfully completing the project.
Certainly, radiotelemetry locations can be plotted
and presented; however, the researcher who inter-
prets the data in context with species' ecology will
get the most out of the time, effort, and finances
invested in radiotelemetry studies.

DISCUSSION

Although there are many positive reasons for
incorporating radiotelemetry as part of a research ef-
fort, a researcher should not first decide to use radio-
telemetry and then seek reasons to justify its use. A
researcher should first rigorously consider the spe-
cific questions being addressed or investigated and
decide whether those questions can be answered
without radiotelemetry. Whether radiotelemetry is
truly needed too often becomes a subjective proce-
dure because of the attractiveness of radiotelemetry
technology. Therefore, we strongly recommend that
the objectives of the study, as well as the reasons for
using radiotelemetry, be well-defined and associated
costs and time be evaluated in relation to the type of
data needed. Many other techniques for marking
and studying the movements of animals are available
and may be more appropriate. These include tattoos,
brands, fur removal, aluminum or plastic tags or
bands, neck collars, fluorescent bone marker, radio-
active markers, microtaggents, and dyes (Day et al.
1980). Trap recapture or direct observation, in com-
bination with one of the above marking techniques,
may prove more appropriate than radiotelemetry
given specific research designs, budgets, and person-
nel constraints.

To the novice, radiotelemetry may appear to be
an efficient, simple, and exciting way to study wild-
life. However, when considering initiation of a radio-
telemetry project, even a researcher experienced in
radiotelemetry is reminded of radiotelemetry's limi-
tations and the time, cost, and technical problems
that can be involved.

Radiotelemetry has been used in many specific
ways in wildlife studies. Investigations of home range
and general movement most commonly have been
performed, and these studies actually are often the
base process in the many types of wildlife research
that may incorporate radiotelemetry. Home range
analysis with radiotelemetry repeatedly has resulted
in a much expanded view of species' home range
and potential for movement than previously detected
by trapping or direct observation (Taylor and Quy
1978). Basic studies of movement may focus on how
movement patterns and home range size change as
a function of sex and age, population density, sea-
sons of the year, time of day, or available habitat
types or habitat diversity (Mech et al. 1966; Trent
and Rongstad 1974; Hemker et al. 1984).

Beyond the initial documentation of movement
are important opportunities to investigate habitat use
and habitat requirements and to find migration
routes, wintering areas, nesting sites, and foraging
areas (Marshall et al. 1962; Curtis and Braun 1983;
Loft et al. 1984). Also, the positive or negative im-
pact of various land management or land-use prac-
tices on wildlife may be discernible from radio-

694

Radiotelemetry

Figure 8. Movement pattern and habitats intercepted by an adult male common barn-owl with 3 young, 7.5
weeks old, in southwest New Jersey. The owl was continuously radio-tracked 2043-2253 h on 12 August
and 2107-2330 h on 13 August 1982 (total tracking-time = 4 h, 33 min.) (from Colvin 1984).

equipped individuals that are subjected to a chang-
ing environment. When radio-tracking and habitat
analysis are conducted in combination, a perceptive
view of discrete habitat requirements and relation-
ships between habitat and population maintenance
may be achieved (Kohn and Mooty 1971; Jenkins
and Starkey 1984; Pierce and Peek 1984; Riley and
Dood 1984).

Examination of population dynamics, including
age-specific survival rates and mortality factors, has
been studied often with radiotelemetry (Stoddart
1970; Cook et al. 1971; Barrett 1984). A radio trans-
mitter allows continued survivorship to be docu-
mented or, conversely, mortality to be documented
essentially when it occurs. Thus, for example, a more
accurate representation of the occurrence of various
mortality factors in a population can be determined,
compared to when radiotelemetry is not used, and
emphasis is placed on those mortality factors that are

most easily identified (e.g., car collision). Predator-
prey relationships also can be studied and provide
additional insight into population dynamics (Mech
1967; Kolenosky 1972; Franzmann et al. 1980; Fuller
and Keith 1980).

Radiotelemetry also provides great advantages in
endangered species research because radio-equip-
ping a single individual can potentially locate con-
specifics in an efficient and non-disruptive manner
(Mech 1977; Fagerstone et al. 1985). In addition,
because endangered, rare, or threatened wildlife
often have highly specific habitat requirements, these
"micro" habitats may be more clearly and quickly
identified with radiotelemetry and thus protected.

Other uses of radiotelemetry have included
monitoring the status and movements of animals in-
volved in translocations or reintroductions (Fritts
et al. 1984). Also, population censusing has been

Radiotelemetry

695

conducted by determining the proportion of radio-
equipped animals not observed in a direct count and
adjusting the total population count upwards propor-
tionately (Floyd et al. 1979). Additionally, vertebrate
pest research often has included radiotelemetry in
evaluation of movements of pest species, efficacy of
wildlife control procedures on target species, and
toxic hazards to non-target species (Taylor 1978;
Fagerstone et al. 1981; Hegdal and Blaskiewicz 1984;
Heisterberg et al. 1984). In a more specialized form,
radiotelemetry has been used to investigate the phys-
iological adaptation of free-ranging animals to their
environment by remotely monitoring, for example,
heart rate (Kanwisher et al. 1978; Follmann et al.
1982).

From a review of the literature, it may appear
that uses of radiotelemetry are limited only by the
creativity of researchers and available hardware.
As the technology continues to improve, so will the
opportunities to attempt new uses and to further
address complex wildlife and ecological issues.

CONCLUSION

There are four major components of a radiotelem-
etry study:

( 1 ) Justified use of the technology based on a
specific research design;

(2) Consideration of hardware, personnel, cost,
and sampling strategies;

(3) Implementation of field research and data
collection; and

(4) Analysis and interpretation of data.

Each of the above components is equally important
in the planning and success of a radiotelemetry
study. Often, because of the specialization of these
components and the time involved, a team approach
is essential to adequately plan, conduct, and evaluate
radiotelemetry studies. The experience of the "team"
will strongly affect the efficiency, accuracy and, thus,
the outcome of the project. In final form, discussion
of radiotelemetry should be well-integrated with
information on species biology and ecology to best
understand and illustrate the role that movement,
migration, and habitat use play in the life-history
strategy of a species.

Radio-tagged mourning dove.

Burn owl instrumented with a transmitter and whip an-
tenna.

696

Radiotelemetry

LITERATURE CITED

AMLANER, C.J., Jr. 1978. Biotelemetry from free-ranging
animals. Pages 205-228 in Stonehouse, B., ed. Animal
Marking: Recognition Marking of Animals in Research.
Macmillan and Company, London.

AMSTRUP, S.C. 1980. A radio-collar for game birds. J.
Wildl. Manage. 44:214-217.

ANDERSON, D.J. 1982. The home range: A new nonpara-
metric estimation technique. Ecology 63:103-112.

BARRETT, M.W. 1984. Movements, habitat use, and preda-
tion on pronghorn fawns in Alberta. J. Wildl. Manage.
48:542-550.

BRAY, O.E. and G.W. CORNER. 1972. A tail clip for attach-
ing transmitters to birds. J. Wildl. Manage. 36:640-
642.

, RE. JOHNSON, and A.L. KOLZ. 1975. A removable

cartop antenna system for radio-tracking birds. Bird-
banding 46:15-18.

BRUGGERS, R., J. ELLIS, J. SEDGWICK, and J. BOURASSA.
1981. A radio transmitter for monitoring the move-
ments of small passerine birds. Proc. Int. Conf. Wildl.
Biotelemetry 3:69-79.

, MM. JAEGER, and J.B. BOURASSA. 1983. The ap-
plication of radiotelemetry for locating and control-
ling concentrations of red-billed quelea in Africa.
Tropical Pest Manage. 29:27-32.

CARR, A. 1965. The navigation of the green turtle. Sci.
Am. 212:79-86.

COCHRAN, W.W. 1975. Following a migrating peregrine
from Wisconsin to Mexico. Hawk Chalk 14:28-37.

. 1980. Wildlife telemetry. Pages 507-520 in Schem-

nitz, S.D., ed. Wildl. Manage. Techniques Manual.
Wildl. Soc, Inc., Washington, DC.

COLVIN, B.A. 1984. Barn owl foraging behavior and sec-
ondary poisoning hazard from rodenticide use on
farms. Ph.D. dissertation. Bowling Green State Univ.,
Bowling Green, OH. 326pp.

COOK, R.S., M.W. WHITE, DO. TRAINER, and W.C.
GLAZENER. 1971. Mortality of young white-tailed
deer fawns in south Texas. J. Wildl. Manage. 35:47-56.

COVICH, A. 1 977. Shapes of foraging areas used by radio-
monitored crayfish. Am. Zoology 17:205 (Abstract).

CURTIS, P.D. and C.E. BRAUN. 1983. Radiotelemetry loca-
tion of nesting band-tailed pigeons in Colorado. Wil-
son Bull. 95:464-466.

DAY, G.I., S.D. SCHEMNITZ, and R.D. TABER. 1980. Cap-
turing and marking wild animals. Pages 61-68 in
Schemnitz, S.D., ed. Wildl. Manage. Techniques Man-
ual. Wildl. Soc, Inc., Washington, DC.

DKON, KR. and J.A. CHAPMAN. 1980. Harmonic mean
measure of animal activity areas. Ecology 61:1040-
1044.

DUNN, J.E. and P.S. GIPSON. 1977. Analysis of radiotelem-
etry data in studies of home range. Biometrics 33:85-
101.

FAGERSTONE, KA., D.E. BIGGINS, and T.M. CAMPBELL,
III. 1985. Marking and tagging of black-footed ferrets
(Mustela nigripes). Pages 10.1-10.10 in Anderson,
S.H. and D.B. Inkley, eds. Proc. Black-footed Ferret
Workshop, Sept. 18-19, Univ. Wyoming, Laramie.

, G.H. MATSCHKE, and D. ELIAS. 1981. Radiotelem-
etry to evaluate effectiveness of a new fumigant car-
tridge for controlling ground squirrels. Proc. Int.
Conf. on Wildl. Biotelemetry. Univ. Wyoming, Lara-
mie. 3:20-25.

FITZNER, RE. and J.N. FITZNER. 1977. A hot melt glue

technique for attaching radio transmitter tail packages
to raptorial birds. North Am. Bird Bander 2:56-57.

FLOYD, T.J., L.D. MECH, and ME. NELSON. 1979. An
improved method of censusing deer in deciduous-
coniferous forests. J. Wildl. Manage. 43:258-261.

FOLLMANN, E.H., A.E. MANNING, and J.L. STUART. 1982.
A long-range implantable heart rate transmitter for
free-ranging animals. Biotelemetry Patient Monitoring
9:205-212.

FORD, R.G. and D.W. KRUMME. 1979. The analysis of
space use patterns. J. Theoretical Biol. 76:125-155.

FRANZMANN, A.W., C.C. SCHWARTZ, and R.O. PETER-
SON. 1980. Moose calf mortality in summer on the
Kenai Peninsula, Alaska. J. Wildl. Manage. 44:764-768.

FRITTS, S.H., W.J. PAUL, and L.D. MECH. 1984. Move-
ments of translocated wolves in Minnesota. J. Wildl.
Manage. 48:709-721.

FULLER, T.K and LB. KEITH. 1980. Wolf population
dynamics and prey relationships in northeastern Al-
berta. J. Wildl. Manage. 44:583-602.

GARSHELIS, D.L. andJ.A. GARSHELIS. 1984. Movements
and management of sea otters in Alaska. J. Wildl.
Manage. 48:665-678.

GILMER, D.S., L.M. COWARDIN, R.L. DUVAL, L.M. MECH-
LIN, C.W. SHAIFFER, and V.B. KUECHLE. 1981. Proce-
dures for the use of aircraft in wildlife biotelemetry
studies. U.S. Dep. Inter., Fish and Wildl. Serv. Resour.
Publ. 140. 19pp.

, V.B. KUECHLE, and I.J. BELL, Jr. 1971. A device for

monitoring radio-marked animals. J. Wildl. Manage.
35:829-832.

HACKETT, D.F. and B. TREVOR-DEUTSCH. 1982. Radiote-
lemetric assessment of grid-trapping techniques in a
study of the eastern chipmunk {Tatnias striatiis L. ).
Biotelemetry Patient Monitoring. 9:213-226.

HARRINGTON, F.H. and L.D. MECH. 1982. Patterns of
homesite attendance in two Minnesota wolf packs.
Pages 81-105 in Harrington, F.H. and P.C. Paquet, eds.
Wolves of the World. Noyes, Park Ridge, NJ.

HEGDAL, PL. and R.W. BLASKIEWICZ. 1984. Evaluation
of the potential hazard to barn owls of Talon {brodi-
facoum bait) used to control rats and house mice.
Environmental Toxicological Chemistry 3:167-179-

and T.A. GATZ. 1978. Technology of radio-tracking

for various birds and mammals. Pages 204-206 in
Symp. on Application of Remote Sensing Data to
Wildlife Management. PECORA IV. Sioux Falls, SD.

HEISTERBERG, J.F., C.E. KNITTLE, O.E. BRAY, D.F. MOTT,
andJ.F. BESSER. 1984. Movements of radio-instru-
mented blackbirds and European starlings among win-
ter roosts. J. Wildl. Manage. 48:203-211.

HEMKER, T.P., F.G. LINDZEY, and B.B. ACKERMAN. 1984.
Population characteristics and movement patterns of
cougars in southern Utah. J. Wildl. Manage. 48:1275-
1284.

HOSKINSON, R.L. 1976. The effect of different pilots on
aerial telemetry error. J. Wildl. Manage. 40:137-139.

IMBODEN, C. 1975. A brief radiotelemetry study on
moreporks. Notornis 2:221-230.

JANSEN, D.K 1982. A new potting material for radiotelem-
etry packages. Copeia 1982:189.

JENKINS, KJ. and E.E. STARKEY. 1984. Habitat use by

Roosevelt elk in unmanaged forests of the Hoh Valley,
Washington. J. Wildl. Manage. 48:642-646.

JENNRICH, R.L and F.B. TURNER. 1969. Measurement of
non-circular home range. J. Theoretical Biol. 22:227-
237.

Radiotelemetry

697

JOHNSON, D.H. 1980. The comparison of usage and avail-
ability measurements for evaluating resource prefer-
ence. Ecology 61:65-71.

KANWISHER, J.W., T.C. WILLIAMS, J.M. TEAL, and KO.

LAWSON, Jr. 1978. Radiotelemetry of heart rates from
free-ranging gulls. Auk 95:288-293-

KENWARD, RE., G.J. HIRONS, and F. ZIESEMER. 1982.
Devices for telemetering the behavior of free-living
birds. Pages 127-129 in Cheesman, C.L. and R.B.
Mitson, eds. Telemetric Studies of Vertebrates. Zoo-
logical Soc. London Symp. 49.

KEUCHLE, V.B., DP. DEMASTER, and D.B. SINIFF. 1979.
State of the art of and needs of the earth platform.
Proc. Int. Symp. Remote Sensing Environment 13505-
518.

KOHN, BE. and J.J. MOOTY. 1971. Summer habitat of

white-tailed deer in north-central Minnesota. J. Wildl.
Manage. 35:476-487.

KOLENOSKY, G.B. 1972. Wolf predation on wintering

deer in east-central Ontario. J. Wildl. Manage. 36:357-
369.

KOLZ, A.L. 1975. Mortality-sensing wildlife transmitters.
Int. ISA Biomedical Sci. Instrument Symp. 12:57-60.

, J.W. LENTFER, and H.G. FALLEK 1980. Satellite ra-
dio-tracking of polar bears instrumented in Alaska.
Pages 743-752 in Amlaner, C.J., Jr. and D.W. Mac-
Donald, eds. A Handbook on Biotelemetry and Radio-
Tracking. Pergamon Press, Oxford.

LEE, J.E , G.C. WHITE, R.A. GARROTT, R.M. BARTMANN,
and AW. ALLDREDGE. 1985. Accessing accuracy of a
radiotelemetry system for estimating animal locations.
J. Wildl. Manage. 49:658-663.

LOFT, E.R., J.W. MENKE, and T.S. BURTON. 1984. Seasonal
movements and summer habitats of female black-
tailed deer. J. Wildl. Manage. 48:1317-1325.

MACDONALD, D.W., F.G. BALL, and N.G. HOUGH. 1980.
The evaluation of home range size and configuration
using radio-tracking data. Pages 405-424 in Amlaner,
C.J., Jr. and D.W. MacDonald, eds. A Handbook on
Biotelemetry and Radio-Tracking. Pergamon Press,
Oxford.

MADISON, DM. 1978. Movement indicators of reproduc-
tive events among female meadow voles as revealed
by radiotelemetry. J. Mammal. 59:835-843.

MARSHALL, W.H., G.W. GULLION, AND R.G. SCHWAB.
1962. Early summer activities of porcupines as deter-
mined by radio-positioning techniques. J. Wildl. Man-
age. 26(l):75-79.

and J.J. KUPA. 1963. Development and use of radio-
telemetry techniques for ruffed grouse studies. Trans.
North Am. Wildl. Nat. Resour. Conf. 28:443-456.

MECH, L.D. 1967. Telemetry as a technique in the study
of predation. J. Wildl. Manage. 31:492-496.

. 1977. Productivity, mortality, and population trend

in wolves from northeastern Minnesota. L Mammal.
58:559-574.

. 1983. Handbook of animal radio-tracking. Univ.

Minnesota Press, Minneapolis. 107pp.

, R.C. CHAPMAN, WW. COCHRAN, L. SIMMONS,

and U.S. SEAL. 1984. A radio-triggered anesthetic-dart
collar for recapturing free-ranging mammals. Wildl.
Soc. Bull. 12:69-74.

, KL. HEEZEN, and D.B. SINIFF. 1966. Onset and

cessation of activity in cottontail rabbits and

snowshoe hares in relation to sunset and sunrise.
Animal Behavior 14:410-413-

MERSON, M.H., RE. BYERS, and L.D. LETA. 1982. A porta-
ble antenna base for fixed-station, radio-tracking.
Wildl. Soc. Bull. 10:44-45.

NICHOLLS, T.H. and D.W. WARNER. 1972. Barred owl
habitat use as determined by radiotelemetry. J. Wildl.
Manage. 36:213-224.

OSGOOD, D.W. 1970. Thermoregulation in water snakes
studied by telemetry. Copeia 1970:568-571.

PIERCE, J.D. and J.M. PEEK. 1984. Moose habitat use and
selection patterns in north-central Idaho. J. Wildl.
Manage. 48:1335-1343-

RILEY, S.J. and A.R. DOOD. 1984. Summer movements,
home range, habitat use, and behavior of mule deer
fawns. J. Wildl. Manage. 48:1302-1310.

SALTZ, D. and P.U. ALKON. 1985. A simple computer-
aided method for estimating radio-location error. J.
Wildl. Manage. 49:664-668.

SANDERSON, G.C. 1966. The study of mammal move-
ments— a review. J. Wildl. Manage. 30:215-235.

SCHUBAUER, J.P. 1981. A reliable radiotelemetry tracking
system suitable for studies of Chelonians. J. Herpetol-
ogy 15:117-120.

SMITH, G.J., JR. CARY, and O.J. RONGSTAD. 1981. Sam-
pling strategies for radio-tracking coyotes. Wildl. Soc.
Bull. 9:88-93.

STEARNS, S.C. 1976. Life-history tactics: A review of the
ideas. Quarterly Review Biol. 51:1-45.

STODDART, L.C. 1970. A telemetric method for detecting
jackrabbit mortality. J. Wildl. Manage. 34:501-507.

TARTER, ME. and R.A. KRONMAL. 1976. An introduction
to the implementation and theory of nonparametric
density estimation. Am. Statistician 30:105-112.

TAYLOR, K.D. 1978. Range of movement and activity of
common rats {Rattiis norvegicus) on agricultural
land. J. Applied Ecol. 1 5:663-677.

and R.J. QLIY. 1978. Long-distance movements of a

common rat {Rattus norvegicus) revealed by radio-
tracking. Mammalia 42:63-71.

TIMKO, RE. and A.L. KOLZ. 1982. Satellite sea turtle
tracking. Marine Fish. Review 44:19-24.

TRENT, T.T. and O.J. RONGSTAD. 1974. Home range and
survival of cottontail rabbits in southwestern Wiscon-
sin. J. Wildl. Manage. 38:459-472.

U.S. DEPARTMENT OF THE ARMY. 1958. Universal
Transverse Mercator Grid. Technical Manual TM 5-
241-8. U.S. Govt. Printing Office 1969-390-947/404.
66pp.

VERTS, B.J. 1963. Equipment and techniques for radio-
tracking striped skunks. J. Wildl. Manage. 27:325-339

WHITE, G.C. 1979. Computer-generated movies to display
biotelemetry data. Proc. Int. Conf. Wildl. Biotelemetry
2:210-214.

WINTER, J.D., V.B. KUECHLE, D.B. SINIFF, and JR.

TESTER. 1978. Equipment and methods for radio-
tracking freshwater fish. Misc. Rep. 152. Agric. Experi-
ment Station, Univ. Minnesota.

WOLCOTT, T.G. 1980. Optical and radio optical tech-
niques for tracking nocturnal animals. Pages 333-338
in Amlaner, C.J., Jr. and D.W. MacDonald, eds. A
Handbook on Biotelemetry and Radio-Tracking. Per-
gamon Press, Oxford.

698

Radiotelemetry

34

FOOD HABITS

Allen Y. Cooperrider

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

Editor's Note: An adequate food supply is one of the
most basic habitat requirements of all species.
Therefore, some understanding of food habits is a
prerequisite to evaluating habitat and providing
good habitat management. Similarly, forage utili-
zation is an important factor in habitat monitor-
ing. The two closely related topics are discussed
in this chapter, emphasizing herbivore food habits
and vegetation utilization.

"Food is the burning question in animal society, and
the whole structure and activities of the community
are dependent upon questions of food supply."

— Charles Elton, from Animal Ecology

INTRODUCTION

Measuring food habits and forage utilization may
appear to be rather specialized monitoring tech-
niques; however, such information is frequently re-
quired in monitoring studies, particularly when
working with herbivores. These animals are capable
of destroying, damaging, or altering their habitat
through overgrazing or overbrowsing. Such impacts
may affect not only the species itself but many oth-
ers as well. Similarly, the species responsible for the
damage may be either a wild or domestic ungulate,
or combinations of one or more of these. On public
lands of western North America, where many species
of wild and domestic ungulates coexist on ranges
being managed for multiple uses, information on
forage utilization and food habits can be quite
valuable.

Furthermore, the need for an understanding of
food habits of species other than ungulates will prob-
ably increase. Twenty years ago, Gullion (1966)
suggested the importance of understanding food hab-
its of game birds. Today, biologists are charged with
managing hundreds of vertebrate species, of which
the majority of food habits are largely unknown.

In this chapter, I describe methods for measur-
ing food habits and forage utilization, discuss prob-
lems associated with such measurements, and
present and evaluate methods of analyzing such data.
This chapter emphasizes techniques for herbivores
since these are more commonly required in monitor-
ing studies by management biologists. Techniques
for omnivores and carnivores are covered briefly and
reference is made to publications with more detailed
information on this subject. Similarly, techniques for
measuring forage utilization will be described in
less detail than those for food habits since utilization
techniques are well-described in numerous, readily
available publications.

In the literature on subjects such as food habits,
forage selection, preference, and utilization have
been used quite loosely and even interchangeably.
More seriously, this has often resulted in biologists

Food Habits

699

confusing concepts and stating erroneous conclu-
sions. In particular, workers have confused food hab-
its or diet with forage preference. To avoid further
confusion, I therefore begin with brief definitions of
each of these measurements.

Food habits or diets refer to the actual foods
that are eaten by one or more animals of the same
species during a given season or time period within
a defined geographic area. Thus, food habits are char-
acteristic of an animal species, but only for a given
time and place; extrapolation of food habit data to
other times and regions, without supplementary data,
is frequently misleading. Food habits are typically
expressed as percentages of the overall diet, which
consists of each individual food item (plant or ani-
mal species or species group; Figure 1 ); this is the
most useful and easily interpretable measure. The
percentages may be by weight or volume depending
on the technique. Usually the two are pretty well
correlated; however, food habits are also sometimes
reported as a frequency (percentage of samples con-
taining each food item) or, more rarely, as the
weight of each food item in the daily (or seasonal)
diet.

Forage utilization refers to the amount of vege-
tation (or food supply) removed by foraging animals
from a defined area of land during a defined period.
Utilization measurements are rarely made except
on plant forages, although the same principles apply.
They are thus attributes of the land, vegetation, or
food supply and not of an animal or species of ani-
mal. Utilization measurements are usually expressed
as the percentage of the available forage or annual
forage production removed during a given season or
year (Figure 2). They may be broken down by plant
species or group and the percentage may be by
weight or volume.

Forage utilization may also be expressed in a
variety of other ways such as the percentage of
plants or twigs that are grazed or browsed, or the
height of grazed down plants. In practice, this utiliza-
tion may be attributed to one species of animal or
to several. When several species of large herbivores
occupy the same range at the same time, determin-
ing the relative amount used by each animal could
be difficult or even impossible.

Food habits of mule deer on Trickle Mountain,
Colorado, Spring 1978.

tr

03

-C

o
b_

Food habits of bighorn sheep on Trickle
Mountain, Colorado, Spring 1978.

"Grass and grasslike plants

Muhly (Muhlenbergia sp.) 16%

Fescue (Festuca sp.) 11%

Sedge (Cyperace) 8%

Other 17%

Forbs

Sagebrush (Artemisia sp.) 22%

L Other 19%

52%

7%

Estimated food habits of elk,
Trickle Mountain, Colorado, 1978.

a GRASS

□ FORBS JAN I

■-E3 BROWSE I978

100%

JULY
1978

Figure 1. Food habits refer to the percentage of each forage species, items, or classes in the diet for a
given period and location. These habits may be depicted in a pie chart, table, or graph.

700

Food Habits

Mountain
Mahogany
30%

Wildrye

1%

Western
Wheatgrass
85%

Figure 2. Forage utilization refers to the percentage of annual production used by one or more animals.

The total amount of forage produced on an area
during a given year or season is called forage pro-
duction. This production may be used synonymously
with availability or corrected for forage physically
or seasonally unavailable to animals for various rea-
sons. Measurement of forage production and availa-
bility are not described in this book in any detail
because they are covered elsewhere (U.S. Depart-
ment of the Interior, Bureau of Land Management
1984). However, they have been introduced here to
clarify the concept of forage.

In contrast, forage preference is the relative
degree to which animals select one forage species
over another. It is thus an attribute of an animal or
species. However, to quantify this variable, both food
supply and food habits need to be measured. The
simplest measurement of forage preference is the
percentage ratio of a given food item in the diet to
the entire food supply (available forage) that is pres-
ent (Table 1 ). Clearly, both measurements must
occur at the same time and in the same geographical
area. (Food items with preference values greater
than 1 are preferred.) The important concept to re-
member is that food habits are not synonymous with
forage preference even though they are often treated
the same in much of the older wildlife and range
management literature. Put simply, forage preference
is what an animal likes to eat; food habits are what
it actually eats. For further discussion on this see
Petrides(1975).

MEASUREMENT OF FORAGE UTILIZATION

Two basic procedures are available for deter-
mining forage utilization: ( 1 ) difference methods and
( 2 ) grazed-plant methods. Since many variations on
these techniques have been developed and used,
no single technique can be prescribed as "best."

Rather, the investigator will have to determine the
most appropriate and efficient method for the prob-
lem at hand.

Difference Methods

Difference methods require estimating the
amount of forage or herbage produced with and
without grazing. The difference then is used to esti-
mate the amount used. Estimates can be obtained
in two different ways: single plots, before and after
grazing, and paired plots, one grazed the other not.

In the before-and-after approach, the amount of
forage is estimated or measured on an area before
and after a grazing period or season. Any acceptable
technique for measuring or estimating forage pro-
duction can be used; however, variances in the esti-
mates of differences can be reduced substantially
by using permanent or marked plots for both meas-
urements. In this case, the first (before) measure-
ment cannot be made by using a removal method
such as clipping. To reduce variances in utilization
estimates, yet still use clipping methods, many work-
ers have used paired plots. Using this method, a se-
ries of paired plots are selected that appear to be
similar in species composition and total forage. One
plot is then selected randomly for clipping prior to
grazing; the other is clipped at the end of the season.

In the paired plot approach, cages or other ex-
closures are used to estimate ungrazed forage. Before
the grazing season or period, paired plots with simi-
lar species composition and total forage production
are selected, and one or the other plot is randomly
chosen to be caged or excluded from grazing. At the
end of the grazing period, the amount of forage on

Food Habits

701

Table 1. Forage production, food habits, and forage preference of bighorn sheep on Trickle Mountain,
Colorado, 1978 (from Bailey and Cooperrider 1982).

Percent of

Percent in

Forage

Diet of

Preference Index

Vegetation

Production

Bighorn Sheep

for Bighorn Sheep

GRASS

Arizona fescue

12

7

0.58

(Festuca arizonica)

Blue grama

15

4

0.26

(Boutelova gracilis)

Muhly

11

10

0.91

{Muhlenbergia sp.)

Sedge

4

5

1.25

(Cyperaceal)

Other

11

12

0.92

FORBS

41

6

0.15

BROWSE

Rabbitbrush

3

*

<0.01

(Chrysothamnus sp.)

Sagebrush

1

25

25.00

(Artemisia sp.)

Saltbrush

*

11

> 100.00

{Atriplex sp.)

True Mountain Mahogany

*

6

> 100.00

(Cercocarpus sp.)

Other

2

14

7.00

Note: The preference index is calculated by dividing percent in diet by percent of forage production.

trace

both the ungrazed and grazed plot are measured
or estimated through clipping or other appropriate
methods.

Both types of difference methods have problems
and pitfalls. Obtaining good estimates of forage pro-
duction is time-consuming and expensive. If meas-
urements have to be made twice, then the costs
in time and money increase. Furthermore, forage
production estimates will yield even less precision.
Forage production estimates are rarely better than
within ± 10%. Therefore, difference methods
should not be used unless utilization is expected to
be high (50% or more). Difference methods thus
may be good for determining high rates of utilization
(50% to 100% ), or for determining if utilization is
high or not. These methods, however, are not partic-
ularly good for determining small differences in utili-
zation, for example, distinguishing between 40%
and 50% utilization.

Grazed-Plant Methods

Grazed-plant methods are used for both herba-
ceous plants and woody plants, but plants are exam-

ined only at the end of the grazing season. These
methods fall into three general categories:

( 1 ) frequency methods,

(2) height or length conversion methods, and

(3) form-class or ocular estimates.

Frequency methods require the biologist to count a
number of plants to determine how many have been
grazed or not. Previously developed regression tables
can then be used to convert the percentage of
grazed plants to percentage of utilization (Cook and
Stubbendieck 1986). Alternatively, the average
weight of grazed portions can be determined and
used to calculate utilization (U.S. Department of the
Interior, Bureau of Land Management 1984:21). This
technique can also be used with browse species by
counting the number of unbrowsed and browsed
leaders.

Height or length conversion methods require
the biologist to determine the average height that
herbaceous species have been grazed or the average
remaining leader-length for browse species. These
measurements are then converted into percentage of

702

Food Habita

Measuring browse plant leader-length.

utilization through previously developed tables or
regression equations (Cook and Stubbendieck 1986;
U.S. Department of the Interior, Bureau of Land Man-
agement 1984).

Form-class or ocular measurements are used
primarily with browse species. They are subjective,
requiring judgments about the degree of past or
current browsing of individual browse plants based
on their shapes. These judgments are then used to
infer the degree of utilization.

Problems with Utilization Techniques

Techniques for determining forage utilization
are not precise, even though many methods are rela-
tively time-consuming and expensive. Furthermore,
the accuracy of most methods is limited because
of factors such as regrowth of grazed plants, loss of
forage from factors other than grazing, and many
other causes. For a discussion of some of these fac-
tors see Martin (1970), Cook and Stubbendieck
(1986), and Pieper (1978).

Because of the high variances, many biologists
prefer utilization measurements of only a few plant
species known to be key forages. Good estimates on
a few key species are usually preferable to poor esti-
mates on many, particularly if the key species are
chosen carefully and judiciously. See U.S. Depart-
ment of the Interior, Bureau of Land Management
( 1984) for a discussion on choice of key species.

In general, utilization measurements may be
useful in answering general questions such as "Are
perennial grasses being overused on summer
ranges?"; "Is bitterbrush being used heavily in the fall
by mule deer (Odocoileus bemionus)?"; or "Are the
elk (Cervus elaphns) using grass or browse on the
winter range?" They are not, however, very useful in

determining slight variations in levels of utilization.
Furthermore, some questions such as the last one
above can be answered more easily and cheaply by
food habit studies rather than utilization studies. The
biologist must determine on a case-by-case basis
which type of study and particular method will be
most appropriate and efficient in terms of time and
money.

MEASUREMENT OF FOOD HABITS

This section concentrates on techniques ap-
plicable to estimating herbivore diets. However, par-
allel techniques for carnivores are mentioned as
appropriate.

Food habits of wild animals can be determined
through three basic methods:

( 1 ) direct or indirect observation of the animal,

(2) stomach or rumen analysis, and

(3) fecal analysis.

Direct or Indirect Observation

Direct observation of animals has been used
with varying success for many years to estimate diets
of free-ranging or tame animals (Neff 1974; Wallmo
et al. 1973; Riney 1982). Feeding on individual food
items is quantified as "bite counts" or "feeding min-
utes," resulting in data which are then converted,
with or without correction factors, to percentage in
the diet. It requires equipment no more complicated
than binoculars (or spotting scopes) and a field note-
book. This method can be extremely time-consum-
ing, however, particularly if animals are difficult to
locate or difficult to approach without flushing them.
The technique works best with diurnal or crepuscu-
lar animals that occupy relatively open habitat be-
cause biologists can approach them close enough to
observe their feeding habits without causing them to
leave. Few animals meet all these criteria, although
the technique has been successful with animals like
bighorn sheep (Ovis sp.) and in National Parks and
other areas where animals become accustomed to
the near presence of human observers.

Because of the difficulty in locating and observ-
ing animals, some biologists have preferred to work
with tame, semi-tame, or constrained animals. Typi-
cally, these animals are raised in captivity so they are
accustomed to humans. To determine food habits,
the animals are taken to an area and allowed to for-
age. The biologist then stays near the animal and
observes what it eats. There are two potential prob-
lems with using tame animals: differences between
tame and wild animals and cost.

Food Habits

703

Food habits of tame animals may be different
from those of wild animals due to ( 1 ) previous die-
tary experience; (2) nutritional status immediately
before a foraging trial (e.g., if the animal has a full
stomach or has been starved before the trial); and
(3) choice of a feeding site/foraging area by the cap-
tors rather than the animal. Some believe these dif-
ferences can be overcome. However, use of tame
animals remains an expensive method for measuring
food habits and will probably be too expensive
in most management (as opposed to research)
situations.

The precision and accuracy of data from direct
observation of animals have been questioned by
some biologists. However, some believe that direct
observation is not only adequate for management
but is adequate for measuring food habits and estab-
lishing a standard to measure the accuracy of alterna-
tive techniques (Gill et al. 1983).

An alternative method of animal observation is
the "feeding site" examination. This technique is
similar in many ways to the utilization methods pre-
viously described, except the measurement obtained
is food habits rather than forage utilization. Basically,
animals are located in the field; then the exact loca-
tion where the animal has been feeding is examined
and the plants or twigs grazed are tallied. The num-
ber of plants of each species that have been grazed is
then converted into a percentage to provide an esti-
mate of the food habits of the animal. Such feeding
can be relatively easy to observe in some situations,
such as when animals are pawing through fresh snow
to eat plants. This can also be done while trailing
animals. At other times, accurately determining what
the animal has eaten can be extremely difficult, if
not impossible. Like other observation methods,
feeding site examinations can be very time-
consuming.

Direct observation or feeding-site observations
are used primarily with herbivores. However, Riney
(1982:132) suggests that direct observation is also
easy with larger carnivores. In North America, large
carnivores are rarely found in high enough densities
such that direct observation would be feasible. Simi-
lar problems arise with smaller mammals as well as
most bird, reptile, and amphibian species. A notable
exception is diurnal raptors at nest sites. In this case,
forages brought to nest sites by parent birds can be
observed relatively easily and recorded, providing an
estimate of food habits during the important nesting
season.

Stomach or Rumen Analyses

Stomach or rumen analyses have frequently
been used to determine food habits. The technique
has been used on virtually every major group of
vertebrates ranging from songbirds and small rodents

to moose and elephants. Stomach or rumen samples
are taken from animals that have been killed, or ob-
tained in other ways; the contents are then analyzed
in the field or sent to a laboratory for later analysis.
Medin (1970) and Korschgen (1980) describe the
basic technique in detail, including preservation
of materials and identification of food items. Food
habits can then be determined and quantified by ( 1 )
tabulating numbers of each food item, (2) tabulating
frequency of occurrence of food items, ( 3 ) measur-
ing volume, or (4) measuring weight.

Microscopic techniques, including the microhis-
tological technique of Sparks and Malechek (1968)
and the microscope-point technique of Heady and
Van Dyne (1965) have also been used, primarily
with herbivores.

Stomach samples create major problems in that
they normally require shooting or otherwise sacrific-
ing animals. In addition to being time-consuming,
shooting animals for the sole purpose of studying
their food habits is rarely warranted, at least with the
larger or more economically important animals or
with any rare or endangered species.

Recently, fistula techniques have been devel-
oped to remove forage samples from the rumen or
esophagus of captive ungulates without killing them.
This allows multiple samples to be taken from the
same animal. These techniques were developed pri-
marily by range and animal scientists working with
livestock, but have been used successfully with only
a few species of wildlife. Although the technique is
useful in research, the time and money required
to raise and keep fistulated animals prohibit this
technique from being used to obtain information for
routine land and wildlife management.

Fecal Analysis

Fecal analysis has been used more for estimating
range herbivore food habits in the past 10 years
than any other procedure (Holechek et al. 1982a).
This happened in spite of the fact that the accuracy
of the technique has been seriously questioned. The
procedure is similar to stomach analyses except that
a fecal sample rather than a stomach sample is
collected.

In the laboratory, microscopic techniques are
normally used to sample herbivores, whereas macro-
scopic techniques are used for carnivores. The tech-
niques for identifying food items in the laboratory
are the same ones used for stomach samples and,
therefore, have the same advantages and disadvan-
tages in terms of time, cost, precision, and accuracy.
Typically, the management biologist need not learn
how to perform the laboratory analyses, since non-
profit regional laboratories will perform the analyses

704

Food Habits

at cost. As with many types of specialized laboratory
techniques, it is usually more efficient for the man-
agement biologist to pay for such work than to try
to do it independently.

Cooperrider et al. (1982) describe field proce-
dures for herbivores, and Korschgen ( 1980) de-
scribes the macroscopic techniques for carnivores as
well as herbivores.

Even though the fecal analysis technique for use
with herbivores has been criticized (see Holechek
et al. 1982a; Gill et al. 1983), the technique has
some important advantages. Holechek et al. (1982a)
list the following seven advantages:

• It does not interfere with the normal habits of
animals.

• It permits practically unlimited sampling.

• It places no restriction on animal movement.

• It has particular value where animals range over
mixed communities.

• It is the only feasible procedure to use when
studying secretive and/or endangered species.

• It can be used to compare the diets of two or
more animals at the same time.

• Actual sampling requires very little equipment.

In addition to these important considerations, Coo-
perrider et al. (1982) suggest that the technique
is relatively cost-effective and is a practical tech-
nique for management biologists working in govern-
ment agencies.

Holechek et al. ( 1982a), reviewing the many
criticisms of the technique to that date, compiled a
similar list of disadvantages of the fecal analysis tech-
nique for herbivores and concluded that inaccuracy
is the greatest overall limitation. More recently, Gill
et al. (1983) reiterated this criticism and suggested
that the technique be critically evaluated.

Fecal analysis techniques are commonly used
with carnivores, both large and small. Such use has
generated far less criticism and controversy, al-
though the accuracy of such procedures is not well
known either. Before discussing accuracy, I will dis-
cuss the nature of diets of wild animals since this has
an important bearing on both accuracy and preci-
sion, as well as the way such data are used.

VARIABILITY IN FOOD HABITS OF WILD
ANIMALS

A common misconception of novice biologists
and lay persons is that food habits of wild animals
are rigid or fixed. Statements like "mule deer eat
bitterbrush in winter" or "elk are grass-eaters" are

quite common. Many similar statements can be
found in scientific literature. Such statements tend to
oversimplify the real situation.

Most vertebrate animals are opportunistic in
their feeding habits and are capable of eating and
surviving on a wide variety of foods. This is in con-
trast to insects, for example, many of which have
become adapted to feeding on only one or a few
closely related plant or animal species.

Animals have limitations imposed by the struc-
ture of their digestive tracts and their ability to lo-
cate, capture, or ingest adequate quantities of certain
foods. Coyotes cannot survive on dry grass, and
horses do not thrive by eating crickets. Nevertheless,
a herbivore such as Rocky Mountain mule deer has
been reported in 99 separate studies to eat over 700
different plant species (Kufeld et al. 1973).

Given the opportunistic nature of most verte-
brates, particularly of most North American herbi-
vores, and the diversity of habitats and
environmental conditions that these animals are ex-
posed to over their geographic ranges, it is not sur-
prising that food habits vary greatly in time and
space. Cooperrider et al. ( 1980), for example, have
demonstrated that food habits of one bighorn sheep
herd vary significantly between seasons, sites, and
years. Differences between regions are likely to be
even greater. Although these sorts of phenomena
have not been analyzed explicitly for many carni-
vores, the diversity of the food habits data reported
suggests that the same is true for many of these spe-
cies. Food habits of animals are not an attribute of
a species, but rather a species in a given locality
during a specific time. Extrapolation of food habit
data to other populations, localities, or times may be
very misleading.

On the other hand, vertebrate feeding is far
from random. Overwhelming data suggest that ani-
mals of the same species have similar if not identical
preferences for food items. Thus forage preference as
opposed to food habits, to a certain extent, is an
attribute of an animal species. The prime determi-
nants of food habits, what the animal actually eats,
are forage availability and forage preference.

Forage availability is quite different from site to
site; and changes over time from removal, phenol-
ogy, weather, and other factors account for much of
the variation in food habits. Given that forage prefer-
ence is an attribute of an animal species and that it
together with forage availability are the prime deter-
minant of food habits, two important conclusions
can be drawn. First, if forage preferences and forage
availability are known or estimated, a biologist
should be able to predict, albeit very crudely, food
habits for a given site or range. Second, if food habits
for a species are known for an area, then food habits

Food Habits

705

should be the same in areas with similar forage avail-
ability. Both types of inferences require some initial
estimation of food habits, however.

ACCURACY AND PRECISION OF FOOD
HABIT MEASUREMENTS

The question of accuracy and precision of food
habit techniques continues to be argued, yet few
investigators give the subject enough priority or ex-
pend much effort in testing the techniques. How-
ever, in fairness, testing the accuracy of a technique
is extremely difficult without having a "standard"
technique that all investigators agree is both accu-
rate and precise and from which other techniques
can be evaluated. Even the obvious solution of feed-
ing animals by hand can introduce biases, in addition
to being costly and time-consuming. Therefore, there
are few conclusions about accuracy and precision of
given techniques that all biologists will agree upon.
The reader who wants or needs to pursue the sub-
ject further is encouraged to investigate the litera-
ture. Holechek et al. (1982a, 1984) provide a good
review of techniques for herbivores.

All three techniques (observation, stomach or
rumen analysis, and fecal analysis) have limited accu-
racy. Direct observations can be deceptive because
the number of "bites" or the time spent feeding on a
given plant relative to the total bites or time may
not be directly proportional to the percentage by
weight of a species that is ingested. Problems with
tame animals have already been discussed. For fur-
ther discussion see Wallmo et al. (1973) and Wallmo
and Neff (1970).

Stomach and fecal analyses have limited accu-
racy for basically the same reasons. The items
counted or sampled from stomachs or feces may not
reflect the proportion by weight in the diet. In all
three cases, correction factors have been developed
to adjust these figures. The development and use
of these correction factors, however, continues to be
more of an art than a science since they are fre-
quently not written down, recorded, or reported in
the literature. Thus it is difficult to replicate results.

Technical error, of course, is always a concern
and can be a problem with any of these techniques.
Field biologists have tried to at least check technical
or other types of errors with stomach and fecal anal-
yses by sending in replicate samples, i.e., two sam-
ples drawn from the same stomach, rumen, pellet
group, etc. This is a good recommended practice;
however, the biologist should be cautious about
interpreting such information when it is returned.
Many biologists have sent replicate samples to a
laboratory and, after receiving results that were not
"the same," have rejected the technique.

Determining if results are "the same" requires
knowledge of the normal variation among samples.
Whenever sampling from a "population" (in the sta-
tistical sense), results are expected to be different.
Most food habit techniques, including microhistolog-
ical, involve some sort of sampling; two analyses of
the same sample should not be expected to be ex-
actly the same. Depending on the precision of the
technique and sampling error, results may be quite
different. There are statistical procedures to deter-
mine the probability that two samples are really
"different" providing one has estimates of variance
or can obtain them. In summary, a basis for deter-
mining whether replicate samples are "the same" or
"not the same" should be established before a tech-
nique is rejected or accepted, based upon such
evidence.

The problem of accuracy will remain until suffi-
cient studies are devoted to quantifying the problem.
In the meantime, biologists must judge subjectively
or intuitively if accuracy of a technique is within
reasonable limits. Management biologists must also
judge if the technique will be accepted or if the
results will be deemed credible by the managers,
agencies, courts, etc., that will be using, reviewing,
or making decisions based on such information. The
accuracy of the fecal analysis technique for herbi-
vores has been criticized more than others, yet many
of the most vocal critics have continued to use it
and publish papers based on its use. As mentioned
earlier, it is also the most widely used method, and
results from such analyses have stood up in court
hearings.

Knowledge of the precision of most techniques
is also quite limited, although it is easier to quantify
and evaluate. The subject is complicated by consid-
erations of the level of precision needed. More specif-
ically, the basic sampling unit is usually one stomach
sample, one pellet group, or one day or session of
bite counts. All of these more or less estimate the
diet of an animal for 1 or 2 days. This daily diet can
usually be estimated with good precision; however,
the "daily diet" of one animal is not usually of much
interest to management.

An animal's diet may change drastically from day
to day, drainage to drainage, animal to animal, etc.
Therefore, what is usually of interest is the seasonal
or monthly food habits of a herd or some other logi-
cal grouping of individuals. Reasonable precision in
estimating this can usually be obtained if the biolo-
gist can obtain enough samples. With fecal samples,
this is usually not a problem, but with techniques
that require killing or fistulating animals, it may be
difficult or even impossible. Dearden et al. (1975),
Cooperrider et al. (1982), and Holechek and Vavra
(1983) discuss precision of microhistological and
fistula estimates of herbivore food habits.

706

Food Habits

In deciding if accuracy and precision are within
acceptable limits, the biologist must keep clearly in
mind the reason data are being collected and the
purpose of these data. Precision and accuracy re-
quired or feasible for management generally are not
as stringent as those required for research. Similarly,
precision and accuracy required to obtain prelimi-
nary insight into range utilization and foraging dy-
namics may be much less than for a statistically
sound monitoring program. For example, a few fecal
samples of deer and elk may be collected and ana-
lyzed to select plant species for intensive monitoring
of utilization on winter range. Insight gained from
limited, slightly inaccurate data may be quite helpful
in focusing a monitoring project on relevant
problems.

USE AND INTERPRETATION OF FOOD
HABITS AND FORAGE UTILIZATION DATA

Because food habits or forage utilization data are
rarely useful by themselves, the ways such data are
used (and misused) in monitoring programs and
in influencing management decisions are discussed
below.

Forage utilization data are principally used to
determine whether key plant species are being over-
used. If such a determination is made and more than
one major herbivore is present on the range, then
further study will be required to determine relative
use by various species. This can be done with food
habit studies or, if the animals are present at differ-
ent times or seasons, it can be done with more de-
tailed utilization studies.

quality of the forage and the nutritional require-
ments of the animal. In some cases, subjective
knowledge may be adequate. For example, if winter
food habits indicate bighorn sheep are eating primar-
ily winterfat and fringed sagebrush, and these are
both known to be good nutritious winter forages,
then the biologist can reasonably infer that the diet
is adequate. In other cases, measurements of nutri-
tional content of the plants may have to be obtained.
The whole area of nutritional analysis is beyond the
scope of this chapter; for herbivores, a reader inter-
ested in pursuing the subject is referred to several
good books and publications on the subject re-
viewed and listed in Zarn ( 1981 ) and Holechek et al.
(1982b).

Whether a diet is good or adequate does not
mean that range conditions are good. Animals could
be receiving an adequate diet while severely overus-
ing many plant species on the range. To determine
this, the biologist needs information on animal densi-
ties, forage production, or forage utilization.

Similarly, food habit data alone are not adequate
to determine competition for forage. Forage compe-
tition refers to common use of forage by two or
more animal species, and dual use of forages that are
limiting factors for one or more of the animal spe-
cies. Food habits of sympatric species are commonly
compared by investigating dietary overlap or similar-
ity. Dietary overlap is quantified using similarity in-
dexes such as shown in Table 2. High dietary
overlap is often used as evidence of severe competi-
tion for forage, and similarly low overlap is used to
suggest absence of forage competition. These infer-
ences may be unwarranted depending on additional
data.

In some cases, biologists attempt to use utiliza-
tion measurements to make inferences about food
habits. This procedure can be awkward and inaccu-
rate for several reasons, even where only one princi-
pal herbivore is present on the range (Martin 1970;
Cook and Stoddart 1953). For example, annual forbs
may constitute a significant portion of the herbi-
vores' diet, but if the forbs have already dried up and
disappeared at the time utilization studies are con-
ducted, such studies will not detect them and infer-
ences about food habits will be inaccurate. If food
habits need to be determined, a technique for meas-
uring food habits should be chosen.

Food habits data have two principal uses: ( 1 )
determination of the adequacy of the habitat to sup-
ply food of sufficient quality and ( 2 ) analysis of com-
petition for forage between two or more sympatric
species.

Biologists cannot determine the adequacy of the
forage supply from knowledge of food habits alone.
They must know something about the nutritional

High overlap in forage use in a noncompetitive
situation can occur on a range with an abundance of
forage preferred by two species. Conversely, low
overlap with severe competition can occur where
one animal species uses preferred and limited for-
ages in an area before a second animal species ar-
rives. This situation, sequential competition, can
occur when cattle occupy a range in summer that is
used by wild ungulates in winter. If the cattle re-
move the preferred, presumably highest quality, for-
age for the wild ungulates, overlap in food habits can
be low while competition for forage, extreme.

Both of these uses of food habit data point to
the value of understanding forage preference. A biol-
ogist can, without great risk, infer that forage prefer-
ence is well -correlated with nutritional value. Those
species that are preferred generally contain the high-
est levels of basic nutrients such as digestible energy
or digestible protein. If an animal is primarily eating
forages that are known to be highly preferred by that
species, a biologist may infer that the diet is good.
Similarly, if two animal species are present on the

Food Habits

707

Table 2. Dietary overlap between bighorn sheep and cattle, Trickle Mountain, Colorado, 1978 (from Bailey
and Cooperrider 1982).

Percent of

Vegetation

Percent of
Forage in
Cattle Diet

Forage in
Bighorn Sheep
Diet

Dietary
Overlap

GRASSES AND GRASSLIKE

PLANTS

Fescue

17

7

7

(Festuca sp.)

Muhly
(Muhlenbergia sp.)

14

10

10

Sedge
(Cyperaceal)

16

5

5

Wheatgrass
(Agropyron sp.)

21

0

0

Other

19

11

11

FORBS

4

6

4

BROWSE

Sagebrush
{Artemisia sp.)

2

25

2

Saltbrush

0

11

0

(Atriplex sp.)

Other

7

14

7

Note: Dietary overlap determined by taking the lesser of the two figures in each row as circled; total dietary overlap is 46%.

708

Food Habits

same range and are using large quantities of a forage
item that is in short supply and that is a preferred
forage for at least one animal species, then a biolo-
gist may infer that forage competition is taking place.
A thorough determination that the species are really
competing requires more information, particularly
evidence that populations of one or the other spe-
cies are limited by availability of the preferred forage
species.

Determination of forage preference requires
information on both food habits and forage availabil-
ity. This information is used in a variety of ways to
calculate preference indexes — the simplest of which,
and the only one with a straightforward interpreta-
tion, is the ratio of a given forage in the diet to the
ratio of that same forage in the habitat. Other prefer-
ence indexes include information on such factors as
frequency on the range and frequency in the diet, an
interpretation that is unclear at best (Chamrad 1979;
Krueger 1972). For a thorough review of forage
preference of ungulates, see Skiles (1984).

Food habits, as described earlier, can be consid-
ered to be primarily a function of forage preference
and forage availability. Spatial arrangement of forage
is of course important, but is difficult to quantify.
All forage availability techniques, therefore, treat
available forage as if it is equally available. Neverthe-
less, if forage preferences are known, then a biologist
should be able to predict, albeit very crudely, the
food habits of an animal species in that area. Many
biologists are capable of doing this, through knowl-
edge of forage preference and forage availability

gained by experience and undocumented observa-
tions. However, attempts to use forage preference in-
dexes and forage availability information in
quantitative models to predict actual diets have not
been very successful, at least with herbivores
(Loehle and Rittenhouse 1982). Hopefully, future
work in this area will result in preference indexes
and quantitative models that will allow the biologist
to predict diets. Developing more powerful and
useful predictive models will probably depend on
development of methods to quantify the spatial
arrangement of forage in a way that can be measured
simply and efficiently, yet is still useful for predic-
tion. Until such models are available, biologists will
need to continue to measure food habits directly.

SUMMARY

Measurement of forage utilization and food hab-
its and the use of such data in monitoring programs
are often required. However, forage selection and
utilization by wild animals is a complex process, and
most measurement techniques have limited accuracy
and precision. In addition, many techniques require
time, money, and equipment that are generally not
available in some agencies or cannot be justified
by the problems being addressed. Nevertheless,
measurements of food habits and forage require-
ments will continue to be needed for special situa-
tions. Biologists should carefully choose appropriate
techniques, considering the type of data needed,
the precision and accuracy required, the acceptabil-
ity of these data to those who must use it, and the
relative cost in terms of money and personnel.

Food Habits

709

LITERATURE CITED

CHAMRAD, AD. 1979- Preference ratings aid in food habit
evaluations and interpretations. Pages 273-276 in
Proc. First Welder Wildl. Foundation Symp.

COOK, C.W. AND L.A. STODDART. 1953 The quandary of
utilization and preference. J. Range Manage.
6:329-335.

AND J. STUBBENDIECK 1986. Methods of measur-
ing herbage and browse utilization. Pages 1 20- 121
in Cook, C.W. and J. Stubbendieck, eds. Range Re-
search: Basic Problems and Techniques. Soc. Range
Manage., Denver, CO. 317pp.

COOPERRIDER, AY., J.A. BAILEY, AND R.M. HANSEN.
1982. Cost efficient methods of estimating ungulate
food habits by fecal analysis. Pages 399-406 in Arid
Land Resource Inventories: Developing Cost-Efficient
Methods. U.S. Dep. Agric, For. Serv. Gen. Tech. Rep.
WO-28. 620pp.

, S.A. MCCOLLOUGH, AND J.A. BAILEY. 1980. Varia-
tion in bighorn sheep food habits as measured by
fecal analysis. Pages 29-41 in Proc. Biennial Symp.
Northern Wild Sheep and Goat Council, Salmon, ID.
668pp.

DEARDEN, B.L., RE. PEGAU, AND R.M. HANSEN. 1975.
Precision of microhistological estimates of ruminant
food habits. J. Wildl. Manage. 39:402-407.

GILL, R.B., L.H. CARPENTER, R.M. BARTMANN, D.L.

BAKER, AND G.G. SCHOONVELD. 1983. Fecal analy-
sis to estimate mule deer diets. J. Wildl. Manage.
47:902-915.

GULLION, G.W. 1966. A viewpoint concerning the signifi-
cance of studies of game bird food habits. Condor
68:372-376.

HEADY, HE AND G.M. VAN DYNE. 1965. Prediction of
weight composition from point samples on clipped
herbage. J. Range Manage. 18:144-149-

HOLECHEK, J.L. AND M. VAVRA. 1983. Fistula sample
numbers required to determine cattle diets on forest
and grassland ranges. J. Range Manage. 36:323-326.

, , AND R.D. PIEPER. 1982a. Botanical com-
position determination of range herbivore diets: A
review. J. Range Manage. 35:309-315.

, , AND . 1982b. Methods for deter-
mining the nutritive quality of range ruminant diets: A
review. J. An. Sci. 54:363-376.

, , AND . 1984. Methods for determin-

ing the botanical composition, similarity and overlap
of range herbivore diets. Pages 425-471 in Develop-
ing Strategies for Rangeland Management. Westview
Press, Boulder, CO. 2022pp.
KORSCHGEN, L.J. 1980. Procedures for food-habits anal-
yses. Pages 113-127 in Schemnitz, S.D., ed. Wildl.

Manage. Techniques Manual, 4th Ed., The Wildl. Soc,
Washington, DC. 686pp.

KRUEGER, W.C. 1972. Evaluating animal forage prefer-
ence. J. Range Manage. 25:471-475.

KUFELD, R.C., O.C WALLMO, AND C. FEDDEMA. 1973.
Foods of the Rocky Mountain mule deer. U.S. Dep.
Agric, For. Serv., Res. Pap. RM-111. 31pp.

LOEHLE, C. AND L.R. RITTENHOUSE. 1982. An analysis of
forage preference indices. J. Range Manage.
35:316-319.

MARTIN, S.C. 1970. Relating vegetation measurements to
forage consumption by animals. Pages 93-100 in
Range and Wildlife Habitat Evaluation — A Research
Symposium. U.S. Dep. Agric, For. Serv., Misc. Publ.
1147. 220pp.

MEDIN, D.E. 1970. Stomach content analyses: Collections
from wild herbivores and birds. Pages 133-145 in
Range and Wildlife Habitat Evaluation — A Research
Symposium. U.S. Dep. Agric, For. Serv., Misc. Publ.
1147. 220pp.

NEFF, D.J. 1974. Forage preferences of trained mule deer
on the Beaver Creek watershed. Arizona Game and
Fish Dep., Spec. Rep. 4. 6lpp.

PETRIDES, G.A. 1975. Principal foods versus preferred
foods and their relations to stocking rate and range
condition. Biol. Conserv. 7: 161-169.

PIEPER, R.D. 1978. Methods for measuring utilization.

Pages 123-147 in R.D. Pieper, ed. Measurement Tech-
niques for Herbaceous and Shrubby Vegetation. Dep.
Animal and Range Science, New Mexico State Univ.,
Las Cruces. 147pp.

RINEY, T. 1982. Study and management of large mammals.
John Wiley and Sons, New York, NY. 552pp.

SKILES, J.W. 1984. A review of animal preference. Pages
153-213 in Developing Strategies for Rangeland Man-
agement. Westview Press, Boulder, CO. 2022pp.

SPARKS, DR. AND J.C. MALECHEK 1968. Estimating per-
centage dry weight in diets using a microscopic tech-
nique. J. Range Manage. 21:264-265.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF

LAND MANAGEMENT. 1984. Rangeland Monitoring—
Utilization studies. U.S. Dep. Inter., Bur. Land Manage.,
Tech. Ref. 4400-3. 105pp.

WALLMO, O.C, R.B. GILL, L.H. CARPENTER, AND D.W.
REICHERT. 1973- Accuracy of field estimates of deer
food habits. J. Wildl. Manage. 37:556-562.

AND D.J. NEFF. 1970. Direct observation of tamed

deer to measure their consumption of natural forage.
Pages 105-109 in Range and Wildlife Habitat Evalua-
tion— A Research Symposium. U.S. Dep. Agric, For.
Serv., Misc. Publ. 1147. 220pp.

ZARN, M. 1981. Wild ungulate forage requirements — A
review. U.S. Dep. Inter., Bur. Land Manage. 103pp.

710

Food Habits

35

WEATHER AND
CLIMATE

James A. Bailey

Department of Fishery and Wildlife Biology
Colorado State University
Fort Collins, CO 80523

"Climate plays an important part in determining the
average numbers of species, and periodical seasons
of extreme drought or cold I believe to be the most
effective of all checks."

Charles Darwin, from The Origin of the

Species

"Climate is what we expect, weather is what we
get."

Robert Heinlein

Editor's Note: Weather is an extremely important
factor to consider in inventories or monitoring
studies. In the western U.S., extreme fluctuations in
weather are quite normal. Weather affects wildlife
directly and indirectly: it can alter population size
or cause changes in animal behavior that will bias
study results.

This chapter identifies weather factors and describes
how they are used to enhance the precision of stud-
ies and promote early detection of trends in wild-
life populations and habitats.

INTRODUCTION

Weather affects wild animals directly by causing
mortality, or by influencing behavior and physiology
in ways that influence reproduction or mortality.
Weather also affects wildlife indirectly by altering
the production or availability of habitat resources
and by influencing competitors, predators, and dis-
ease organisms. Consequently, much year-to-year
variation in wildlife abundance and in the productiv-
ity or condition of wildlife and wildlife habitat is due
to stochastic (random) variation in weather.

In addition, weather may affect the observability
or activity of wildlife and thus may affect data ob-
tained for wildlife monitoring. The objective of mon-
itoring is to detect trends of improvement or decline
in wildlife or in wildlife habitat. However, the sto-
chastic direct and indirect effects of weather on
wildlife, and the stochastic effects of weather on
monitoring data, will reduce precision of the moni-
toring system and may delay detection of directional
trends in wildlife populations or in habitat.

The objectives of this chapter are to describe
some effects of weather on wildlife and on monitor-
ing data, and to promote the use of weather data as a
covariant in analyzing wildlife-monitoring data. Use
of covariant analysis can enhance precision and early
detection of directional trends in wildlife popula-
tions and habitats.

WEATHER AND WILDLIFE

Direct and Indirect Effects of
Weather on Wildlife

A variety of direct and indirect impacts of
weather on wild vertebrates has been reported (Ta-
ble 1 ). The smaller species, in particular, may die
from acute exposure to severe weather. Mortality of
larger species during severe weather may result from
a combination of direct exposure of the animals or
indirect effects on the availability of habitat re-
sources. Drought or severe runoff events may drasti-
cally alter current velocity, turbidity, and amount of

Weather and Climate

711

Table 1. Some reported effects of weather upon wildlife and on wildlife-monitoring data.

Mammals

Birds

Amphibians, Reptiles

Fish

1. Direct effects:

Jacobs and Dixon

Yocum 1950

Larimore 1959

2 1. Acute mortality

1981

Jones 1952

Smith 1960

Jehl and Hussell 1966

Stout and Cornweil

1976
Graber and Graber

1979
Stout and Cornweil

1976
Graber and Graber

1979

Jojhn 1964
Tramer 1977
Matthews and Styron
1981

2. Exposure with

Martmka 1967

Nehring and

habitat loss

Meslow and Keith

1971
Connolly 1981

Anderson 1984

3. Effects on

Wight and Conaway

Hammond and

Starrett 1951

reproduction

1961
Picton 1984

Johnson 1984

Gagnark and Bakkala

1960
John 1963
Crecco and Savoy

1984

4. Effects on

Lovaas 1958

Graber and Graber

habitat

Richens 1967

1979

selection

5 Effects on

Gilbert et al. 1970

Rutherford 1970

Obert 1976

Stehr and Branson

movement

Bruns 1977

1938

patterns

Starrett 1951
Rinne 1975
May et al. 1976
Ottaway and Clark
1981

II. Indirect effects:

Smoliak 1956

Gullion 1960

Hoddenbach and

1. Upon habitat

Gallizioli 1965

Francis 1970

Turner 1968

Shaw 1965

Brown and Smith

Turner et al. 1969,

Beatley 1969, 1976

1980

1976

Day 1971

Medica et al. 1975

Shiflet and Dietz 1974

Whitford and Creusere

Turkowski and Vahle

1977

1977

Nixon et al. 1980

Stephenson and

Brown 1980

Mueggler 1983

2. Upon

Longhurst and

Vaught et al. 1967

competition

Douglas 1953

Graber and Graber

or disease

1979

712

Weather and Climate

Table 1. Some reported effects of weather upon wildlife and on wildlife-monitoring data (concluded).

Mammals

Birds

Amphibians, Reptiles

Fish

III. Effects on

monitoring

data:

Lovaas 1958

Diem and Lu 1960

Obert 1976

1 . Changed

location

Gilbert et al. 1970

Rutherford 1970

affects

Graber and Graber

duration

1979
Best 1981

2. Persistence or

Wallmo et al. 1962

Emlen and DeJong

detectability of

1981

sign

3. Activity and

Newman 1959

Gullion 1966

Lillywhite 1982

detectability of

Morse and Balser

Robel et al. 1969

animals

1961
Progulske and Duerre

1964
LeResche and

Rausche 1974

Brown 1971
LaPerriere and

Haugen 1972
Porter 1973
Morrison and Slack

1977
Shields 1979
Conner and Dickson

1980
Bock and Root 1981
Dawson 1981
Robbins 1971

4. Trap results

Bailey 1969
Phillips et al. 1972

Raveling 1966

Vogt and Hine 1982

Kirkland 1962
May et al. 1976
Hubert and Schmitt
1982 a b

5. Observer

Bock and Root 1981

diligence

Dawson 1981

suitable fish habitat in streams and rivers, often caus-
ing acute mortality or downstream transport of fish.
Weather may affect animal behavior in ways that in-
fluence reproduction, habitat selection, or movement
patterns and distribution. Effects of weather on ani-
mal condition may persist for several months after a
weather event (Clutton-Brock and Albon 1983).

Impacts of weather on the production or availa-
bility of habitat resources, especially forage, have
been described. In arid areas, great variation in pre-
cipitation, and consequently in the quality and abun-
dance of forage, has been correlated with reproduc-
tive success of mammals, birds, and reptiles. Weather
may also influence wildlife indirectly by altering
competition or by influencing exposure to disease.

The direct and indirect effects of weather on
wildlife tend to be more acute (capable of quick ef-
fect) and severe for small animals than for large ani-
mals, and for populations living near the edges of
their geographic ranges (Siivonen 1956; Birch 1958;
Graber and Graber 1979).

Winter-stressed mule deer (when the ears are collapsed,
the animal cannot be saved).

Weather and Climate

713

Effects of Weather on Wildlife Data

Numerous reports exist on weather conditions
influencing the results of wildlife monitoring activi-
ties (Table 1 ). Weather may influence the concentra-
tion of wildlife and thus the probability that animals
will be detected and counted. Weather may alter the
locations of animals among habitats or among re-
gions of the geographic range. Consequently, animals
may move into habitats where they are less detecta-
ble or into habitats or regions away from the areas
being monitored.

Weather factors have modified the persistence
of wildlife sign and the transmission of sounds used
in census methods. Activity patterns of animals have
been altered in ways that influence detection and ob-
servability (Figure 1 ). Success of trapping animals
and even sex-age structures of trapped samples of
animals may be determined by weather. Wildlife ob-
servers may become less diligent during poor
weather.

In response to these sources of bias, ornitholo-
gists have stressed a need to schedule bird surveys
so that data are collected only during suitable
weather. However, this option is not available in all
wildlife monitoring. A few authors (Robel et al.
1969; Best 1981; Dawson 1981 ) have suggested that
"adjustment factors" might be developed to account
for weather bias. This approach is proposed in this
chapter.

40-

o
EC

30 -

..••'

0)

a

-1^<^^ ' '

•o
m

20

. • * '

c

CO

a

10

I

I I I I

f40

45

I I I I
50 55 60 65 70 75

Temperature (Fahrenheit)

80°

Figure 1. Effect of air temperature 1/2 hour before
sunrise on numbers of cardinals {Cardinalis
cardinalis) counted on standard breeding-bird
survey routes. Dashed lines indicate 95%
confidence limits for data in each 5°-interval set
( after Robbins 1981 ).

WEATHER AND WILDLIFE MONITORING

Wildlife monitoring produces time-series data.
These data provide information on trends in popula-
tion size, density, or distribution; in population qual-
ity; or in habitat condition (Hocutt and Stauffer
1980; Worf 1980). Trends are usually measured over
several years. Whereas the objective of wildlife mon-
itoring is to detect persisting trends in wildlife or
habitats, weather causes temporary changes in wild-
life or habitats, or temporary bias in the monitoring
data, that vary stochastically (improvement or de-
cline) from year to year.

These temporary effects of weather may obscure
long-term trends, especially if a biologist hopes to
detect trends from a relatively short time-series of
wildlife or habitat data. For example, 1 or 2 years of
favorable weather may enhance a wildlife population
even though the long-term trend in habitat condition
or population is downward. Consequent delay in de-
tecting the downward trend may result in several
years' delay in developing a management response.

This problem can be diminished if effects of one
or a few important weather factors on wildlife can
be determined and measured during the early years
of monitoring. Once the wildlife -weather relation-
ships have been quantified, the temporary effects of
weather can be isolated from long-term trends in
time-series data. For example, although the indicated
trend in wildlife or habitat condition for the most re-
cent year or two may not be downward, the recent
conditions may be poor when compared with only

714

Weather and Climate

those previous years having similar weather
conditions.

In the same manner, positive responses of wild-
life or habitat to management programs can be de-
tected more quickly if year-to-year effects of variable
weather can be identified.

Selecting the Key Weather Factor

There are no rules to follow in selecting key
weather factors for use in data analysis. Seven sug-
gestions are-

( 1 ) Use local or regional "conventional wisdom."

( 2 ) Consider those weather factors known to affect
production or availability of limiting habitat
resources .

( 3 ) Remember that autumn populations of species
with high biotic potentials may be determined
primarily by spring and summer weather affect-
ing reproductive success, rather than by winter
weather affecting overwinter survival and
spring-breeding densities.

(4) Consider weather factors known to bias the dis-
tribution or observability of animals and there-
fore the accuracy of the monitoring system.

(5) Consider the amount of seasonal precipitation
in arid areas as a key factor.

(6) Consider depth, persistence, and quality of
snow in northern and alpine areas as key
factors.

(7) Note that plants often respond to cumulative
weather effects, whereas small wildlife at the
edges of their geographic ranges are most apt to
be directly affected by acute weather
phenomena.

Key factors will vary among species, regions,
and even sites (Mueggler 1971, 1983), and among
monitoring systems. Selecting the key factor will be
a local decision.

Conventional Wisdom. Much local and regional
"conventional wisdom" exists regarding effects of
weather on many species of wildlife. These widely
accepted but often unquantified impacts of weather
on wildlife are frequently used to determine the tim-
ing of yearly wildlife studies so as to gather data un-
der similar weather conditions each year. When
these attempts to standardize weather conditions fail,
the assumed weather effects are used to subjectively
interpret the resulting trends in wildlife data. For ex-
ample, in northern regions, population indices based
on "trend counts" of ungulates along fixed routes
(aerial or ground counts) are often obtained during
that part of each winter when snow is deepest and
most widespread. When this strategy is successful,
animals are presumed to be equally concentrated

and therefore equally distributed for being counted
each year. When the strategy fails, as in a winter
without much snow, trends in the counts are usually
presumed to be due largely to the weather effect
and are therefore ignored. Thus, failure to quantify
the weather effect results in complete loss of 1
year's data.

Limiting Habitat Resources. Each vertebrate spe-
cies has its own set of habitat requirements, and hab-
itat resources vary temporally and geographically.
Consequently, there is much taxonomic, temporal,
and geographic variation in which habitat resources
limit wildlife populations. However, if the limiting
resources are known for a local population, then — by
definition — the performance of the population (e.g.,
abundance, reproductive success, or survival ) will be
correlated with variation in the limiting resources.
Furthermore, population performance may be corre-
lated with weather factors that influence the produc-
tion or availability of the limiting resources. These
weather factors, perhaps precipitation influencing
forage production or snow depth determining forage
availability, may then be used in analyzing trends of
population performance.

Biotic Potential. Species with high biotic potentials,
such as small game or songbirds, have high rates of
population turnover. In a favorable season of repro-
duction, these species may produce large autumn
populations, regardless of variation in spring-breed-
ing abundance (Steen 1944). A biologist monitoring
autumn populations may use weather factors affect-
ing the breeding season and reproductive success in
analyzing trends in the population. However,
weather factors that determine overwinter survival
and spring-breeding density will not be useful when
spring-breeding density is not well correlated with
autumn abundance.

Bias. Weather will influence the distribution, habitat
selection, movement patterns, and activity of wild-
life, and consequently may influence the results of
monitoring efforts. Effects of snow conditions on
trend counts of ungulates are described above. Nu-
merous studies have shown effects of weather on re-
sults of bird surveys (Table 1 ), despite the fact that
these surveys are usually conducted within a narrow
range of weather conditions. Weather may be known
to affect observability of animals, or even sex-age
classes of animals, and thus may affect measures of
abundance or of population composition. Such
weather factors are likely candidates for use in
analyses.

Seasonal Precipitation. Several studies have shown
wildlife data to be correlated with seasonal precipita-
tion, especially in arid areas (Shaw 1965; Gallizioli
1965; Day 1971; Medica et al. 1975; Turner et al.
1976; Whitford and Creusere 1977; Brown and
Smith 1980) where primary production is stimulated

Weather and Climate

715

by thresholds of soil moisture and temperature (Beat-
ley 1974). Most authors have used total precipitation
for 6 or more months preceding the wildlife survey
as a weather factor in data analysis.

Snow. In northern or alpine areas, snow depths,
quality, and persistence have been correlated with
wildlife data (Severinghaus 1947; Loveless 1967; Gil-
bert et al. 1970; Bartman and Steinert 1981; Adams
and Bailey 1982). When information on daily snow
depths is available, a good measure of snow impacts
on animals and on habitat is the total number of win-
ter days with more than a critical depth of snow on
the ground. Critical snow depths for deer have been
determined to be 38-50 cm (15-20 in.), about the
height of a deer's brisket (Severinghaus 1947; Bart-
man and Steinert 1981 ). Similar depths may be esti-
mated for other ungulates. However, for non-ungu-
lates, and for study areas with modest snow
accumulation, this criterion may not discriminate
among years very well. (There may be very few days
with > 38 cm [15 in. ] of snow on the ground in any
year; yet there may be considerable important varia-
tion among years in persistence of snow depth <38
cm [15 in.]). In these situations, the mean snow
depth for all winter days is an arbitrary criterion that
will discriminate among years.

Cumulative vs. Acute Effects. When monitoring
production of a habitat resource, such as forage, or a
wildlife variable closely dependent on production of
a habitat resource, consider selecting a key weather
factor based on cumulative weather effects. For ex-
ample, forage production may depend upon cumula-
tive rainfall over 2-to-several months before and dur-
ing the growing season (Fisher 1924). Negative
effects of early-season weather upon forage produc-
tion may be completely overridden by positive late-
season effects. Consequently, forage production
would not be correlated with weather data for a cer-
tain month, but may be correlated with cumulative
weather data for 2 or more months.

In contrast, when monitoring a wildlife popula-
tion that is expected to be directly affected by
weather during a short period, the key weather fac-
tor correlated with population data is apt to be ( 1 )
attainment of a threshold of weather severity at any
time during a long season, or ( 2 ) weather during a
short season when the animals are especially vulner-
able each year. The wildlife involved are apt to be
small-bodied species, and populations near the edges
of their geographic ranges (Graber and Graber
1979). For example, a severe period any time during
the winter may acutely depress populations of small

Snow depths greater than 18 inches (45.7 cm) hamper mule deer and pronghorn movements.

716 Weather and Climate

birds, and a mild late-winter period cannot compen-
sate once birds have died.

In this respect, large wild mammals are interme-
diate between plant and small wildlife. Large mam-
mals draw upon substantial body reserves during se-
vere weather and are less subject to acute effects of
weather than are small wildlife. Population data for
large mammals are often correlated with weather
data based on cumulative effects.

Local Verification of Key Weather Factors

Selection of a key weather factor for use in ana-
lyzing wildlife monitoring data should be based upon
a demonstrated correlation between the weather fac-
tor and the local wildlife or habitat data. One may
assume that correlations based on conventional wis-
dom, on assumptions about limiting habitat re-
sources, or on results from other areas will be cor-
rectly applicable to a new monitoring project. But
this assumption should be tested with local time-
series data before assuming there is much confidence
in the results.

When a time series of wildlife data is available,
it is tempting to assess numerous weather conditions
for possible correlations with the wildlife data. Very
many intercorrelated weather parameters can be
generated from weather records. For instance, snow
can be represented by several measurements relating
to thresholds, persistence, and constancy, and each
of these measurements can be tabulated for several
separate as well as overlapping periods each winter.
Correlations of each of these resulting parameters
with the monitoring data can be assessed easily with
computers. Some, often several, "statistically signifi-
cant" correlations usually are found. Some of these
may represent real direct or indirect relationships
between weather and wildlife. Other correlations are
probably due only to chance, since statistical signifi-
cance is usually based on an error rate (probability
of rejecting the hypothesis of no correlation when
it is true ) that applies separately to each weather
parameter.

Can future wildlife monitoring data be evaluated
by using any of these correlated weather parameters
as covariants? Meslow and Keith ( 1971 ) discussed
some aspects of this problem. If there is a biologi-
cally plausible hypothesis regarding a correlation be-
tween weather and the wildlife data, if the hypothe-
sis is developed before any correlation in the data is
evaluated subjectively or statistically, if only a few
weather measurements are used to represent the
weather parameter, and if a conservative (small) er-
ror rate is used, there should be little risk in accept-
ing the correlation as a basis for evaluating future
monitoring data. Otherwise, any strong correlations
between a weather parameter and the monitoring

data should be viewed as a new hypothesis to be tested
( 1 ) for biological plausibility and ( 2 ) with several years
of new data, before the indicated relationship is used in
assessing the results of monitoring.

COVARIANT ANALYSIS

In the examples below, linear correlation and
regression are used to estimate relationships be-
tween wildlife characteristics and weather factors. In
reality, relationships such as these may be curvilin-
ear, and abrupt thresholds may exist. With suitable
statistical methods, non-linear relations between
weather and wildlife may be used in covariance anal-
ysis of monitoring data. However, detection and veri-
fication of non-linear relationships will usually re-
quire abundant and precise estimates of both
variables, and these are not common in wildlife mon-
itoring. Consequently, most analyses will involve lin-
ear methods.

Example 1 — Browse Utilization and Snow Depth

During the 1960s, the Montana Fish and Game
Department measured winter utilization of service-
berry (Amelanchier alnifolia ) browse by white-
tailed deer {Odocoileus virgmianus) in the Goat
Creek drainage of the Swan Valley. Utilization was
measured on several marked plants at the end of
each winter. Presumably, browse utilization was
being monitored as an index to the relative balance
between numbers of deer and the availability of
browse. A trend toward higher levels of utilization
would indicate either increasing numbers of deer or
a declining availability of browse and might require a
management response. The purpose here, however,
is only to address the problem of promptly detecting
significant trends in levels of utilization when those
levels fluctuate up and down each year ( Figure 2 ).
These data are extracted from Hildebrand ( 1971 ).

In 1965, 74% of scrviceberry twigs on the tran-
sect had been browsed — the highest level of utiliza-
tion in 6 years (Figure 2). But had a significant
change in utilization occurred?

In the North, wintering white-tailed deer are
very sensitive to snow conditions and tend to con-
centrate especially in conifer stands, as snow be-
comes deep. Browse transects were typically placed
in these areas of concentration, often called "key
areas." Consequently, among-years variation in utili-
zation of browse on key areas often represented vari-
ation in concentration of deer caused by annual vari-
ation of snow conditions.

The number of days with >48 cm (19 in.) of
snow on the ground each winter, recorded at the
nearby Seeley Lake Ranger Station, was significantly

Weather and Climate

717

correlated with percentage utilization of serviceberry
for the years 1960-64 (Figure 3). Based on the
regression for those years, the expected utilization of
serviceberry in 1965, when there had been 69 days
with > 48 cm ( 19 in.) of snow on the ground, was
52.5 ± 12.3% (95% conf. limits). The observed uti-
lization was 74% . Considering the regression from
1960 to 1964, the probability of this 74% utilization
being due only to sampling error was <0.03. (I have
calculated probabilities using untransformed utiliza-
tion data, justified for 1960-64 when percentages
were near 50.) This was strong evidence that some-
thing had changed in the balance between deer
numbers and browse resources at Goat Creek. As it
turned out, 1965 was the first of several years that
exhibited comparatively high levels of utilization at
Goat Creek (Figure 3)

In 1965, a biologist monitoring browse utiliza-
tion in the Swan Valley may have considered 74%
utilization on serviceberry as a signal for a closer
look at conditions in the Valley. Certainly in 1966,
when utilization was 78%, with only 25 days of deep
snow, the biologist should have been convinced that
something was amiss; and use of data on snow
depths as a covariant in analyzing a time-series of
browse-utilization data would have enhanced prompt
detection of the change.

100

o\

V— '

>
cr
oc

LU
CO

LU

O
>

DC
LU
C/)

LL

o

<

N

K
3

50

1960 1961 1962 1963 1964 1965

Tagged branch of browse plant.

Figure 2. Utilization of serviceberry browse by
white-tailed deer on the Goat Creek transect,
Swan Valley, Montana (Data from Hildebrand
1971)

Furthermore, use of this covariant analysis of
monitoring data would have enhanced detection of
an impact of land-use practices on a wildlife popula-
tion. Clear-cutting had begun on private lands in the
Swan Valley in 1955. Winter range was being elimi-
nated, concentrating deer on smaller amounts of
range (Hildebrand 1971).

Example 2 — Age Ratios and Snow Depth

Mountain goats (Oreamnos americanus) were
introduced to the Sawatch Range, Colorado, in 1948
and onto Sheep Mountain in that range in 1950 (Ad-
ams and Bailey 1982). Age ratios of the Sheep Moun-
tain herd have been observed as an index to monitor
reproductive success since 1966. Age ratios varied
greatly among years. Bailey and Johnson (1977)
noted that these ratios for 1966-74 were correlated
with the depth of snow each previous May at a high-
elevation location observed by the U.S. Soil Conser-
vation Service (Figure 4). It is biologically plausible
that depth and persistence of snow on winter ranges
will increase demands upon nanny goats and de-
crease forage availability, and that these impacts dur-
ing late gestation might influence reproductive suc-
cess, especially survival of neonates.

The Sheep Mountain population of mountain
goats grew steadily during 1965-7-4. Most rapid
growth apparently occurred after 1970 (Adams and
Bailey 1982). Based on concepts of density -depen-
dence, especially with introduced ungulates (Caugh-
ley 1970), a decline in reproductive success (repro-
duction + neonatal survival ) was expected. A

718

Weather and Climate

100-

£

1967* .1969

>
cr
en

UJ
CD
UJ

o

1968

1965.

1964. ^*-^i

OF SERV
o

1960* ^ ■***" ^\

^*~ 1960-1964

Z

O
H
<

N

_J
H

. -- ' 1 963 Y = 36.8 + 0.23 X (p < 0.03)

1961

0-

c

i i i i

1 25 50 75 100

NO. OF DAYS WITH > 48 CM SNOW ON GROUND

Figure 3. Use of snow-depth at Seeley Lake,

Montana, as a covariant in analysis of browse-
utilization data. A significant change in the
relationship began in 1965.

decline in age ratios was first suggested in 1976, and
age ratios for 1976-79 were significantly different
from age ratios for 1966-75 (Figure 5; Adams and
Bailey 1982). To test this hypothesis, an "artificial
variable" representing the two time periods was used
in a multiple-regression analysis (X = 0 for 1966-75;
X = 1 for 1976-79). At a mean snow depth of 97
cm (39 in.) for the 14 years, age ratios had declined
by 47% in 1976-79 compared with 1966-75.

Note that the lowest age ratios had occurred in
1968 and 1970, before the decline indicated above.
But 1968 and 1970 were very severe winters. It is
only after using a weather variable in covariant anal-
ysis, so that winters of similar severity may be com-
pared, that the decline in age ratios becomes
obvious.

OBTAINING WEATHER DATA

Existing Records

Finklin (1983) described the sources of climatic
data available to wildland managers. I have ab-
stracted his suggestions.

Year-round weather data are published by the
U.S. Weather Bureau and its successor agencies in-
cluding the National Oceanic and Atmospheric

0.8 -

CO

N^

_i

N. •

< 0.7 -

\'66

\ •

< 0.6 -

• \ '74

cr

LU

'67 K

'69 \

Q 0.5-

^v

_i

^v

O

Ny

co 0.4.

• N.

Q

'72 \.

X.

\^

0.3 -

N.

O

>.

h-

>v

< 0.2-

• * \

oc

'68 '70

_l

_l

< 0.1 -

LL

0.0 -

1 1 1 1 1 1 1

0 10 20 30 40 50 60 70

SNOW DEPTH ON 1 MAY Cinches)

Figure 4. Relationship between 1 May snow depth
from the previous winter at Monarch Pass and
kidrolder animal ratios from fall aerial surveys of
mountain goats on Sheep Mountain and vicinity
(Bailey and Johnson 1977).

.'66

• 1966-75

07.

<

2
z

.'67

<

cr

06.

Q

'69 v^ <

74

O

CO

05.

• 72 \

•75

o

• 77

H

0.4.

<

cr

UJ

'•<76

O

<

cc

•

LU
2

78

2

Z>

02.

CO
LU

79

•68 -70

H

5

01 .
0.0

(

) 30

60 90 120

150 180

1 MAY SNOW DEPTH (cm)

Figure 5. Relationship between 1 May snow depth
from the previous winter at Monarch Pass and
kid:older animal ratios of mountain goats on
Sheep Mountain-Gladstone Ridge. Sawatch
Range, Colorado, 1966-79 (Adams and Bailey
1982).

Weather and Climate

719

Administration (Hatch 1983). These data are avail-
able in many major libraries or from the National
Climatic Center, Federal Building, Asheville, NC
28801. Information from local stations may be found
in monthly and annual "Climatological Data" sum-
maries for each state. Each station provides daily pre-
cipitation and maximum and minimum temperatures.
Some stations also record evaporation, soil tempera-
tures, monthly windspeed, relative humidity, sun-
shine, snowfall, and snow depth.

Fire -weather observations, maintained by the
U.S. Forest Service, provide an additional data base,
but for the fire season only. These include afternoon
temperature, relative humidity, wind, daily precipita-
tion, and daily maximum and minimum tempera-
tures. Such data are archived on tapes at the National
Fire Weather Data Library, Fort Collins, CO, and are
available via access to the USDA computer at the

Fort Collins Computer Center. (See Finklin 1983 for
additional sources.)

Snow survey data for 1 1 western states are avail-
able from the USDA Soil Conservation Service, Port-
land, OR 97209 or from state offices of the Soil Con-
servation Service. Periodically, state or regional data
are compiled into a Summary of Snow Survey
Measurements.

Stream flow records, especially for major rivers,
may be obtained from state or regional offices of the
U.S. Geological Survey. These are compiled for states
or for major watersheds within states and are in
some libraries. Records include historic average
rates, extreme rates, and daily rates.

See Finklin ( 1983) for other, minor sources of
weather data.

Solar-powered weather station.

Climatological data summaries.

720

Weather and Climate

Obtaining Local Data

It may be desirable to monitor one or a few
weather variables on a study area as part of the mon-
itoring program. This would occur if existing
weather stations are not nearby of if they do not ad-
equately represent the study area because of differ-
ences in elevation, aspect, habitat, or topography.
Also, key weather factors may not be recorded at the
nearby cooperating Weather Bureau station. (For ex-
ample, not all stations maintain records of snow
depths. ) Descriptions of commonly used weather in-
struments, of weather stations, and of weather data
collection and analysis are available from Day and
Sternes (1970) and Critchneld (1974).

Wildlife biologists have developed a few special
instruments for measuring impacts of weather on an-
imals. Although deep snow may hinder movements
of large mammals, snow quality — in particular, crust-
ing— may diminish or exacerbate problems of mov-
ing in snow, depending on the density and location
of crusts in the snowpack. Hepburn ( 1978) de-
scribed a penetrometer for measuring the effective
depth of snow for large mammals when the snow-
pack is crusted. Energy demands of wintering big
game may depend, in part, on the combined effects
of temperature, wind, and radiation exchange with
the environment (Moen 1968, 1973). Verme (1968)
has described a simple device for measuring these
combined effects upon energy demands.

Weather and Climate

721

LITERATURE CITED

ADAMS, L.G. and J.A. BAILEY. 1982. Population dynamics

of mountain goats in the Sawatch Range, Colorado. J.

Wildl. Manage. 46:1003-1009.
BAILEY, J A. 1969. Trap responses of wild cottontails. J.

Wildl. Manage. 33:48-58.
and B.K JOHNSON. 1977. Status of introduced

mountain goats in the Sawatch Range of Colorado.

Pages 54-63 in Samuel, W. and W.G Macgregor, eds.

Proc. First Int. Mountain Goat Symp. Kalispell, MT.
BARTMANN, R.M. and S.F. STEINERT. 1981. Distribution

and movements of mule deer in the White River

Drainage, Colorado. Colorado Div. Wildl. Spec. Rept.

51. 12pp.
BEATLEY, J.C. 1969. Dependence of desert rodents on

winter annuals and precipitation. Ecology 50:721-724.
. 1974. Phenological events and their environmental

triggers in Mohave desert ecosystems. Ecology 55:856-

863.

1976. Rainfall and fluctuating plant populations in

relation to distributions and numbers of desert ro-
dents in southern Nevada. Oecologia 24:21-42.

BEST, LB. 1981. Seasonal changes in detection of individ-
ual bird species. Studies in Avian Biology 6:252-261.

BIRCH, L.C. 1958. The role of weather in determining the
distribution and abundance of animals. Cold Spring
Harbor Symposia in Quantitative Biology 22:203-215.

BROWN, D.E. and R.H. SMITH. 1980. Winter-spring pre-
cipitation and population levels of blue grouse in Ari-
zona. Wildl. Soc. Bull. 8:136-141.

BROWN, W.H. 1971. Winter population trends in the red-
shouldered hawk. Am. Birds 25:813-817.

BRUNS, EH. 1977. Winter behavior of pronghorns in rela-
tion to habitat. J. Wildl. Manage. 41:560-571.

CAUGHLEY, G. 1970. Eruption of ungulate populations,
with emphasis on Himalayan thar in New Zealand.
Ecology 51:53-72.

CLUTTON-BROCK, T.H. and S.D. ALBON. 1983- Climatic
variation and body weight of red deer. J. Wildl. Man-
age. 47:1197-1201.

CONNER, R.N. and J.G. DICKSON. 1980. Strip transect
sampling and analysis for avian habitat studies. Wildl.
Soc. Bull. 8:4-10.

CONNOLLY, G.E. 1981. Limiting factors and population
regulation. Chapter 7 in Wallmo, O.C., ed. Mule and
Black-tailed Deer of North America. Univ. of Nebraska
Press, Lincoln.

CRECCO, V.A. and T.F. SAVOY. 1984. Effects of fluctua-
tions in hydrographic conditions on year-class
strength of American shad (Alosa sapidissima) in the
Connecticut River. Can. J. Fish. Aquatic Sci. 41:1216-
1223-

CRITCHFIELD, H.J. 1974. General climatology (3rd Edi-
tion). Prentice-Hall, Inc. 446pp.

DAWSON, D.G. 1981. Counting birds for a relative mea-
sure (index) of density. Studies in Avian Biology 6:12-
16.

DAY, G.I. 1971. Wildlife research in Arizona, 1970-71.
Pages 153-157 in Wildlife Research in Arizona. Ari-
zona Game and Fish Dept. Phoenix.

DAY, J A. and G.L. STERNES. 1970. Climate and Weather.
Addison-Wesley Publ. Co. Reading, MA. 407pp.

DIEM, KL. and KH. LU. I960. Factors influencing water-
fowl censuses in the Parklands, Alberta, Canada. J.
Wildl. Manage. 24:113-133.

EMLEN, J.T. and M.J. DEJONG. 1981. The application of
song detection threshold distance to census opera-
tions. Studies in Avian Biology 6:346-352.

FINKLIN, A.I. 1983. Summarizing weather and climatic
data — a guide for wildland managers. U.S. Dep. Agric,
For. Serv. Gen. Tech. Rep. INT- 148. 43pp.

FISHER, R.A. 1924. The influence of rainfall on the yield
of wheat at Rothamsted. Royal Soc. London Phil.
Trans., Series B. 213:89-142.

FRANCIS, W.J. 1970. The influence of weather on popula-
tion fluctuations in California quail. J. Wildl. Manage.
34:249-266.

GAGNARK, H.A. and R.G. BAKKALA. I960. A comparative
study of unstable and stable ( artificial channel ) spawn-
ing streams for incubating king salmon at Mill Creek.
California Fish and Game 46:151-164.

GALLIZIOLI, S. 1965. Quail research in Arizona. Arizona
Game and Fish Dep., Phoenix. 12pp.

GILBERT, P.F., O.C. WALLMO, and R.B. GILL. 1970. Effect
of snow depth on mule deer in Middle Park, Colorado.
J. Wildl. Manage. 34:15-23.

GRABER, J.W. and R.R. GRABER. 1979. Severe winter

weather and bird populations in southern Illinois. Wil-
son Bull. 91:88-103.

GULLION, G.W. I960. The ecology of Gambel's quail in
Nevada and the arid southwest. Ecology 41:518-536.

. 1966. The use of drumming behavior in ruffed

grouse population studies. J. Wildl. Manage. 30:717-
729.

HAMMOND, M.C. and D.H. JOHNSON. 1984. Effects of
weather on breeding ducks in North Dakota. U.S. Dep.
Inter., Fish and Wildl. Serv. Tech. Rep. 1. 17pp.

HATCH, W.L. 1983- Selective guide to climatic data

sources. Key to Meteorological Records Documenta-
tion No. 4.11. National Oceanic, Atmospheric Admin.
Asheville, NC. 338pp.

HEPBURN, R.L. 1978. A snow penetration gauge for stud-
ies of white-tailed deer and other northern mammals.
J. Wildl. Manage. 42:663-667.

HILDEBRAND, PR. 1971. Biology of white-tailed deer on
winter ranges in the Swan Valley, Montana. M.S. The-
sis, Univ. Montana. Missoula. 91pp.

HOCUTT, C.H. andJ.R. STAUFFER, JR., eds. 1980. Biologi-
cal monitoring of fish. Lexington Books. Lexington,
KY. 41 6pp.

HODDENBACH, G.A. and F.B. TURNER. 1968. Clutch size
of the lizard Ufa stansburiana in southern Nevada.
Amer. Midi. Nat. 80:262-265.

HUBERT, W.A. and D.N. SCHMITT. 1982a. Factors influ-
encing catches of drifted trammel nets in a pool of the
Upper Mississippi River. Proc. Iowa Acad. Sci. 89:153-
154.

and . 1982b. Factors influencing hoop net

catches in channel habitats of Pool 9, Upper Missis-
sippi River. Proc. Iowa Acad. Sci. 89:84-88.

JACOBS, D. and KR. DIXON. 1981. Breeding-season pre-
cipitation and the harvest of cottontails. J. Wildl. Man-
age. 45:101 1-1014.

JEHL, J. and D. HUSSELL. 1966. Effects of weather on re-
productive success of birds at Churchill, Manitoba.
Arctic 19:185-191.

JOHN, KR. 1963. The effect of torrential rains on the re-
productive cycle of Rhinichthys osculus in the Chiri-
cahua Mountains, Arizona. Copeia 2:286-291.

. 1964. Survival of fish in intermittent streams of the

Chiricahua Mountains, Arizona. Ecology 45:112-119.

722

Weather and Climate

JONES, G. 1952. Hail damage to wildlife in southwest
Oklahoma. Wilson Bull. 64:166-167.

KIRKLAND, L. 1962. A tagging experiment on spotted and
largemouth bass using an electric shocker and the Pe-
tersen disc tag. Proc. Southeast. Assoc. Game and Fish
Comm. 16:424-432.

LA PERRIERE, A J. and A.O. HAUGEN. 1972. Some factors
influencing calling activity of wild mourning doves. J.
Wildl. Manage. 36:1193-1199.

LARIMORE, R W. 1959. Destruction and re-establishment
of stream fish and invertebrates affected by drought.
Trans. Am. Fish. Soc. 88:261-285.

LE RESCHE, RE. and R.A. RAUSCH. 1974. Accuracy and
precision of aerial moose censusing. J. Wildl. Manage.
38:175-182.

LILLYWHITE, H.B. 1982. Tracking as an aid in ecological
studies of snakes. Pages 181-191 in Scott, NJ., Jr., ed.
Herpetological Communities. U.S. Dep. Inter., Fish and
Wildl. Serv. Wildl. Res. Rep. 1 3.

LONGHURST, W.M. andJ.R. DOUGLAS. 1953. Parasite in-
terrelationships of domestic sheep and Columbian
black-tailed deer. Trans. North Am. Wildl. Nat. Resour.
Conf. 18:168-188.

LOVAAS, A.L 1958. Mule deer food habits and range use,
Little Belt Mountains, Montana. J. Wildl. Manage.
22:275-283.

LOVELESS, CM. 1967. Ecological characteristics of a mule
deer winter range. Colorado Div. Game, Fish and
Parks. Tech. Publ. 20. Denver, CO. 1 24pp.

MARTINKA, C.J. 1967. Mortality of northern Montana
pronghorns in a severe winter. J. Wildl. Manage.
31:159-164.

MATTHEWS, W.J. and J.T STYRON, JR. 1981. Tolerance of
headwater vs. mainstream fishes for abrupt physico-
chemical changes. Am. Midi. Nat. 105:149-158.

MAY, N., L. TRENT, and P.J. PR1STAS. 1976. Relation offish
catches in gill nets to frontal periods. Fishery Bull.
74:449-453.

MEDICA, PA., R.B. BURY and F.B. TURNER. 1975. Growth
of the desert tortoise ( Gopherus agassizi ) in Nevada.
Copeia 4:639-643-

MESLOW, E.C. and LB. KEITH. 1971. A correlation analy-
sis of weather versus snowshoe hare population pa-
rameters. J. Wildl. Manage. 35:1-15.

MOEN, A.N. 1968. Surface temperatures and radiant heat
loss from white-tailed deer. J. Wildl. Manage. 32:
38-344.

. 1973- Wildlife Ecology. W.H. Freeman and Co. San

Francisco, CA. 458pp.

MORRISON, ML. and R.D. SLACK. 1977. Population trends
and status of the Olivaceous Cormorant. Am. Birds
31:954-959.

MORSE, M.A. and D.S. BALSER. 1961. Fox calling as a
hunting technique. J. Wildl. Manage. 25:148-154.

MUEGGLER, W.F. 1971. Weather variations on a mountain
grassland in southwestern Montana. U.S. Dep. Agric,
For. Serv. Res. Pap. INT-99. 25pp.

. 1983. Variation in production and seasonal devel-
opment of mountain grasslands in western Montana.
U.S. Dep. Agric, For. Serv. Res. Pap. INT-316. 16pp.

NEHRING, R.B. and R. ANDERSON. 1984. Recruitment
and survival of young-of-the-year brown trout {Salmo
trutta L. ) in the South Fork of the Rio Grande River
versus parent spawner density, stream discharge and
fry habitat. Proc. Colorado- Wyoming Chapt. Am. Fish.
Soc. 19:35-43.

NEWMAN, D.E. 1959. Factors influencing the winter
roadside count of cottontails. J. Wildl. Manage.
23:290-294.

NLXON, CM., M.W. McCLAIN, and LP. HANSEN. 1980. Six
years of hickory seed yields in southwestern Ohio. J.
Wildl. Manage. 44:534-539.

OBERT, H. 1976. Some effects of external factors upon the
reproductive behavior of the grass frog, Rana tempor-
aria. Oecologia. 24:43-55.

OTTAWAY, E.M. and A. CLARK 1981. A preliminary inves-
tigation into the vulnerability of young trout {Salmo
trutta L. ) and Atlantic salmon {Salmo salar L) to
downstream displacement by high water velocities. J.
Fish. Biol. 19:135-145.

PHILLIPS, R.L., R.O. ANDREWS, G.L. STORM, and R.A.

BISHOP. 1972. Dispersal and mortality of red foxes. J.
Wildl. Manage. 36:237-248.

PICTON, H.D. 1984. Climate and the prediction of repro-
duction of three ungulate species. J. Applied Ecol.
21:869-879.

PORTER, D.K 1973- Accuracy in censusing breeding pas-
serines on the shortgrass prairie. M.S. Thesis, Colorado
State Univ. Ft. Collins. 107pp.

PROGULSKE, DR. and DC. DUERRE. 1964. Factors influ-
encing spotlighting counts of deer. J. Wildl. Manage.
28:27-34.

RAVELING, D.G. 1966. Factors affecting age ratios of sam-
ples of Canada geese caught with cannon nets. J.
Wildl. Manage. 30:682-691.

RICHENS, V.B. 1967. Characteristics of mule deer herds
and their range in northeastern Utah. J. Wildl. Manage.
31:651-666.

RINNE, J.N. 1975. Changes in minnow populations in a
small desert stream resulting from naturally and artifi-
cially induced factors. Southwest. Nat. 20:185-195.

ROBBINS, C.S. 1981. Bird activity levels related to
weather. Studies in Avian Biology 6:301-310.

ROBEL, R.J., D.J. DICK and G.F. KRAUSE. 1969. Regres-
sion coefficients used to adjust bobwhite quail whistle
count data. J. Wildl. Manage. 33:662-668.

RUTHERFORD, W.H. 1970. The Canada geese of south-
eastern Colorado. Colorado Div. Game, Fish and Parks.
Tech. Publ. 26. 65pp.

SEVERLNGHAUS, CW. 1947. Relationships of weather to
winter mortality and population levels among deer in
the Adirondack region of New York. Cong. 12:212-
223.

SHAW, H. 1965. Investigations of factors influencing deer
populations. Pages 125-143 in Wildlife Research in Ar-
izona, 1964. Arizona Game and Fish Dep. Phoenix.

SHIELDS, W.M. 1979. Avian census techniques: an analyti-
cal review. Pages 23-51 in Dickson, J.G., R.N. Conner,
RR. Fleet, J.C Kroll, and J.A. Jackson, eds. The Role of
Insectivorous Birds in Forest Ecosystems. Academic
Press. New York, NY.

SHIFLET, T.N. and HE. DIETZ. 1974. Relationship be-
tween precipitation and annual rangeland herbage
production on southeastern Kansas. J. Range Manage.
27:272-276.

SHVONEN, L. 1956. The correlation between the fluctua-
tions of partridge and European hare populations and
the climatic conditions of winters in southwest Fin-
land during the last thirty years. Papers on Game Res.,
Finnish Game Foundation 17:1-30.

SMITH, A. I960. Hail: great destroyer of wildlife. Audubon
Mag. 62:170-171, 189.

Weather and Climate

723

SMOLIAK, S. 1956. Influence of climatic conditions on for-
age production of shortgrass rangeland. J. Range Man-
age. 9:89-91.

STARRET, W.C. 1951. Some factors affecting the abun-
dance of minnows in the Des Moines River, Iowa.
Ecology' 32:13-27.

STEEN, M.O. 1944. The significance of population turn-
over in upland game management. Trans. North Am.
Wildl. Conf. 9:331-335.

STEHR, W.C. and J.W. BRANSON. 1938. An ecological
study of an intermittent stream. Ecology 19:294-
310.

STEPHENSON, R.L. and D.E. BROWN. 1980. Snow cover
as a factor influencing mortality of Abert's squirrels. J.
Wildl. Manage. 44:951-955.

STOUT, I.J. and G.W. CORNWELL. 1976. Nonhunting mor-
tality of fledged North American waterfowl. J. Wildl.
Manage. 40:681-693-

TRAMER, E.J. 1977. Catastrophic mortality of stream fishes
trapped in shrinking pools. Am. Midi. Nat. 97:469-
478.

TURKOWSKI, F.J. and JR. VAHLE. 1977. Desert rodent
abundance in southern Arizona in relation to rainfall.
U.S. Dep. Agric, For. Serv. Res. Note RM-346. 4pp.

TURNER, F.B., PA. MEDICA, JR. LANNON, JR., and G.A.
HODDENBACH. 1969. A demographic analysis of
fenced populations of the whiptail lizard, Cnemido-
phorns tigris, in southern Nevada. Southwest. Nat.
14:189-202.

and B.W. KOWALEWSKY. 1976. Energy

utilization by a desert lizard (Uta stansburiana).
US/IBP Desert Biome Monogr. 1. 57pp.

VAUGHT, R.W., H.C. MCDOUGLE, and H. H. BURGESS.
1967. Fowl cholera in waterfowl at Squaw Creek Na-
tional Wildlife Refuge. Missouri. J. Wildl. Manage.
31:248-253.

VERME, L.J. 1968. An index of winter weather severity for
northern deer. J. Wildl. Manage. 32:566-574.

VOGT, R.C. and R.L MINE. 1982. Evaluation of techniques
for assessment of amphibian and reptile populations in
Wisconsin. Pages 201-217 in Scott, N.J. , Jr., ed. Herpe-
tological Communities. U.S. Dep. Inter., Fish and Wildl.
Serv. Wildl. Res. Rep. 1 3

WALLMO, O.C., A.W.JACKSON, T.L. IIAILEY, and R.L
CARLISLE. 1962. Influence of rain on the count of
deer pellet groups. J. Wildl. Manage. 26:50-55.

WHITFORD, W.G. and F.M. CREUSERE. 1977. Seasonal and
yearly fluctuations in Chihuahuan Desert lizard com-
munities. Herpetologica 335-4-65.

WIGHT, H.M. and C.H. CONAWAY. 1961. Weather influ-
ences on the onset of breeding in Missouri cottontails.
J. Wildl. Manage. 25:87-89.

WORF, D.L., ed. 1980. Biological monitoring for environ-
mental effects. Lexington Books, Lexington, KY. 227
pp.

YOCUM, C.F. 1950. Weather and its effect on hatching of
waterfowl in eastern Washington. Trans. North Am.
Wildl. Nat. Resour. Conf. 15:309-319.

724

Weather a/id Climate

VI ANALYSIS AND
PRESENTATION

36 Data Management

37 Statistical Analysis

38 Habitat Evaluation Systems

39 Evaluation and Interpretation

40 Economic Analysis

41 Written Communications

42 Verbal Presentations

m mm

36

DATA
MANAGEMENT

Larry Peterson and Iris Matney

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

"A frequent predicament of field workers is to accu-
mulate so many notes that time is lacking to analyze
them, or to have notes string out over such a long
period that the earlier ones are lost or hard to segre-
gate by the time a sufficient volume are at hand to
warrant a conclusion."

— Aldo Leopold, Game Management ( 1948)

Editor's Note: The field of data management is rap-
idly growing and changing. Computers are already
widely used in business and science and are used
increasingly in resource management. Furthermore,
computers are no longer merely number crunchers;
data processing now also includes image process-
ing and word processing. While data management
technology is quite sophisticated, its application
is becoming more and more user friendly. This
chapter provides an overview of the processes in-
volved in data management, but the reader will
have to keep up with specifics. A major obstacle to
the new user of modern computer systems is the
jargon associated with it. This chapter introduces
many of the terms and concepts used in the field, so
that the biologist can communicate with the data-
processing specialist when necessary.

INTRODUCTION

Aldo Leopold (1948) warned biologists about
problems of data management. This same problem
exists today. By using computers, wildlife biologists
can devote their attention to the collection of data
and their time to the interpretation and analysis of
information. They can provide options or recommen-
dations to management while a computer performs
the tasks of adding, subtracting, sorting, and other
data manipulations. The computer is an important
tool in the management of our natural resources.

Data processing for the wildlife biologist can
mean very simple or complex procedures. It can
mean dealing with a whole new group of people
who have their own jargon and who may not be fa-
miliar with the biologist's needs or activities. This
chapter discusses some of the most elemental con-
cepts of data management. We have suggested some
basic references at the end of the chapter to expand
on the topics presented here.

The data processing field is now frequently
called information management. It is relatively new
and has been developing rapidly. The first large-scale
electronic digital computer was built in 1946. It
contained 18,000 vacuum tubes and was pro-
grammed by connecting various wires between units
of the computer and setting up 6,000 switches so
the program would run (Shelly and Cashman
1980:2.3) By today's standards, this computer was
big, slow, unreliable, and difficult to program.

By 1954, the transistor brought about the sec-
ond generation of computers which were faster,
smaller, and less expensive. In 1964, third generation
computers using Solid Logic Technology (SLT) be-
gan to appear.

'Current Address: Tckton Software, Inc., 1495 Canyon Blvd.. Suite
100, P.O. Box 7300, Boulder, CO 80306

Data Management

727

The minicomputer and the new era of inte-
grated electronics began with the use of silicon
chips. Small processing units have brought the com-
puter into many small businesses, homes, and
schools. In fact, some colleges require incoming
students to have microcomputers. It is a rare indi-
vidual, even a wildlife biologist, who has not been
affected in some way by a computer. The Inter-
nal Revenue Service, utility companies, grocery
stores, payroll departments, and organizations — all
use computers.

A relatively new development in computer tech-
nology, of particular interest to wildlife biologists
and resource managers, is the use of Geographic
Information Systems (GISs). A GIS consists of soft-
ware and hardware for performing spatial analyses.
Thus, the operations that biologists do on maps, such
as calculating areas, amount of edge, distance from
waters, etc., can now be automated. This technology

is not discussed in detail in this chapter; however,
the principles described apply equally well to such
systems as to the traditional systems that process
alphanumeric data. For further information on GISs,
see de Steiguer and Giles ( 1981 ), Steenhof ( 1982),
and Mayer (1984).

A common problem in thinking about data man-
agement is to consider data only as numerical data,
and computers or automated systems as only "num-
ber crunchers" or systems for dealing with numeri-
cal data. Spatial data or maps analyzed by GISs are as
much data as text processed on word processors.
Unfortunately, the technologies for automating such
data have developed separately in many cases, such
that they must be used independently. However, the
future trend is to develop integrated software and
hardware so the user can easily use all three types of
data in one system, just as biologists now use maps,
tables, and written reports. The following discussion

MICROPROCESSORS AND MICROCOMPUTERS

A microprocessor is a one-chip central pro-
cessing unit (CPU). That is, all of the components nec-
essary to perform the CPU functions are contained on
one silicon chip. The small, relatively inexpensive mi-
croprocessors are used in a variety of products, includ-
ing hand-held calculators, vending machines, and
microwave ovens.

A microprocessor together with additional chips
containing circuitry for timing, temporary storage, and
interfaces for input and output devices forms a micro-
computer. First used in industrial automation, the
microcomputer has moved into a diverse environment
ranging from business and banking to transportation
and manufacturing. Most personal and home computers
are microcomputers. A microcomputer will be able to
fill many of a biologist's computing needs. However,
complex programs or large amounts of data may re-
quire larger computers to process the data efficiently.

MINICOMPUTERS

Minicomputer is a term which was first used to
describe a machine that was smaller and less expensive
than mainframe computers typically manufactured at
the time. Since that time, the range in size, ability, and
cost of computers has increased so much that the term
has lost its original significance. Although the term is
still used, its meaning varies greatly. A machine called a
minicomputer by some users may be referred to as a
small computer by others. Similarly, some people may
call a microcomputer a minicomputer. The usual dis-
tinction between a "micro" and a "mini" is whether or
not the processing unit is on one chip or many.

MAINFRAME COMPUTERS

Full-scale computers, termed mainframe com-
puters, are capable of processing large amounts of data
at very fast speeds with access to billions of characters
of data. Depending on the amount of processing done
at one location, several operators working in shifts may
be required. Also, an alternate power supply and ade-
quate air conditioning are generally required.

GEOGRAPHIC INFORMATION SYSTEM (GIS)

A Geographic Information System (GIS) stores,
displays, retrieves, and analyzes spatial data. It is similar
to other automated information systems except that it
handles spatial data instead of just words and numbers.
Since spatial data can be associated with specific geo-
graphic locations, the system can model or simulate
land uses and resource values. Many different systems
are available either as public domain software (not
copyrighted) or as proprietary systems (copyrighted
and sold under license) that must be purchased. MOSS
(Map Overlay Statistical System) is a public domain
system that is used by the U.S. Bureau of Land Manage-
ment and several other agencies. ARC/INFO, on the
other hand, is a proprietary system, available only under
license from the manufacturer.

728

Data Management

of data management applies to all types of data even
though the examples given deal mostly with alpha-
numeric data.

A computer consists of a group of devices that
process data without intervention by people at every
step. The machinery that makes up the computer
system is called hardware. Software or programming
is merely a set of instructions telling the machine
what to do.

Computers can do only a limited number of
things. Programs must be written to tell them ex-
actly what to do and when to do it. Computers do
these functions quickly and reliably and are espe-
cially good for repetitive tasks, such as —

( 1 ) Mathematical tasks, including addition, sub-
traction, multiplication, and division;

(2) Logical operations, which may include com-
paring two values to determine whether one
is larger or smaller than the other; and

(3) Input/output operations, such as accepting
data for processing or outputting data in a
printed report.

A wildlife biologist does these same tasks. There
is nothing a computer can do with data that cannot
be done by hand. But a computer does calculations
and comparisons much faster than wildlife biologists.
In fact, all the data processing and computational
work done by computers in the world today can no
longer be done by hand: about 400 billion people
would be needed to tackle this workload (Shelly and
Cashman 1980:2.1).

The computer is as available as trucks, aerial
photos, or helicopters, and it makes people more ef-
ficient. One does not need to know how to manufac-

WORD PROCESSING

One of the major uses of microcomputers by many
businesses is word processing. Systems range from the
editing typewriter to communicating word processors
that provide file transfer or electronic mail capability.
Some systems also have the ability through other soft-
ware to perform some data processing applications such
as payroll and accounting.

HARDWARE, SOFTWARE, AND FIRMWARE

Computer functions are often discussed in terms
of hardware and software. Hardware is a term used
for the physical components of the system. Input de-
vices, output devices, main storage, and auxiliary stor-
age devices are all part of the computer hardware.
Software is the collection of programs ( groups of writ-
ten instructions) that tell the computer what to do.
Each computer system has an operating system (pro-
gram) to manage scheduling, input, computing, and
output of the application programs it processes. An ap-
plication program performs a specific task such as pro-
ducing a wildlife inventory report or moving wildlife
inventory data to storage. Programs that perform com-
mon procedures such as sort routines are called utility
programs. These are "packaged" programs that per-
form functions common to many organizations such as
accounting, statistics, and modeling.

With the extensive advancement in computer
hardware and software, the distinction between the two
is not always easy to define. Therefore, the term firm-
ware is used to refer to a combination of hardware and
software used to control the operations of a computer.
Firmware instructions (microcode) are used to control
some of the permanently wired circuits (called hard-
wired circuits).

COMPUTER SYSTEM

A computer system consists of input/output units,
a processing unit, and secondary storage devices. Data
are entered into input units such as card readers, mag-
netic tape drives, or key-entry terminals. Input units
convert data to a form understandable to the computer.
The data are then processed by a Central Processing
Unit (CPU). The results of this processing are known
as output. Output may be in a form understandable
to people, such as printed reports, or it may remain in a
form, readable only by computers, for input to another
computer, such as data stored on tape. Output is made
available to people through printers, magnetic tape
drives, or terminals.

Secondary
Storage

Input

Processing

Control

ALU*

Main
Storage

-Output

Computer system

Arithmetic/logic unit*

Data Management

729

ture computers or to be a programmer any more
than one needs to know how to repair a transmis-
sion or fly a helicopter to use those tools. Other
specialists are available to assist with these tasks. It is
important, however, to recognize computer capabili-
ties and limitations.

Modern biologists often have computers or ter-
minals on their desks or in an adjoining room. To
benefit from these, the biologists need to learn how
to turn them on, use program applications, and exe-
cute certain commands.

Many computer programs are being designed to
be "user friendly." Knowing complicated program-
ming languages is no longer necessary. In many in-
stances, specific steps are outlined in user guides.

The computer is part of a technology that is be-
coming as important to the wildlife biologist as other

common tools in the natural resource field. The
wildlifer can use computers to assist in analyzing
large volumes of data and make better decisions
regarding wildlife habitats. Many wildlife manage-
ment organizations already use computers to process
wildlife species or habitat data. To get the informa-
tion required, biologists must decide —

• what data to collect;

• how much data to collect;

• how to store and maintain the data;

• how to process the data.

An understanding of computers will affect these
decisions.

Information processing remains a rapidly devel-
oping field, whether it be hardware, laser printers,

COMPONENTS

A computer consists of five components:

( 1 ) Input units

(2) Main storage

(3) Arithmetic/logic unit (ALU)

(4) Control unit

(5) Output units

Input units are used to enter data into the com-
puter. Data are usually entered through a keyboard,
optical scanner, or other device. They are transferred to
computer memory for processing and may then be
stored on media such as magnetic tape or disk.

The ALU and control unit make up the central
processing unit (CPU). These are the electronic com-
ponents that cause processing to occur by interpreting
instructions, performing calculations, moving data
around in main computer storage, and controlling in-
put/output operations.

The control unit directs and coordinates the entire
computer system. This includes the entry and removal
of data from main storage. The control unit may coordi-
nate such activities as rewinding a tape reel or routing
data from main storage to the ALU.

The ALU performs arithmetic and logic operations
on data. The arithmetic circuitry performs calculations
and related functions such as rounding. The logic cir-
cuitry performs logical operations such as comparing
two sets of characters or testing a number to see if it is
positive, negative, or zero.

Main storage (also called memory or internal stor-
age) is frequently considered a part of the CPU. It is
somewhat like an electronic filing cabinet with address-
able locations. Each location is capable of holding data
or instructions. This allows for certain locations in main
storage to be used in specific ways, such as for calcula-
tion and instruction interpretation. When main storage
is not considered part of the CPU, it is referred to as
one of the secondary or auxiliary kinds of storage avail-
able.

Output units are used to obtain information in
various forms. Some types of output are produced as
printed reports or microfiche. Output is also displayed
on video screen terminals.

DATA PROCESSING CYCLE

The data processing cycle consists of inputting,
processing, and outputting data, which transforms data
into information that is organized and usable.

The input step identifies the data to be processed
and the instructions to be used for processing. During
processing, data may be checked for completeness
and accuracy. If they are not valid, the output will have
no value. This is the basis for the term "garbage in —
garbage out." Therefore, at this step, it is important to
validate the data entered so the user can make neces-
sary corrections before other processing is completed.

During the processing step, manipulations or
computations will be performed according to instruc-
tions. For example. State codes could be expanded from
an input abbreviation (NV) to output a full name (Ne-

730

Data Management

computers small enough to fit in a hand, or better
software programs. Few people know what the fu-
ture holds, but if history is any indication, significant
new developments will continue and will affect how
people will do their work.

Because data management is a rapidly changing
field, we have kept this chapter very general; users
should keep up with the latest references or ask
computer experts for the most up-to-date informa-
tion. Not every wildlifer needs to be a statistician or
programmer, but he or she should become familiar
enough to converse with statisticians and data pro-
cessing personnel to know when more highly trained
assistance in these fields is needed.

This chapter presents some factors to consider
when working with data processing or information
management. We have tried to keep the main text as
free from undefined jargon as possible. Many special

topics and concepts are described in sidebars, and
special terms are defined in the glossary. We have
not discussed particular program or machine opera-
tions, as these are best covered in detailed manuals
or user guides — and any discussion would soon be-
come outdated.

USING THE COMPUTER

Accurate, timely information is vital in making
decisions. Computers are often used to process data
into the information required. Computers are helpful
when —

( 1 ) A large volume of data is to be processed;

(2) Complex comparisons are involved;

vada), or a code for an animal species (ODHE) will
be output as mule deer. An example of a computation
would be the accumulation of each animal species in
each State.

The output step provides the processed informa-
tion. During the data processing cycle, the data are
organized so they will be meaningful to those who re-
ceive them. Because decisionmaking occurs in many
areas and at many levels, one person's information may
be another person's data.

Processing

Information

Stored

X

Printed

Data processing cycle

To illustrate the processing of data, consider a
hypothetical study a wildlife biologist might conduct to
determine the number of mule deer, elk, and desert
bighorn sheep tagged for a special study in several
States. Our example contains only 10 records so the
relationship between input and output will be easier to
see. We used the following codes for the input data:
NV for Nevada, NM for New Mexico, and UT for Utah.
The animal species codes are ODHE for mule deer,

CEEL for elk, and OVCAME for desert bighorn sheep.
The data are as follows:

NV

ODHE

UT

CEEL

NV

ODHE

NV

CEEL

NM

CEEL

NM

OVCAME

UT

OVCAME

NV

ODHE

UT

CEEL

UT

OVCAME

Only after the above data were processed would
they become useful information. The output might
appear as follows:

Number o-f Vertebrate Species Tagged by State
as oi 1/1/84

State

Species

Num

Nevada

Mule Deer
Elk

3

1

New Mexico

Elk

Desert Bighorn Sheep

1
1

Utah

Elk

Desert Bighorn Sheep

2
2

Notice that the total for all species is 10, which is
the same as the number of records input.

Data Management

731

(3) Calculations need to be performed;

(4) Reordering (resequencing) is required.

Any time data are processed, the following
series of functions must be performed, regardless of
whether the process is performed manually or
automatically:

( 1 ) Define required data elements,

(2) Identify data sources,

(3) Develop a collection process,

(4) Define where data are to be processed,

(5) Define programming requirements,

(6) Define data entry procedures,

(7) Define input procedures,

(8) Identify processing requirements,

(9) Define output procedures.

When data are processed by a computer, these
functions can be performed by a systems analyst,
programmer analyst, or wildlife biologist. These
functions are discussed further in the following
paragraphs.

Define Required Data Elements

To identify data required for input, the biologist
should identify and analyze each item on the output
document. One way is to list each item, then analyze
it to determine if the data are read in, calculated, or
generated. Figure 1 provides an example of an out-
put document. The data elements requiring analysis
are numbered 1 through 7.

DATA REPRESENTATION

Data must be in a form recognizable to the com-
puter to be accepted by a data processing system. Sym-
bols are used to represent data. The symbols on this
page are recognizable by people. They represent a
language. Electronic signals are the symbols used in
computer language. When people communicate with a
computer, they must use special devices to translate
people-language into machine-language.

Data going into a computer are recorded in var-
ious ways, depending on the type of input device being
used. They can be recorded as punched holes on input
media such as cards, magnetic spots on magnetic tape
and disk, characters and lines on paper documents, and
so on. Not all data are recorded on an intermediate
input media such as cards, magnetic tape, or disk. On
some terminals, they may go directly to main storage
with no input medium.

When data are entered through an input device,
such as a card reader or a terminal, the device trans-
lates each character of input into a series of electronic
signals. A signal, called a bit, is in one of two states — on
or off, just like a light bulb. When individual signals are
grouped together as a set of signals, they are called a
byte. Each byte represents an addressable location
in the computer's main storage. When data have been
transferred to a location in main storage, they can be
located and used in calculations or manipulations as
necessary.

A character also is represented by one byte. Each
character has a different set of bits. One character is
different from another according to the bits which are
on or off within its set. The combination of bits and the

number of bits used to represent a character make up
a code. Because codes used by computers vary, a code
set is used to convert data from one form of representa-
tion to another. Codes that represent alphabetic, nu-
meric, and special characters form a code set. Examples
of commonly used code sets are Extended Binary
Coded Decimal Interchange Code (EBCDIC),
American Standard Code for Information Inter-
change (ASCII), and Binary Coded Decimal (BCD).

Codes are also used to indicate the end of data.
The computer accepts all these signals and stores them
in main storage. They can then be referenced and ma-
nipulated within the processing unit. When processing
is completed, the electronic signals can be converted
back into characters readable by people through output
devices. Because of the similarity between the input
and output functions, some devices are designed to per-
form both functions.

The above discussion refers to code sets used for
character representation and should not be confused
with coding of computer programs.

DATA STRUCTURES

The data used during the execution of a program
are contained in a data structure known as a file. Data
must be organized in such a way that they can be iden-
tified and located during the execution of a program.
The smallest data structure is a character (byte). A
group of characters is a data element; a group of data
elements is a record; and a group of records is referred

732

Data Management

Stream

River

River

Segment

Mean

Rip

arian

Riparian

Name

Mile (St

art)

Mile (End)

Lenqth (mi )

m

dth

(ft)

Acreaoe

North Creek

7.2

9.4

2.2

20

11.1

South Creek

0.0

5.0

5.0

35

42.4

East Creek

4.1

4.3

0.2

40

1.9

West Creek

3.7

4.7

1.0

25

6.1

TOTALS

8.4

61.5

Figure 1. Sample output document.

to as a file. For example, a person's name, consisting
of a specific number of characters, could represent
a data element. When several data elements containing
information about the person are collected, a record
is formed. The data elements in the record could con-
tain the State codes and species codes. Several records
with different data create a file. A file may be located on
any media such as cards, tape, or disk.

Each field within a record contains the same num-
ber of characters as the corresponding field on another
record. Because a space is a valid character, some fields
may contain more spaces than others. For example, if
a field contains 10 characters, the data "ABCD" would
use 4 characters and 6 spaces to complete the field.

MAIN STORAGE CAPACITY

A computer's main storage capacity determines
the amount of data and instructions that can be held
in the system at any one time. The more main storage
available, the faster storage and retrieval of data can
occur. Although it would be desirable to have as much
main storage as possible, it is expensive, and cost is
often a limiting factor.

Computer storage is measured in Ks; one K repre-
sents 1,024 bytes (characters). Computer main storage
sizes range from as little as 4K on a microcomputer
to as large as 16.000K (16 million bytes) on a large
computer. One million bytes is called a megabyte.

APPLICATION PROGRAMS

Application programs perform specific user
functions such as payroll, wildlife inventory, or material
sales. Many programming languages are available. Three
of the most commonly used are COBOL (Common
Business Oriented Language), FORTRAN (Formula
Translator), and BASIC (Beginner's All-Purpose
Symbolic Instruction Code). These languages make
essential and difficult computer programming tasks

The sequence of instructions that make up the
program is read into the computer's main storage. The
instructions are processed in sequential order and the
results produced. Once the program has begun to pro-
cess, it will continue, without any human intervention,
until completed. The instructions will be processed
exactly as the program indicated.

PACKAGED PROGRAMS

An application program may be a generalized, pre
written program that performs a function requested
by the user. This type of program, known as a pack-
aged program, is increasingly available to the user.
Many such programs are available for business and
scientific functions. They usually contain alternatives
that users may select for their particular situation.
Sometimes these programs require modifications to
meet the users' specific requirements. The cost for a
packaged program is often less than having a specific
program written. Consideration must be given to the
computer system the program will be executed on.

Data Management

733

Identify Data Sources

Develop a Collection Process

An analysis of the data elements indicates that
data elements 1, 2, 3, and 5 are read in from a file.
Data element 4 is calculated by subtracting data
element 2 from data element 3- Data element 6 is
calculated by converting data element 4 to feet, cal-
culating the area, and converting the results to acres.
Data element 7 is an accumulation of the indicated
data elements.

The source of the data to be read in must be
identified. The data may be in a form ready for input
to the computer, such as a file located on disk or
magnetic tape. The data may be located on source
documents, such as field observation forms, but the
data must be converted for input to the computer. If
the data are not readily available, arrangements must
be made to collect them.

A method for collecting data must be estab-
lished. A special form may need to be designed for
this purpose. Careful consideration should be given
to the design of data entry forms.

Define Where Data are to be Processed

When data are to be processed manually, ar-
rangements must be made as to where they will be
processed and who will do the processing. If a com-
puter is to process the data, arrangements must be
made with the computer facility for processing. Be-
cause of the various sizes and capabilities of com-
puters, it is important to know that the computer to
be used is capable of processing the job. Where the
computer is located will determine the method used
in making the input data available.

READERS

Readers perform the input function in the data
processing cycle. They read data into the computer's
main storage where it will be available for processing.
The devices discussed here are only capable of reading
data. Punched cards were once the most important
medium for data entry. In some locations they are still
used despite their disadvantages of slow input, suscepti-
bility to operator error, and machine malfunctions.
Because of their disadvantages, cards are being replaced
as an input media by direct entry to an electronic stor-
age medium.

Another type of reader is the paper-tape reader.
This device senses data as holes punched in paper tape.
Another reader, the magnetic-ink character reader,
translates magnetic-ink characters into electronic sig-
nals. It is widely used by the banking industry. Optical-
character-recognition devices are often used to scan
or read special marks, bars, letters, numbers, and char-
acters from special printed documents which are then
converted to electronic signals for computer process-
ing.

To transfer graphics, drawings or images such as a
map to the computer, a digitizer is used. A cursor
or some other reading mechanism that can be passed
over the surface converts the image to digital data.
These data can then be displayed on a cathode ray
tube (CRT) screen or processed by a computer system.

INPUT CONSIDERATIONS

Input data arc often preprocessed or edited to
ensure their validity. Because a process for updating ac-

counts receivable will have different edit requirements
than a wildlife inventory update, there will be different
edit criteria for each. However, the following are gen-
eral considerations that are usually included in an edit:

( 1 ) Tests to ensure numeric data are numbers and
alpha data are characters.

(2) Tests for reasonableness of the data. For exam-
ple, the months in a year should not exceed 1 2
or be less than 0.

( 3 ) Range tests to see if data, such as acreage and
herd numbers, fall within a range of acceptable
values.

(4) Checklists of acceptable values to see if data are
valid, such as correct species codes.

(5) Tests for data consistency. The data entered
may be valid by themselves, but invalid in rela-
tion to other data entered. For example, when a
wildlife inventory record has an entry for a
hypothetical occurrence and a verified occur-
rence, only one may be entered. An entry for
both, or blanks for both, indicates error.

(6) Checks for transcription and transposition er-
rors. These errors usually occur when the data
are being key-entered.

The instructions performed during edit processing
are included in a computer program. When input data
are processed, each record is checked using the in-
struction in the program. Any invalid data are identified
so that they may be corrected. The process may be
repeated as many times as necessary to obtain valid
data.

754

Data Management

Define Programming Requirements

Each processing step must be performed specifi-
cally and in the correct sequence. Because a com-
puter can only perform calculations, comparisons,
and data movements, all steps must consist of these
types of instructions. The computer will process
these instructions in sequential order. For example,
if data are to be calculated, they must be read in
before the calculation instruction is processed.

Programming is not as simple as it first appears,
since every instruction has to be given to the com-
puter. However, once these instructions have been
prepared by a programmer and stored on storage
media, the user needs only to issue a few commands
and the instructions will be executed. Programs to
do these calculations may already be available in
some agencies or elsewhere. Furthermore, many soft-
ware packages for common tasks are available, par-

ticularly for microcomputers. These include
spreadsheet and data base management software.
Such software requires simple commands to set up
and perform tasks, but does not require knowledge
of a programming language. Many biologists use such
software without assistance from a computer pro-
grammer or specialist.

Define Data Entry Procedures

The data to be processed must be prepared for
input. Data may need to be transferred to a data
entry form if it is not on a document that can be
directly used for data entry. If a form is required, the
person filling out the form and the method of enter-
ing the data to be used should be identified.

Define Input Procedures

Data must be converted to a form usable by the

METHODS OF PROCESSING

One of several tasks performed by the computer
when processing data is to interpret instructions. An-
other is to set up a connection to the input or output
devices indicated in the instructions. The total tasks
required to process a given set of data are known as a
job.

Two basic processing methods are used by a com-
puter system: batch processing and interactive pro-
cessing. In batch processing, data are collected into
groups (batches) and stored before being processed.
Batch jobs are scheduled and prioritized by the com-
puter operators or the computer's operating system,
depending on the installation. They may be deferred to
complete the processing when computer workloads
are lower.

In interactive processing, data are entered directly
into the computer and processed immediately. The
user interacts with the program while it is processing,
providing input and receiving output on a terminal.

OUTPUT CONSIDERATIONS

A printed report is one of the main types of out-
put. It may be produced on various output devices and
media. Printers can produce reports on a variety of
forms or on microfiche. There are three major cate-
gories of reports: detail reports, summary reports,
and exception reports.

Detail reports imply that each input record is read
to determine if it will be printed on the report. A sum-

mary report provides a summary of the input data. It
is usually used by management for making decisions.

Exception reports contain information determined
to be an "exception" or other than normal. They are
produced when specific information is needed. Their
advantage is that they save time and money. Large re-
ports that required many hours to produce and are
printed on many pages do not have to be produced
when the required information can be produced in a
few minutes on a few sheets of paper.

INQUIRY

Output from interactive processing is usually ob-
tained by inquiry. An inquiry is a request from the
terminal operator to a computer system for information.
This method provides the user with the most current
information in the shortest period of time. Often a
printed copy of the information on the terminal screen
can be obtained. Usually the screen will prompt the
user for entries and provide status information. Inquiry
systems are usually designed for a wide variety of users
and are easy to use. A well-designed inquiry system
will provide prompts to guide the user. It will acknowl-
edge all the user's entries and respond almost immedi-
ately. A wildlife biologist might use this type of system
to determine the number and type of species in a cer-
tain location. This information can be requested and
received from the computer system by the user without
the need for a specific program to be written. Often
programs have already been written and are available to
assist the user in manipulating or extracting information
from their data.

Data Management

735

computer. Various input methods are available to
accomplish this. The amount of data, the media the
data are on, and the organization of the data (fields,
records) are factors that should be considered. If
input is directly through a terminal, computer access
procedures must be established.

Identify Processing Requirements

The manipulations that must be performed to
convert the data to information must be identified.
All functions must be specified precisely and the
sequence in which they are to be performed must
be identified exactly. These are the instructions
that will be followed during the execution of the
program.

Define Output Procedures

The method of distributing the output and the

type of output must be identified. Output may be in
several forms. Due consideration should be given
to the types available. Future use of the output
should also be considered.

SYSTEM LIFE CYCLE

Just as computer technology has evolved from
vacuum tubes to semiconductor chips, so have meth-
ods for developing systems. The first computer appli-
cations, such as payroll, were relatively simple.
Problems in developing systems began to appear
when applications such as wildlife were put on com-
puters. These systems had a wide variety of data
and complex relationships. System failures and lack
of structure prompted managers to emphasize con-
trol of the development process. During the 1970s,
computer specialists increasingly employed the sys-

PRINTERS AND PLOTTERS

Printers convert electronic signals received from
the computer system's processor to letters and words
on paper (hard copy) that can be understood. Nearly
every computer system has a printer; many have sev-
eral. Because printing requirements vary from user
to user, many things must be considered when select-
ing and installing a printer. Some considerations are the
volume of printing, the number of multiple copies
required, the use of special forms, and the print quality
required.

One way to classify printers is by the manner in
which they print information. Serial printers print one
character at a time; line printers print one line of infor-
mation at a time; and page printers print a complete
page of information at a time. Most computer sheets
(11 x 14 in.) contain 66 horizontal lines printed 6 lines
per inch with 132 characters in each line.

Another way to classify printers is by the way
information is printed. When the printing elements are
pressed (impacted) against the paper, the method is
known as impact printing. Nonimpact printers use ther-
mal (heat), chemical, electrical, or optical techniques
to form images.

The way a printer forms characters is another
method of classification. A solid character is like the
characters used by a typewriter. Dot-matrix characters
are formed with a series of dots. Impact printers use
solid or dot-matrix characters. Nonimpact printers use
only dot-matrix techniques.

Plotters are output devices that convert numerical
information to graphic form. The output is produced
by the plotter through use of pens or pencils or by
electrostatic means. Many types of statistical informa-
tion can be graphically illustrated by using a plotter.

MAGNETIC TAPE

Magnetic tape units are used for both input and
output. They are capable of reading and writing on
magnetic tape. They can read the equivalent of 300,000
punch cards a minute.

The typical magnetic tape reel contains 2,400 ft of
0.5-in., oxide-coated mylar tape. Data are recorded on
the tape as magnetized spots in parallel channels or
tracks along the length of the tape. The density or num-
ber of characters that can be written on a tape varies
from 800 bytes per inch (bpi) to 6,250 bpi. The equiv-
alent of 1.5 million fully punched cards can be stored
on one reel of tape.

Magnetic tapes provide a compact, relatively inex-
pensive media for retaining data. They require protec-
tion from dust, humidity, and temperature fluctuation.
To assure that data stored on a tape remain readable,
they should be transferred (copied) to another tape
once a year. It is not always necessary to use a new
tape. The data on a tape can be erased, making it possi-
ble for the tape to be reused many times.

736

Data Management

tem life cycle as a methodology for system develop-
ment. It is still widely used today.

Although it has been developed for large-scale
application, the principles and processes are applica-
ble to any system. A wildlife biologist, setting up a
system for storing and retrieving data on annual
waterfowl, counts on the office microcomputer and
follows the same basic procedure.

The evolution of an automated objective, such
as wildlife inventory, from an initial idea to the re-
quired output, consists of several phases. These
phases (which may be referred to by different termi-
nology) consist of initiation, development, and
operation.

During the initiation phase, the general require-
ments for meeting the objective are established.

The feasibility and cost of automating are docu-
mented; necessary approvals are obtained and justifi-
cations prepared; alternative processing procedures
are identified; and requests for computing resources
are made.

The development phase consists of definition,
design, programming, and testing. During the devel-
opment phase, the user's requirements are defined.
From these requirements, specifications are pre-
pared, the system is designed, and programs are writ-
ten and tested, or appropriate software is located or
purchased and tested. Documentation is prepared
and user and operator guides provided as necessary.
Testing results are reviewed by the user to assure
the results are adequate. All deficiencies are identi-
fied and corrected. When the results produced
meet the requirements, the development phase is
complete.

MAGNETIC DISK AND DISKETTE

Magnetic disks and diskettes are often referred
to as hard disks and floppy disks. The distinction
refers to the flexibility of the material from which they
are made. In this section, we refer to hard disks as disks
and floppy disks as diskettes or flexible disks.

Magnetic disk units that read and write on mag-
netic disks are called disk drives. A magnetic disk is a
thin metal platter coated on both sides with a metal
oxide. Data are recorded as magnetized spots placed in
circles on the surface of the disk. These circular record-
ing positions are called tracks. Even though the tracks
in the center of a disk are smaller, they contain the
same amount of data as the outside tracks. The number
of tracks on one surface of a disk varies from less than
100 on microcomputers to over 800 on large main-
frames. The two recording surfaces of a disk have a
storage capacity ranging from 2 million to about 20
million characters, depending on the number of charac-
ters written on each track. When several disks are com-
bined on a common hub, they are known as a disk
pack.

There are three categories of disk drives: ( 1 ) fixed,
(2) removable, and (3) sealed. A fixed disk drive con-
tains non-removable disks. These units provide large
amounts of quickly accessible storage. A removable disk
drive requires a disk or a disk pack to be mounted on
the drive's vertical shaft before the data on the disk can
be accessed. The need to increase data reliability re-
sulted in the development of a sealed cartridge known
as a Winchester disk. The cartridge contains the disks
and other components used in reading and writing
on the disks.

Another type of disk drive reads and writes on
diskettes (floppy disks). A floppy disk is sealed in a
square, plastic jacket. The most common sizes are 8
inches, 5 1/4 inches, and 3 inches. Since the slits in the
jacket expose the recording surface, users should be
careful not to touch the exposed area; this could dam-
age the recording surface. Diskettes provide convenient,
low-cost data-recording media that are easy to handle
and can be reused.

Devices capable of storing large amounts of data
on media such as magnetic tapes and disks are known
as mass storage devices. Other types of mass storage
devices capable of performing both input and output
functions include magnetic drum and data cartridge.
These types of storage are used by large computer
systems because of their efficient intermediate storage
capabilities.

ACCESS METHODS

There are two methods of accessing data: sequen-
tial and direct. In sequential access, data are read and
written in a sequential order. To access data located
at the end, all data preceding it must be read. The ac-
cessing procedure always starts at the beginning of the
data.

In direct access, also called random access, data
may be written to or read from any location on the
medium. Data at the end of a project can be directly
accessed. That is, none of the preceding data would be
accessed. Devices capable of accessing data directly
are called direct-access storage devices.

Data Management

737

The operation phase includes maintenance, eval-
uation, and modifications, if required.

be more involved for a complex system than a sim-
ple one.

The limitations of the system life cycle have
sometimes resulted in a backlog of applications to be
placed on computers. Complex systems require a
year or more to reach full production if this proce-
dure is followed. Many organizations use structured
analysis and design techniques and very high-level
languages to reduce the time required to develop a
system. Using these techniques, a skeleton system
can be developed to the point of user involvement
within a short period of time. Only a few months
may then be required to bring the system to full
production. The use of these techniques will grow in
the future. The time involved to develop a system
depends on its complexity. Such considerations as
identifying data elements, defining codes to be used,
and designing procedures for inputting data may

SUMMARY

This chapter is not an in-depth discussion on
any of the subjects mentioned. Additional informa-
tion can be obtained from a number of good publica-
tions, such as those listed in the Literature Cited
section at the end of this chapter. Another good
source of information is through computer courses
offered at most colleges and universities. For infor-
mation on particular applications, many organizations
have individuals who provide training and assistance.

Computers are sophisticated tools that require
various amounts of training in their use. The infor-
mation they provide is a very valuable resource.
As with other resources, information must be man-

TERMINALS

Entering data directly into a computer's main
storage is possible from a terminal. Receiving output
(information) from a computer system is also possible
at a terminal. A terminal may be a single device, two or
three connected devices, or a computer system that
sometimes acts like a terminal.

One of the most commonly used terminal devices
is the cathode ray tube (CRT). It consists of a televi-
sion-like screen and a typewriter-like keyboard. Data are
entered through the keyboard by the terminal operator.
Output information is presented in display form on
the screen.

Another popular terminal device is the printer
terminal. This device creates hard copy output instead
of displaying the output information on a screen. It may
or may not have a keyboard for data entry. When it is
used only for receiving data, it is called a receive-only
(RO) terminal.

Terminals that do nothing more than pass data
back and forth are known as dumb terminals. A ter-
minal that can perform limited processing functions,
such as editing data before sending it to the computer,
is known as an intelligent terminal. In addition to a
keyboard and display screen (or printer), an intelligent
terminal has a processor, main storage, and input/output
interfaces. The processor circuitry may be a micropro-
cessor.

Data may be entered and sent to a location(s)
remote from the central computer site, if desired.
Through remote job entry (RJE), data to be proc-
essed is entered directly into the computer via a re-
mote terminal. Processing occurs on the central

computer and output is sent back to one or more re-
mote terminals.

DATA COMMUNICATIONS

The transmission of data from one location to
another is known as data communication. When the
transmission is over a long distance by means of tele-
phone or telegraph lines, microwave, or radio, it is
referred to as telecommunications. Teleprocessing,
on the other hand, is an activity that involves both
data processing and telecommunications.

A system in which data are collected from one or
more points of origin, transmitted to a central location,
and processed to produce results that are distributed
to one or more points of use is known as a teleprocess-
ing system.

The device used to convert electronic signals to a
form for transmission over telephone lines is a modula-
tor-demodulator, usually called a modem. It may be
a separate unit or part of a communication control unit.
Both the receiving end and the sending end of the line
require a modem to convert signals. A special type of
modem is known as an acoustic coupler, which uses a
standard office telephone and does not require special
wiring.

Data are transmitted over data communication
channels (also called data lines or data links). There are
several types of data communication channels, includ-
ing ( 1 ) standard telephone lines, ( 2 ) coaxial cable,
(3) microwave transmission, (4) satellite
communications, and (5) fiber optics. The costs of
these vary; standard telephone lines are the least
expensive.

738

Data Management

aged. Computers are one of the tools that can be
used to manage information and other resources
more effectively and efficiently.

Changes have taken place rapidly over the past
few years and that trend is expected to continue.
The real accomplishments of today may be only
commonplace tomorrow. We believe the future
holds great promise for the use of data processing by
biologists. Exact applications of this technology are
hard to predict, however. The use of hand-held data
terminals, graphics, video text, laser printing, and
fiber optics are just getting started. Many may be
commonplace in the future. Sophisticated communi-
cations technology will soon allow electronic mes-
saging and networking to be part of our lives. If

recent trends continue, hardware prices will de-
crease, speed of computations will increase, size will
decrease, and computers will become more "user
friendly."

Data processing for financial and administrative
activities is simple and widespread compared with
programming for wildlife and other natural resource
fields. It is often difficult to explain all of the re-
source interactions to the programmer who is devel-
oping a particular application. However, biologists
should not allow these complexities to intimidate
them. Remember, the computer is just a tool. It is a
powerful tool which can collect, store, and analyze
data to expand knowledge about the resources
we manage and make better wildlife management
decisions.

Another factor affecting cost is line speed. Data are
transmitted in bits per second (bps). A low-speed de-
vice transmits 40 to 300 bps. Although these devices
usually transmit over telegraph communication lines,
standard telephone lines used for voice communication
(called "voice-grade" lines) can be used (with special
modems) for full-duplex transmission at speeds up
to 2,400 bps. For speeds of 4,800 bps and above, spe-
cial communications are always required.

OPERATING SYSTEM

Each computer has an operating system that
consists of a collection of software to supervise its op-
erations. The operating system is usually provided by
the computer manufacturer. Usually changes cannot be
made to it by the local computer staff. When enhance-
ments are made by the manufacturer, a copy is pro-
vided to the user. Operating systems vary from
computer to computer even among those of one manu-
facturer. They are closely designed to the architecture
of the computer hardware.

Functions of the operating system include super-
vising the execution of programs, directing input/output
devices, and loading data into main storage. Most oper-
ating systems have programs that sort and merge data
and transfer data from one media to another.

Operating systems are often identified by acro-
nyms. For example, Disk Operating System (DOS) is
a commonly used system for microcomputers.

TIME SHARING

Time sharing is a method whereby a computer is
shared by several users for different purposes at the
same time. While in time sharing, a user may interact

with a program while it is processing, submit process-
ing to be performed later, or issue commands to deter-
mine the status of processing. Users appear to be served
simultaneously due to rapid electronic speeds, although
service is actually sequential.

DATA BASE

The term data base, in a generic sense, refers to
all data elements that will be created, used, or manipu-
lated by an automated data system. It does not imply
any particular organization or structure of the data, nor
that a data base management system (DBMS) will be
used. The definition of and the procedures for using a
data base may vary because of the different structures
and organizations.

The updating process for a data base provides for
changing the data for only one data element rather than
a data element in several files. The value of the data
element will be the same on the output from each
program.

A data dictionary or data element dictionary is a
tool that lists all data elements, their definitions, how
and where they are used, and who is responsible for
them. The dictionary helps standardize definitions of
each data element. It is up to user departments to en-
sure that the data they need are truly represented in an
organization's data base.

A major task for many organizations is deciding
the data bases they need, the best locations, the data
that should be stored in them, and the organization of
the data.

Data Management

739

LITERATURE CITED

BOHL, M. 1980. Information processing. Sci. Res. Associ-
ates, Inc., Chicago, IL. 492pp.

. 1981. An end-user's guide to data base. Prentice-
Hall, Inc., Englewood Cliffs, NJ. 144pp.

DE STEIGUER, J.W. and R.H. GILES, Jr. 1981. Introduction
to computerized land-information systems. J. Forestry
79:734-737.

LEOPOLD, A. 1948. Game management. Charles Scribner's
Sons. New York, NY. 481pp.

MAYER, K.E. 1984. A review of selected remote sensing
and computer technologies applied to wildlife habitat

inventories. California Fish and Game 70:102-112.
SHELLY, G.B. and T.J. CASHMAN. 1980. Introducton to

computers and data processing. Anaheim Publishing

Company, Fullerton, CA. 337pp.
SIEGERT, P.P. 1972. Systems and general management — a

rationale. Am. Manage. Assoc. 181pp.
STEENHOF, K. 1982. Use of an automated geographic

information system by the Snake River Birds of Prey

Research Project. Computer-Environment Urban Sys-
tem 7:245-251.
STRACKBEIN, R. and D.B. STRACKBEIN. 1983. Computers

and data processing simplified and self-taught. ARCO

Publishing, Inc., NY. 94pp.
WINKLER, C. 1983- The computer careers handbook.

ARCO Publishing, Inc., NY. 142pp.

740

Data Management

37

STATISTICAL
ANALYSIS

William H. West

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

"Quantification as such has no merit except insofar
as it helps to solve problems."

— P.B. Medawar, Advice to a Young Scientist

Editor's Note: A book on inventory and monitoring
of wildlife habitat would not be complete without
a discussion of statistics. Statistics are the basic
tool that biologists must use for analyzing quanti-
tative data Although many specialized techniques
have been developed for analyzing habitat data, in
many cases the traditional statistics are adequate.
One chapter cannot, of course, substitute for ade-
quate basic training in statistics. It does, however,
describe the use of traditional statistics as well
as some of the neiv techniques such as multivariate
analysis to evaluate habitat.

INTRODUCTION

Statistics should be a standard tool of the biolo-
gist working with inventory or monitoring. While
statistics are sometimes confusing and difficult to un-
derstand, they are procedures a resource manager
or biologist cannot afford to be without. Reliable
data are a must when managing natural resources in
today's society. However, we simply cannot afford to
measure every square inch of a resource! Statistical
inference allows the biologist to measure enough of
a site to make informed decisions based on data
representative of the area.

Statistics provide a way of describing wildlife
habitat as accurately as possible with the least
amount of cost. A biologist can describe a habitat
with a few factors and assign a level of confidence
about these factors. This provides the manager with
a good understanding of the problems faced and
allows a more knowledgeable decision to be made.
Finally, statistics provide a biologist with a means to
detect and test the results of management decisions.
Results that may be too small to be detected by vis-
ual inspection of a site or data can be displayed by
statistics, and variation from natural events, rather
than management actions, can be separated.

In all cases, statistics are merely a tool to be
used by a biologist or manager. Statistics do not
make decisions; they only provide managers with
information to make decisions. A manager or biolo-
gist who depends on statistics to make a decision
is abusing the use and intent of statistical analysis.
Only the manager can weigh all available data, deter-
mine the level of risk he or she is willing to accept,
and make a decision based on that information.

Statistics are tools, and there are special tools to
be used for special jobs. No one tool can be used
in all situations, and it is important to use the right
tool for the job at hand. A statistician can help select
and use the right tool and should be consulted be-
fore designing a study; after the study has started
is too late. Working with a statistician before data
are collected will help clarify questions, result in a
better study design, and save time and money.

Statistical Analysis

741

Use of the right statistical tool can help support
or refute the perceptions of a biologist. What a biol-
ogist sees, or thinks he or she sees, may not actually
be occurring. Natural variation can cloud the situa-
tion and lead a biologist to a wrong conclusion. If
the right statistical tool is used, confusion can be
reduced or at least accounted for.

(3) Association between two (or more) factors
the biologist has measured;

(4) The effect on one factor from changes to
other factors; and

(5) Whether a management action is having the
desired effect.

This chapter is not intended to teach statistics;
the reader is assumed to have some basic knowl-
edge. Rather, the chapter is intended to acquaint the
reader with inventory and monitoring situations
that require or are appropriate for statistical analysis
and to supply the reader with statistical tools appli-
cable to such situations. Basic statistical courses
teach the equations, but seldom teach techniques of
using them. The intent of this paper is to help teach
how to use the equations.

TYPES OF STATISTICS

Generally statistics fall into two categories: de-
scriptive and inferential. Descriptive statistics are
used to describe a population or sample. In statistics
a population is denned as the total items of interest
that might occur (Box et al. 1978). The population
is defined by a biologist before any testing is con-
ducted. How a population is defined depends on the
questions a biologist is interested in answering. If a
biologist is interested in answering questions about
the production of prairie falcons in Colorado, then
the population would be defined as the number of
falcon eyries in Colorado on a certain date. The
numbers describing the population are the parame-
ters of the population (e.g., the mean number of
falcon eyries).

A description of the population is usually based
on a sample from the population. The sample is al-
ways a subset of the population under study. The
125 falcon eyries surveyed during the summer of
1984 would be a sample of the population defined
above. To help meet the assumptions made when
using statistical analysis, the population sample is
usually taken randomly. A random sample is one in
which each member of the population has an equal
and independent chance (probability) of being se-
lected in any sample (Harnett 1982).

Inferential statistics, as the name implies, help a
biologist infer or conclude something about a popu-
lation, based on a sample. Inferential statistics allow
analysis of the following:

(1) True population parameters (e.g., mean and
standard deviation);

Natural variation in populations and differences in
samples taken make it difficult to determine differ-
ences or patterns just by looking at raw data. Infer-
ential statistics help remove or account for the
effects of natural and sampling differences, allowing a
biologist to look more at the population without all
the "noise."

STEPS TO USING STATISTICS

Use of statistics as a tool is most effective when
there is a clear understanding of what is to be ac-
complished. Examining a data set without any clearly
defined purpose to discover relationships or hypoth-
eses is known as "data snooping." Data snooping is
useful, but not the most efficient or effective means
of examining data.

Green ( 1979) outlines 10 principles for design-
ing a study. These principles are presented as a
guide; not all situations may warrant using all 10
steps. A biologist must choose the applicable princi-
ples and use the concepts as guideposts.

(1)
(2)

(3)

State the question concisely.

Take replicate samples within each combi-
nation of time, location, and any other con-
trolled variables.

(2) Population differences;

Take random samples. Putting samples in
representative or typical places is not ran-
dom sampling.

(4) Test whether a condition has an effect. Col-
lect samples where the condition is present
and where it is absent, but all else is the
same.

(5) Do preliminary sampling to provide a basis
for evaluating sampling design and statistical
analysis.

(6) Verify that the sampling device or method is
sampling the population thought to be sam-
pled.

(7) Break large areas into relatively homogene-
ous subareas and allocate samples to each in
proportion to the size of the subarea.

(8) Verify that sample unit size is appropriate to
the size, densities, and spatial distribution
of the organisms being sampled.

742

Statistical Analysis

(9) Test data to determine if the error variation
is homogeneous, normally distributed, and
independent of the mean.

(10) Having chosen the best statistical method to
test an hypothesis, stick with the result
(Green 1979).

Of these principles, the most important are 1
and 10, although the others should not be ignored.
Stating clearly the questions being asked will save
many problems during the course of the study.
Vague questions lead to vague answers, collection of
unnecessary data, wasted time and money, bad con-
clusions, and use of inappropriate statistical tests.
The importance of clear questions cannot be over-
stressed. A good question is the starting point or
foundation of all good studies, and can lead to a
good study design and proper selection and use of
test statistics.

If the questions are properly asked, a biologist
can avoid discarding these data because the results
were incorrect or unexpected. A conclusion may be
undesirable but the results will still be valid if the
question was stated properly and the correct test
chosen. Data and results should never be discarded
on the basis that the results and conclusions were
not what was wanted.

DESCRIPTIVE STATISTICS

Populations can be described in many different
ways. The intent of any description is to convey
an understanding of the population. A population (or
a sample) can be described to a manager by merely
listing all data in the population. For small data sets
(<10 values), this may be a satisfactory method
but listing becomes impractical when the number of
values is large. Giving a manager a long list of num-
bers does not convey much information. It would be
much more meaningful and easier to remember if
only one or two values were listed to describe the
population. The population mean and standard devia-
tion, or sample average and standard deviation, sup-
ply information in a more manageable way.

Measures of Central Tendency

Descriptive statistics are used to convey more
information in less space than listing all raw data.
The first statistic used to describe a data set is some
measure of the central tendency of these data. Three
measures of central tendency are mode, median,
and mean. The mode is a value in a data set (sam-
ple) that occurs most often; the median is a middle
value that arranges data in ascending order; and the
mean is a sum of all values divided by the total num-
ber of values (Harnett 1982).

EXAMPLE 1

A biologist has collected data on the number of
sagebrush plants per square meter (square yard)
on a single mesa top. A total of 20-meter-square plots
were sampled and these following data were collected:

3, 9, 1, 5, 9, 0, 4, 7, 9, 5, 5, 1, 1, 6, 3, 8, 4, 4, 7, 5

The data value that occurs most often (mode) is
5. When these data are arranged in ascending order,
the middle value of the data set (median) is also 5
The mean is 4.8 plants per m2.

With continuous rather than discrete data, the
mode is better calculated using groups or classes of
data rather than each individual data set. If raw data
are used, possibly only one value would occur for
each measurement. By grouping these data, a more
meaningful mode can be determined. However, the
overall usefulness of the calculated mode depends
on how the classes are constructed.

The mode and median are not used as often to
describe a data set as the mean or average. However,
there are situations when the mean of a data set
does not make sense, as when working with nominal
data (functional groups for vegetation classification,
i.e., shrubs, half shrubs, grasses, etc. ) or with ordinal
data in which subjects are ranked, i.e., suitability of
ecosystems for elk habitat (Krebs 1978). The mean
of a population or the average of a sample can be
thought of as the "point of balance of these data,
analogous to the center of gravity" (Harnett 1982),
and is the most commonly used statistical value.

Measures of Dispersion

The mean, mode, and median indicate the cen-
tral tendency of these data but not how these data
are grouped around the center. Measures of disper-
sion tell how these data are grouped or spread
around the center. The range of these data is a meas-
ure of dispersion, indicating the difference between
the high and low values. A common method of de-
scribing these data spreads is the standard deviation
or the variance of these data. Each data set can be
looked at for differences from the average. However,
this method of describing these data has the same
problem as discussed earlier; it conveys little mean-
ing and is difficult to remember. A single summary
value describing the spread of these data is much
more satisfactory. However, simply summing the
deviations conveys even less information than listing,
since the sum of the deviations will always equal
zero.

The average of all the squared deviations is the
variance of these data. The square root of the vari-
ance is the standard deviation of these data. The

Statistical Analysis

743

standard deviation can be thought of as the typical
deviation around the mean (Harnett 1982) and is
more convenient to work with than the variance
because it is in the same units as these original data.

EXAMPLE 2

Using these same data described in the previous
example, the deviations of each plot value from the
average of the entire sample (4.8 plants per m2) result
in the following:

-1.8, 4.2, -3.8, 0.2, 4.2, -4.8, -0.8, 2.2, 4.2, 0.2,
0.2, -3.8, -3.8, 1.2, -1.8, 3.2, -0.8, -0.8, 2.2, 0.2

The sum of these deviations is zero. The sample var-
iance is 7.9 and the standard deviation is 2.8.

Through repeated collection of continuous data,
measurements tend to cluster around the mean
measurement in a very predictable pattern. This pat-
tern is the normal or bell-shaped distribution (see
Figure 1 ). This distribution is described by two pa-
rameters: the mean (u.), describing the peak of the
distribution, and the standard deviation (cr), describ-
ing the spread of these data around the mean.

This distribution has useful properties for deal-
ing with continuous measurements. Approximately
68% of all data lies within plus or minus one stand-
ard deviation of the average, and 95% lies within
plus or minus two standard deviations of the aver-
age. In the above example, 68% of all data lies be-
tween the values 2.0 and 7.6. The standard deviation
and how it describes a sample will be important
when testing a hypothesis or making inferences
about the parent population.

fix)

0. 399/o

^ 0.3413

0.1359 /

•-la"*

*-lo "•

V 0.1359

M-2a M-o V P+o p+2o

Figure 1. Normal distribution with mean and
standard deviation.

Frequency Distributions

While the average and standard deviations are
the most common methods of describing a data set,
other descriptors are also useful in understanding
these data. By grouping these data into equal inter-
vals, dividing the number of observations in each
interval by the total number of observations, and
plotting the calculated frequencies, a frequency dis-
tribution can be derived. The shape of a frequency
distribution helps explain data characteristics. Attri-
butes of these data such as unimodal (one central
peak), bimodal (two central peaks), or multimodal
(many peaks) can be easily displayed with a fre-
quency distribution. The shape of the distribution is
very sensitive to the grouping of these data.

EXAMPLE 3

A biologist, working on the biological assump-
tion that the size of a sagebrush plant crown is
directly related to the age of the plant, collects aver-
age crown diameter (in centimeters) from 25 ran-
domly selected plants from a specific vegetation
type.

5, 5, 6, 7, 8, 8, 8, 9, 12, 12, 15, 16, 18, 18,
22, 24, 25, 25, 25, 26, 26, 27, 28, 29, 29

To attain the following frequency distribution,
group these data into 2-cm intervals and divide the
number in each class by the total number. The sum
of the frequencies for all classes will be 1.

CLASS

NUMBER FREQUENCY

4.50-6.49

(3)

0.12

6.50-8.49

(4)

0.16

8.50-10.49

(D

0.04

10.50-12.49

(2)

0.08

12.50-14.49

(0)

0.00

14.50-16.49

(2)

0.08

16.50-18.49

(2)

0.08

18.50-20.49

(0)

0.00

20.50-22.49

(D

0.04

22.50-24.49

(D

0.04

24.50-26.49

(5)

0.20

26.50-28.49

(2)

0.08

28.50-30.49

(2)

0.08

TOTAL:

(25)

1.00

The frequency distribution shows two peaks
(bimodal) in these data at 6.50 to 8.49 cm and
again at 24.50 to 26.49 cm. The mean was calcu-
lated at 17.32 cm and the standard deviation at
8.68 cm (Figure 2).

The biologist would describe these data as
having a mean diameter of 17.32 cm and a bimodal
distribution with peaks at 6.50 to 8.49 cm and 24.50
to 26.49 cm. Based on the biological assumption
made when these data were collected, the biologist
would conclude the population has old and young
plants, but few plants of medium age. This distribu-
tion may lead the biologist to ask why there was this
type of distribution in age.

744 Statistical Analysis

INFERENTIAL STATISTICS

Inferential statistics use information from a sam-
ple to infer something about the parent population
or to test a theory or hypothesis about the popula-
tion. The type of statistical test used will depend on
the assumptions (statistical assumptions, rather than
biological assumptions) made about the population.

Parametric statistics are used when the popula-
tion can be assumed to have a known distribution.
The normal or bell-shaped distribution is one that is
familiar to the average user. Since standard deviation
of the population is usually unknown, a t-distribution
based on the sample is substituted for the normal
distribution. The t-distribution is similar to the nor-
mal distribution (Huntsberger and Billingsley 1977);
the exact shape of the distribution depends on the
size of the sample taken. The larger the sample size,
the closer the t-distribution will approach a normal
distribution.

An assumption made about the sample is that
the sample items have a normal distribution and are
independent of each other. The selection of an item
(or subject) to sample will not affect the distribution
of other items in the sample (Box et al. 1978).

To ensure the assumptions are met, it is impor-
tant a random sample from the population be used.
Collecting data in representative or typical areas
does not constitute a random sample (Green 1979).
The largest sample that is economically feasible
should be collected so sample distribution more

FREQUENCY DISTRIBUTION

SAGEBRUSH CROWN DIAMETER

FREQUENCY (

5.5 7.5 9.5 11.5 13.5 15.5 17.5 19.5 21.5 23.5 25.5 27.5 29.5

CROWN DIAMETER (CENTIMETERS)

Figure 2. Frequency distribution of sagebrush
crown measurements.

closely approximates the population distribution. A
good discussion on the consequences of violating
these assumptions when using ecological data can be
found in Green ( 1979).

When no assumption can be made about the
distribution of the population, nonparametric statis-
tics are employed. Nonparametric statistics are used
to test the assumption of a normal distribution and
to conduct statistical tests without making assump-
tions about the form of the distribution. Collecting a
random sample has more influence on the results
of any statistical analysis than the assumption of a
population with an unknown distribution (Box et al.
1978). Nonparametric statistics will be discussed
after some of the common parametric statistics.

Confidence Intervals

The first parametric statistic deals with the infer-
ence of the true population mean, based on the sam-
ple average and standard deviation and desired level
of probability. This type of statistic is known as the
confidence interval and assumes normal distribution
of population. The confidence interval, in effect,
indicates the range of potential values of the popula-
tion mean, given the level of error a biologist is will-
ing to accept.

EXAMPLE 4

A biologist has defined a population as the
standing crop (g/m2) of a particular pasture during
the week of July 4, 1983. Sixty plots, each 1 m2,
were sampled, and a sample average of 327 g/m2
and sample standard deviation of 1 12 g/m2 were
calculated. Confidence intervals for the mean at an
80% confidence level and a 95% confidence level
were calculated.

80% C.L on the mean
95% C.L. on the mean

308 g/m2 to 345 g/m2
298 g/m2 to 355 g/m2

The biologist would conclude the population
mean lies somewhere within the ranges calculated
with the indicated probability.

The confidence level expresses the probability
that a parameter falls within the indicated limits but,
more importantly, the level of risk a biologist is will-
ing to accept in coming to a wrong conclusion.
When a biologist says he or she is 80% confident the
mean is within the limits, what is actually being said
is he or she is willing to accept a 20% chance of
being wrong. A 95% confidence means a 5% chance
of being wrong.

The smaller the acceptable risk of being wrong,
the larger the limits. This concept becomes impor-
tant when testing hypotheses using a high confi-
dence level. The higher the confidence level, the
more difficult to disprove or reject an hypothesis. It

Statistical Analysis

745

must also be remembered there is a trade-off be-
tween the size of a sample and the confidence level
desired. A statistic calculated from a small sample
size with a high confidence level may not tell much
about the population parameter being estimated.

EXAMPLE 5

The biologist decides to collect only four sam-
ples from the population (Example 4) and deter-
mines the sample average to be 327 g/m2 and a
sample standard deviation of 1 12 g/m . He or she is
interested in taking little risk when estimating the
population mean and desires a confidence level of
95% or 99%.

95% C.L on the mean = 148 g/m2 to 505 g/m2
99% C.L. on the mean = -0.10 g/m2 to 654 g/m2

The biologist has high confidence the popula-
tion mean is between the limits 0.0 g/m2 and 654
g/m2. However, because of the range of values,
nothing more is known than when started. The small
sample size coupled with the high confidence level
have made the numbers meaningless.

Sample Size

To ensure the statistics will be meaningful for
the desired confidence level and range of limits,
an adequate sample size should be calculated. Before
any sampling is done, a biologist must determine
the level of risk, the maximum range of acceptable
limits, and the funds available for sampling (Harnett
1982). With these conditions in mind, a biologist
can then collect a few samples and determine the
sample size needed to conform to the stated specifi-
cations, based on the preliminary samples. (See
Example 6.)

When choosing an adequate sample size, set
reasonable expectations for the results. Do not make
the range of limits too small! Knowing population
values with a very small range and high confidence
level may be desirable, but collecting these data may
be too expensive. In Example 6, if the biologist had
set a range of 10 g/m2 with a 99% confidence level,
a total of 5,300 samples would have to be collected.
Additionally, because the adequate sample is based
on a preliminary sample, the adequate sample size
should be recalculated periodically throughout the
sampling period to determine if any changes in the
adequate sample size have occurred from changes in
the sample mean or standard deviation.

There is a balance between meaningful results
and the cost of sampling. It is important to strive for
the most meaningful results within the constraints
of time and money. If the commitment of resources
is not going to be made to obtain meaningful results,
then it is better to not sample at all, or to reduce
the population size to fit within the monetary con-

straints. Do not reduce the limits of the results be-
low the level where they convey any information.
Subjective, qualitative information gathered by a site
visit, for example, may be better than quantitative
information derived from an inadequate data set. No
data are just as good as vague data and much less
costly. Example 5 is a situation where the cost of
data collection was minimal, but the information
gathered was useless.

Hypothesis Formulation

The next area of inferential statistics involves
testing of ideas or concepts. Essential to this process
is the formulation of hypotheses. Two hypotheses
are always formed about the ideas being tested. The
null hypothesis (H()) is assumed to be correct until
proven false by the test. The alternate hypothesis
(Ha) is assumed to be correct in the event the null
hypothesis is proven false. The null hypothesis and
alternate hypothesis are complementary.

Green ( 1979) cited some general rules for for-
mulating null hypotheses. First, the null hypothesis
should be the simplest one possible and yet describe
the concept being tested. If the hypothesis is com-
plex, the question asked is too broad. Second, the
null hypothesis must be falsifiable; the data collected
must be able to disprove the concept (no amount
of data can ever prove a concept). Finally, the null
hypothesis should have the fewest number of un-
known explanatory factors. If the concept tested has
many unknown factors affecting the result, conclu-
sions will be difficult to form.

EXAMPLE 6

Before sampling the population described in
Example 4, the biologist decides to determine an
adequate sample size before completing the sam-
ple. The decision has been made to accept a 10%
chance of error with a maximum range in the limits
of 50 g/m2. After collecting 10 samples, a sample
average of 327 g/m2 and sample standard deviation
of 112 g/m2 are calculated.

Solving the equation:

(2tsr

w

Where:

t = the t-variable for the sample at the

stated level of error,
s = the standard deviation of the sample,
w = the width of the desired confidence

interval.

Based on the preliminary sample, the biologist
determines that 67 samples must be collected to
meet the desired specifications. An additional 57
samples are then collected.

746

Statistical Analysis

Hypothesis formulation relates directly to the
questions a biologist is asking. If good questions are
asked, the formulation of hypotheses will be easier
than if poorly formed or vague questions are asked.
The formulation of hypotheses should be made when
the questions are asked. This helps determine what
data should be collected and what tests conducted
to falsify the null hypothesis and answer the
questions.

Included in the formulation of the null hypothe-
sis should be the probability level at which the null
hypothesis will be rejected in favor of the alternate
hypothesis. The alternate hypotheses should also
indicate if all the chance for error is to be placed in
one tail or both tails of the probability distribution
of the sample. The chance for error can be placed in
one tail if there is reason to believe there is no
chance of the error occurring in the opposite direc-
tion. When testing the hypothesis that standing crop
on a salt flat is zero, there is no need to test if the
standing crop is less than zero. Any chance of error
is placed on the side of standing crop being greater
than zero.

t-Test

The most common statistic employed to test a
hypothesis or compare the means of two samples
(populations) is the t-test. The test assumes the pop-
ulation has a normal distribution and the sample
has a t-distribution.

EXAMPLE 7

A biologist has identified two sagebrush flats on
either side of a small canyon (site A and site B).
Both sites are influenced by the same weather pat-
terns and are on the same soil formations. The biol-
ogist wants to know whether the sites are the same
and can be managed as such. The biologist also
wants to know if site B has enough standing crop
(at least 1,120 g/m2) to support the deer herd using
the site.

The two populations are defined as the total
standing crop (g/m2) at sites A and B on the first
week in August. The following specific questions
and hypothesis are formulated:

(1) Is the mean standing crop of population A the
same as population B?

H0: Mean A = Mean B
Ha: Mean A + Mean B

EXAMPLE 7 (concluded)

(2) Is the mean standing crop of population B at
least equal to the minimum standing crop
needed (1,120 g/m2) to support the deer herd
using the site?

H0: Mean B = 1,120 g/m2
Ha: Mean B = 1,120 g/m2

The biologist is willing to accept a 10% chance
of error of estimating the mean incorrectly, with a
range of 60 g/m2. For the first question, the chance
of error will be distributed between both tails of
the t-distribution. For the second question, the prob-
ability will be placed entirely in the lower portion of
the t-distribution since the biologist is only interested
in a value less than 1,120 g/m2.

Ten preliminary samples were collected at both
sites. For site A, the preliminary sample had an
average of 973 g/m2 and standard deviation of 68 g/
m2; from this information an adequate sample size
of 17 plots was calculated. For site B, with a sample
average of 1,010 g/m2 and standard deviation of
107 g/m2, an adequate sample size of 43 was deter-
mined. From the adequate sample sizes, the final
average of site A is 985 g/m2 with a standard devia-
tion of 70 g/m2; the final average for site B is 1 ,022
g/m2 with a standard deviation of 1 16 g/m2.

From these sample data, a t-test statistic of
-1.22 for Question 1 was calculated. Comparing this
value to a table value (1.671) for the t-distribution
of these sample data, the biologist found the
chance of error was not greater than the stated
rejection level. Since the biologist did not reject the
null hypothesis (H0), he or she concluded the two
populations were the same, or the two samples
were from the same population.

For Question 2, a t-test statistic of -5.5 was
calculated and compared to a table value of 1.303
for the t-distribution for these sample data. The
calculated t-test statistic was found to exceed the
permissible error level (10%) The null hypothesis
(H0) was then rejected and the alternate hypothesis
(Ha) accepted. The biologist concluded the standing
crop at site B was not sufficient to supply the mini-
mum needs of the deer herd at 1,120 g/m2.

Analysis of Variance

The t-test is useful when comparing two sam-
ples or when comparing a sample to some hypothe-
sized value. However, when more than two samples
(populations) are compared, an Analysis of Variance
(ANOVA) table is used. The ANOVA table accounts
for the variation of different factors of many popula-
tions simultaneously and is very useful when com-
paring the effectiveness of different management
treatments. The table can separate variation between
treatments and within treatments so a biologist can
more directly evaluate the response of the site to the

Statistical Analysis

747

management actions. The ANOVA table provides
information for the calculation of statistics, such as
the F-statistic, that can be used to compare popula-
tions or test hypotheses.

Construction of an ANOVA table is flexible and
can be built to account for environmental conditions
and the experimental design used. By careful experi-
mental design, a biologist can remove (account for)
some of the variability in these data and examine
only the factors of interest. A good experimental de-
sign is the key to using an ANOVA table, and a statis-
tician should be consulted at the very beginning of
the test to ensure the design is adequate. Major fac-
tors affecting the experimental design are the biolog-
ical and environmental conditions present. A
biologist must have a good understanding of these
factors when the statistician is consulted. Differences
in soils, climate, or species being studied might all
prompt changes in the design of the experiment and
must be considered prior to starting the study.

EXAMPLE 8

Two herbicide dealers want a biologist to use
their products in a proposed spraying of tamarix
plants to reduce plant cover. Both herbicides are
applied to the soil and are absorbed through the
plant roots. Before buying either product, the biolo-
gist wishes to determine which of the two sprays
is more effective.

A small hillside is used in a test of the two
sprays. The biologist notes a difference in the soil
type, depth, and available moisture at the bottom,
middle, and top of the slope; all other environmental
factors (to the best of the biologist's knowledge)
are the same. Because of these differences, the bi-
ologist feels these changes must be accounted
for in the experimental design. The design is, there-
fore, blocked (partitioned) based on the position of
each test plot on the slope. One-ha plots (selected
at random) at the top, middle, and bottom of the
slope are sprayed with Herbicide A; the other similar
plots are sprayed with Herbicide B. Control plots of
untreated tamarix are also located at each position
on the slope to compare with the treated plots

The population the biologist is interested in is
defined as the cm2 of ground covered by the aerial
portions of the tamarix plants per square meter of
ground area, 1 month after spraying has occurred.
The following specific questions are to be answered
by the study:

(1) Is there any difference in the effectiveness of
the three herbicides?

H0: Mean A = Mean B = Mean C
Ha: At least two means are not equal

Where: Mean A = the plots treated with
Herbicide A.

EXAMPLE 8 (continued)

Mean B = the plots treated with
Herbicide B.

Mean C = the untreated plots.

(2) Is there any difference in the cover of tamarix
at each position on the slope after treatment?

H0: Mean L = Mean M = Mean T
Ha: At least two means are not equal

Where: Mean L = the plots at the lower level of the
slope.

Mean M = the plots in the middle of the
slope.

Mean T = the plots near the top of the
slope.

The biologist determined a 10% chance of error
is acceptable and set a 90% confidence level for
the test. One month after spraying the following
results were obtained:

SLOPE

The following ANOVA table was constructed
from these data:

TREATMENT

A

B C

T

37

38 36

M

88

76 81

L

51

42 47

SOURCE OF

SUM OF

MEAN

F

VARIATION

DF*

SQUARES

SQUARE

STATISTIC

Block (position on

slope)

2

3,313.56

1,656.78

139 71

Treatment

(herbicide)

2

67 56

33 78

2.83

Residual

4

47.78

11.94

TOTAL:

8

3,428 90

N/A

N/A

"DF: Degrees of Freedom

Comparing the F-statistic for treatments calcu-
lated from these data (2.83) to the table value (4.32)
for the F-distribution, the null hypothesis (HC)) for
the first question cannot be rejected.

The biologist would thus conclude that neither
treatment is effective in reducing the cover of tama-
rix plants, since treatment with Herbicide A or B
did not significantly reduce the tamarix cover below
that of the untreated plots. The ANOVA table does
not tell the biologist why the herbicides are ineffec-
tive, only that there is no difference in the tamarix
cover of the plots. If the biologist had found signifi-
cant differences in the treatments, he or she could
then go on to test and examine which treatments
were different.

748

Statistical Analysis

EXAMPLE 8 (concluded)

For the blocks (position on slope) the calculated
value (139.71) exceeds the table value (4.32) and
the null hypothesis (H0) is rejected. Thus, the biolo-
gist would conclude that there are significant differ-
ences in cover of tamarix at different positions on
the slope. The conclusion would also be reached
that the decision to block the experiment on the
basis of the position of the plot on the slope was
appropriate, as significant differences were found in
the tamarix cover on the bottom, middle, and top
of the slope.

In addition to determining differences between
treatments, the ANOVA table can also evaluate inter-
actions of treatments when applying two or more
treatments to plots in combination. This use of the
ANOVA table enables a biologist to examine each
kind of treatment individually and in combination
to determine differences and interactions of the
treatments.

If in Example 8 the biologist had also been in-
terested in the effect the type of application (aerial
versus hand application ) had on the effectiveness of
the herbicides, the experimental design could be
altered to examine these interactions. The experi-
mental design would be the same as described ex-
cept each herbicide treatment would have been
applied manually and aerially. The biologist would
be able to answer the same questions as before,
but would also be able to answer the following
questions:

( 1 ) Is there a difference in the effectiveness of
the mode of application?

(2) Is there any interaction between the mode of
application and type of herbicide used?

Regression Analysis

Another situation often encountered is the need
to predict the changes in one factor as other factors
change. This type of situation is handled statistically
by the use of regression analysis, using either simple
linear regression (one dependent variable and one
independent variable ) or multiple linear regression
(one dependent variable and many independent
variables). Regression analysis uses associations be-
tween the dependent and independent variables
to construct an equation describing the dependent
variable based on the independent variables. From
this equation, the biologist can predict what the
dependent variable will be as the independent varia-
bles change. Another application of the equation is
to permit a biologist to predict a variable too expen-
sive or difficult to measure directly (dependent vari-
able) from variables less costly to measure
(independent variables).

Closely associated with regression analysis is
correlation analysis. While regression analysis de-
scribes how one factor will change as other factors
change, correlation analysis determines the degree of
association between the factors.

EXAMPLE 9

The biologist has noted an apparent association
between the density of sagebrush plants and the
density of sage grouse in a certain mountain basin.
Based on this association, the biologist wishes to
predict the number of sage grouse per square kilo-
meter from the number of sagebrush plants per
hectare, and has collected the following sets of
data:

SAGEBRUSH

5ET

PLANTS/ha

SAGE GROUSE/km2

1

200

21

2

600

29

3

500

30

4

100

12

5

700

32

6

700

27

7

100

17

8

500

32

9

900

37

10

200

18

11

300

25

Based on these data the biologist calculates the
regression equation:

Y = 0.0255 X + 14.32

r = 0.91

Where: Y = the number of sage grouse per
square kilometer.

X = the number of sagebrush plants
per hectare

r = the correlation coefficient of the two
variables.

The biologist then asks what the population
density of sage grouse will be if sagebrush density
is reduced to 300 plants per hectare. Replacing the
X in the regression equation, the biologist estimates
there will be 22 sage grouse per square kilometer at
a plant density of 300 plants per ha. The biologist
goes on to ask for the range of estimates in the
sage grouse density if he or she was willing to ac-
cept only a 10% chance of error, and constructs
a 90% confidence interval around the estimate.

90% C.I.: 22 sage grouse ± 6.37 sage grouse

The correlation coefficient given in the previous
example is a measure of the degree of association
between the independent and dependent variables. A
correlation coefficient (r) approaching 1 indicates a

Statistical Analysis

749

high positive association between the two variables;
as one variable increases, the other variable also
increases. A correlation coefficient of -1 indicates a
high negative association, where one variable in-
creases as the other decreases. The closer the r value
is to 1 or -1, the stronger the association; the closer
to zero, the weaker the association.

The correlation coefficient is only an indication
of how the two variables will change in association
with each other. It does not indicate a cause-and-
effect relationship although there may be one. In Ex-
ample 9, the high r value of 0.9 1 does not necessar-
ily mean the change in sagebrush density is causing a
change in sage grouse density.

Even though two variables have a high correla-
tion, one variable will not necessarily be a good
predictor of the other. The two variables may be
changing in association with each other, but the
amount of change is so small the predication equa-
tion does not provide any information. This situation
can be discovered by testing to see if the slope of
the regression line is significantly greater than zero.

The lack of significance in Example 10 may be
due more to the small sample size than a true lack of
slope in these data. A larger sample size could indi-
cate a slope significantly different from zero. How-
ever, this example is typical of the sample sizes
many biologists use or must use.

The example and descriptions of uses of simple
linear regression given here can be expanded to
more than one independent variable by the use of
multiple linear regression. Multiple regression can be
useful in choosing which independent variables of
many can be useful in predication, by testing the
coefficients of each variable to see whether they are
greater than zero (as done in Example 10).

Nonparametric Statistics

Many of the statistics discussed so far have been
based on the assumption that the populations, from
which the samples come, have a normal or some
other distribution. This assumption may not always
be the case (as in Example 3) or a biologist may not
be willing to make the assumption. In such situa-
tions, a distribution-free or nonparametric test can
be used, analogous to the parametric tests already
presented. It is also possible to transform these data
so the assumptions for parametric tests can be made;
however, this is a topic that will not be discussed
here.

Mann-Whitney U Test. The Mann-Whitney U test is
used in situations appropriate for a t-test, but only

EXAMPLE 10

For the last 9 years, the biologist has been
measuring the ambient temperature on April 1 and
the peak standing crop (g/m2) on July 1 , in a certain
mountain valley, hoping the temperature may be
used to predict the standing crop.

YEAR TEMP(°C) STANDING CROP(g/m2

1

11

500

2

12

520

3

12

519

4

13

530

5

14

535

6

14

537

7

9

492

8

11

498

9

5

480

The following regression line is calculated from
these data:

Y = 6.71X + 437.05

r - 0.92

The high correlation coefficient of the equation
(0.92) suggests there is a strong positive associa-
tion between the two variables. The biologist wants
to test whether the temperature is a good predictor
of standing crop. To do this he or she asks the
question and formulates the hypothesis:

Is the data slope of the temperature line signifi-
cantly different from zero7

H0: B = 0
Ha: B + 0

Where: B = the slope of the regression line.

The biologist determines that only a 2% chance
of error is acceptable. A t-test statistic is calculated
for the slope of the line at 2.44. It is then compared
to a table value of the t-distribution of these data
of 2.998. The test statistic is not beyond the level of
error set for the test; therefore, the null hypothesis
cannot be rejected. The biologist would conclude
the slope of the regression line is not significantly
different from zero, and the ambient temperature on
April 1 is not a good predictor of standing crop on
July 1, despite the high correlation between the two
variables.

where no assumption of a normal distribution is
made and samples are from independent populations
(Harnett 1982). (See Example 11.)

Contingency Table. Another nonparametric proce-
dure is the use of the Chi-Square statistic to test the
independence of two or more classifications made
on the same site or subject by using a contingency
table. Data used in this test are the frequency of
classification combinations given to the study sub-
ject. The question asked is "Is one classification of a

750

Statistical Analysis

site independent of another classification of the same
site or does one classification tend to occur consis-
tently with another classification?"

portance as a tool of the ecologist. The intent of this
section is to merely introduce the reader to these
procedures and to describe general areas where
these techniques could be applied.

EXAMPLE 1 1

A biologist has encountered a similar situation
as described in Example 7, but is unwilling or un-
able to make the assumption that the population has
a normal distribution. Again, the biologist wants to
know if the two samples are from the same popula-
tion and, thus, forms hypotheses to test the relation-
ship at a 5% chance of error.

H0: Population A = Population B
Ha: Population A =£ Population B

Ten random samples are located in each site
and standing crop data (g/m2) are collected from
each site:

SITE A

SITE B

900

1,065

1,201

843

857

912

1,013

1,104

923

1,292

951

1,312

1,173

742

1,021

961

914

1,093

777

776

A Mann-Whitney U test statistic is calculated
from these data at 46 and compared with a table
value of 78 (Snedecor and Cochran 1967). The
calculated value exceeds the acceptable error es-
tablished for the test (the calculated value is less
than the table value), and the null hypothesis is
rejected. The biologist would conclude the two sam-
ples are not from the same population.

MULTIVARIATE ANALYSIS

All the analyses dealt with so far have used only
one variable at a time, e.g., standing crop, crown
diameter, plants per unit area. This type of analysis is
known as univariate analysis. Unfortunately, most
habitats and habitat relationships cannot be ade-
quately described by only one variable, but require
the use of many variables. Descriptions using more
than one variable are the area of statistics known as
multivariate analysis. This area of statistical analysis is
receiving greater emphasis from the ecological com-
munity (Johnson 1981 ), despite the complexity of
calculation and difficulty of interpretation. However,
as the availability of computer programs expands and
as biologists become more familiar with these tech-
niques, this area of statistics will attain greater im-

EXAMPLE 12

A specialist has identified 22 separate mountain
valleys that can be grazed by either cattle or elk.
By an arbitrary system, each valley has been classi-
fied as good, fair, and poor elk range and as good,
fair, and poor cattle range. The specialist is inter-
ested in determining if the two classifications are in-
dependent of each other, or if good elk range also
tends to be good cattle range To answer the ques-
tion at the 5% level of rejection, the hypotheses
are formed:

HQ: The classifications are independent.
Ha: The classifications are dependent

Each valley classification is arranged in a con-
tingency table.

CATTLE Fa

Good Fair Poor

Where the values in the table are read as —

OBSERVED FREQUENCY/EXPECTED FREQUENCY

From these data, the calculated Chi-Square
statistic is 6.14, which compared with the table
value (9.49) is not beyond the level of rejection es-
tablished for the test. The null hypothesis cannot be
rejected, and the specialist would conclude the
two classifications are independent of each other.
Good elk range is not always good cattle range.

ELK

Good

- 2/2 64

3/1.2

1/2.16

Fair

- 4/4.84

2/2 2

5/3.96

Poor

- 5/3.52

0/1.6

3/2.88

Although not a true multivariate analysis, the
first technique that comes to mind is multiple regres-
sion. Touched on briefly before, this analysis permits
a biologist to predict a variable of interest (depend-
ent) from many other variables (independent). The
technique can be used to predict a variable too
costly to measure directly from less costly variables;
to predict changes in a variable based on changes
in other variables; or when using stepwise multiple
regression, to select those independent variables that
account for the most variability in the dependent
variable. A common use in habitat analysis is to pre-
dict animal abundance (dependent variable) from
a set of habitat attributes (independent variables).

Statistical Analysis

751

Discriminant Analysis

Discriminant function analysis is used to sepa-
rate observations into groups (Johnson 1981 ). When
a number of groups or sites have been defined on
the basis of certain ecological factors, discriminating
between the different sites is sometimes difficult
because of overlapping distributions of ecological
factors used to describe the sites. A biologist cannot
clearly assign a site to one group or another because
the site could have attributes of both groups.

Based on the variables used to describe the
original groups, discriminant analysis creates new
distributions (hybrid distributions of the original
variables ) that result in greater separation or dis-
crimination between the defined groups (Webster
and Burrough 1974). The hybrid distributions make
it easier to assign a site to one group or another
because of greater definition between the groups.
This is potentially one of the most useful multivar-
iate techniques, especially in the area of habitat
classification.

The discriminant analysis is a transformation of
the original factors for each site into a single value
(score) distribution for each site, with more distinc-
tion between the sites. The biologist calculates the
discriminant score for each new site, plots the new
score on the distributions of the defined sites, and
determines which site the new location more closely
approximates. (See Example 13.)

Principal Component Analysis

The principal component analysis is a multivar-
iate technique used to combine "measurements of
similar nature into a fewer number that may be
more stable" (Johnson 1981 ). The process selects a
few important variables out of many, by taking a
set of similar observations (measurements) and com-
bining them into a single variable. For example,
measurements of tree height, diameter, age, and
stand density may all be combined into a single vari-
able. The first new component constructed from
these original data will account for the largest possi-
ble variability of these original data, and each addi-
tional component constructed will account for the
largest possible amount of the remaining variability
left after construction of previous components.
(See Example 14.)

Principal component analysis is useful in identi-
fying which groups of measurements will account for
the greatest variability in these data. If several soil,
climate, and vegetation measurements are collected
for a site, principal component analysis will identify
which group (soil, climate, or vegetation) accounts
for the greatest variability in these data. Because
all these data are included in the analysis, some

combinations of measurements could be difficult to
interpret. Tree measurements and soil measurements
could be combined into a single principal compo-
nent making it difficult to interpret exactly what the
principal component is telling a biologist.

Canonical Correlation

Very similar to the principal component analysis
is the canonical correlation analysis. The difference
is that canonical correlation analysis uses two sepa-
rate sets of measurements and derives a linear com-
bination of each data set such that the correlation
between the two derived values (components) is

EXAMPLE 13

A biologist defines two habitat types on the
basis of these available data and on past experi-
ence. To define the types, the biologist relied on the
percentage of shrub cover, annual precipitation,
and the annual forage production. When examining
new locations, the biologist had difficulty placing
new locations into one of the two habitat types. One
factor (e.g., forage production) may place the loca-
tion in one type, whereas the other factors (e.g.,
percentage shrub cover and annual precipitation)
would place the same location in a different type. To
help differentiate the new locations into the two
habitat types, the biologist performed a discriminant
analysis on the two previously defined types and
obtained a discriminant equation:

D = aX +bY + cZ + C

Where: X = the standardized shrub cover.

a = the standardized discriminant coeffi-
cient for shrub cover.

Y = the standardized annual precipita-
tion.

b = the standardized discriminant coeffi-
cient for precipitation.

Z = the standardized annual forage pro-
duction.

c = the standardized discriminant coeffi-
cient for forage production.

C = the constant for the discriminant
equation.

D = the discriminant score.

A mean discriminant score was calculated for
each of the two known habitat types. As each new
site was encountered, a discriminant score was
calculated and compared to the means of the two
known groups (habitat sites). Each new site was
then classified into one of the two known habitat
types, based on which of the known discriminant
groups had the highest probability of containing the
discriminant value from the new site.

752

Statistical Analysis

EXAMPLE 14

In the previous example (Example 13), the biol-
ogist decided not to limit these data to the three
sets described and instead decided to use 19 cli-
matic variables (data from Newnham 1968):

NUMBER VARIABLE

Latitude

Elevation, feet

Average daily max. temp.

(°F),

Winter
Spring
Summer
Fall

Average daily mm. temp. (°F),

9

10

Winter
Spring
Summer
Fall

11
12
13
14

Average daily mean temp. (°F),

Winter
Spring
Summer
Fall

15

Average precipitation (inches)

, Winter

16

Spring

17

Summer

18

Fall

19

Average frost-free period,

days

After computat

on, three components

are de-

rived that account for 92.2% of the variation of these

data.

COMPONENT

1

2

3

VARIABLE

COEFFICIENTS

1

-0.63

-0.50

1.00

2

-0.85

0.10

-067

3

0.97

0.10

-0.41

4

0.02

1.00

-0.18

5

-0.11

0.99

-0.29

6

0.94

0.13

-0.45

7

1.00

-0.02

-0.10

8

0.90

0.35

0.50

9

0.84

0.31

0.75

10

099

-0.05

-0.01

11

1.00

0.02

-0.22

12

0.57

0.85

0.17

13

0.38

0.94

0.22

14

0.99

0.03

-0.19

15

0.72

-0.67

-0.23

16

0.59

-0.73

-0.26

17

0.40

-086

0.06

18

0.71

-070

-0.10

19

0.93

0.01

0.05

PERCENT OF
VARIATION

57.4

29.4

5.4

Once each coefficient was determined for each
variable, the biologist multiplied each by its coeffi-
cient and summed the products to arrive at the
value for each principal component. The three com-
ponents would then be used in the discriminant
analysis as described in Example 13.

maximized (Warwick 1975). A second canonical cor-
relation ( component ) is then constructed from the
remaining variability of these original data so the
highest correlation is achieved between the next
two components.

Whereas the principal component analysis de-
rives values from a single set of data to account for
the greatest variation in these data, the canonical
correlation analysis derives values from two sets of
measurements to attain the highest correlation be-
tween the two data sets. If one data set contained
vegetation measurements and a second contained
abiotic measurements, canonical correlation analysis
would identify which of the abiotic measurements
is most highly associated ( correlated ) with the vege-
tation measurements.

EXAMPLE 15

A biologist has collected data on relative hu-
midity (V1), precipitation (V2), wind speed (V3), and
climatic temperatures (V4). Data have also been
collected on vegetation cover (V5), frequency (V6),
production (V7), and standing crop (V8). The biolo-
gist is curious how the two data sets (V1 through V4
and V5 through V8) relate to each other.

The canonical analysis selects the V1 and V2
values from the first set, and V7 and V8 from the
second set to be used in the first canonical variant
(CANVAR1). Since the biologist is only interested
in the effects of climate on vegetation, only the coef-
ficients for these climate data are presented.

CANVAR1

0.527(V1) -0.463(V2)

The V3 and V4 values from the first set and the
V5 and V6 values from the second set are selected
for the second canonical variant (CANVAR2).

CANVAR2 = 0.542(V3) - 0.635(V4)

The first canonical variant has a correlation of
0.889, indicating that these data in the first canoni-
cal variant (V1, V2, V7, V8) share 79% (the square
of the correlation) of the variance of these data. The
second variant has a correlation of 0.824 and
shares 68% of the variance.

After conducting canonical correlation, the biol-
ogist found precipitation-humidity data and produc-
tion-standing crop data had the highest correlation,
followed by wind speed, temperature, and cover-
frequency data.

As with principal component analysis, a biologist
runs the risk of obtaining combinations of data that
are hard to interpret. This example was constructed
in such a way as to present the clearest interpreta-
tion; however, this type of clarity will seldom be
present in most ecological situations.

Statistical Analysis

753

TIME-SERIES ANALYSIS

Ecologists have long recognized that ecological
systems experience a cyclic behavior over time. The
use of time-series analysis allows a biologist to de-
scribe how a variable changes through time and
to predict how the variable will change in the future.
The time frame used with this analysis can be any
unit the biologist wishes: seconds, minutes, hours,
years, or other measures of time (Chatneld 1980). In
most statistical analysis, all measurements are as-
sumed independent; in time-series analysis, observa-
tions taken over time are often not independent
and an observation may be influenced by past
occurrences.

Time-series analysis has three major uses in the
field of ecology. First, it can be used to describe
the type of variation occurring (seasonal, cyclic,
trend, or irregular fluctuations). Second, the ob-
served variation in one time series can be used to
explain the observed variation in another time series
(e.g., the variation in forage production has the same
time series pattern as the variation in deer produc-
tivity). Third, time series can be used to predict a
variable at some future time.

EXAMPLE 16

A biologist has collected 20 years of data on
rabbit population densities in a certain mountain
valley. By plotting these raw data, the biologist can
see a cyclical pattern in the densities; however, due
to the "noise" in these data, the pattern is unclear.
After performing time-series analysis, the biologist
develops a logistic curve to describe these data:

Xt

A

;1 +B exp[-Ct])

Where: A, B, and C = parameters and t =
time.

Upon examining the equation and the plot of
the equation, the biologist can clearly see the cycles
in these data, an overall trend (up or down) to these
data, and whether the amplitude of the cycles is
increasing or decreasing.

The equation also allows the biologist to predict
the density of rabbits at some future time. By com-
paring the time series for the rabbits to time series
for factors affecting rabbit density (e.g., coyote
density), the biologist may also be able to explain
the pattern of the rabbit time series by the pattern of
the time series of other factors.

The use of time-series analysis holds great prom-
ise in the area of habitat management. However, it
is a complicated procedure that requires the help of
a statistician, and large data sets are needed to ade-
quately determine the pattern of fluctuations. As
with multivariate analysis, the advent of computer

programs will make the use of time-series analysis
more available to the biologist.

OPTIMAL ESTIMATION

Better techniques for estimating resource meas-
urements are beginning to appear in the natural re-
sources field. These techniques provide the biologist
with an estimation of the number of samples that
must be made and how often (e.g., every 5 years)
samples must be taken to maintain a predefined level
of variance in these data (Jameson 1985). These
techniques hold great promise for the efficient plan-
ning of monitoring studies over the long term and
can be applied either to univariate or multivariate
conditions.

The process utilizes Bayesian statistics to deter-
mine when the process variance becomes unaccepta-
ble, based on actual sample variation and estimated
change in the variance over time. The change in the
variance over time may be made by either assuming
a constant change or by simulating the changes
based on the conditions present, using a simulation
model.

As time passes, the variance of the process will
increase until the variance surpasses the maximum
tolerable level set by the biologist or manager (Fig-
ure 3). To return the variance to a tolerable level,
a sample must be taken to reduce the overall vari-
ance of the process.

o
z
<

<

>

ALLOWABLE VARIANCE
VARIANCE FOR n = 2
VARIANCE FOR n = 3

VARIANCE FOR n = 5

TIME PERIOD

Figure 3. Changes in process variance over time

for three levels of sampling (Jameson 1985).

754

Statistical Analysis

The biologist has the option of collecting a few
samples at short time intervals or collecting many
samples over longer time intervals. By applying cost
estimates for the sampling process, the biologist
can evaluate the sampling scheme (few samples —
short time frame; many samples — long time frame)
that is most economical.

CONCLUSION

Statistics are an important and powerful re-
source tool. The techniques can be used to describe
habitats, to test ideas and concepts, to evaluate man-
agement alternatives and decisions, and to predict

changes. While they are a powerful tool, statistics are
only a tool and are not capable of making decisions.
Statistics only aid the resource biologist in the deci-
sion process.

To be properly employed and not abused, statis-
tics must be considered from the very beginning of
any data collection or impact study effort. As part of
conceptualizing a project, statistics can help to clar-
ify the questions being asked, avoid the collection of
unnecessary data, and reduce the collection of mean-
ingless data. When used as intended, statistics can
provide the biologist with meaningful and useful
information.

Statistical Analysis

755

LITERATURE CITED

BOX, G.E.P., W.G. HUNTER, and J.S. HUNTER. 1978. Statis-
tics for experimenters — An introduction to design,
data analysis, and model building. John Wiley & Sons,
New York, NY. 653pp

CHATFIELD, C. 1980. The analysis of time series: An
introduction. Chapman and Hill, London. 268pp.

GREEN, R.H. 1979. Sampling design and statistical meth-
ods for environmental biologists. John Wiley & Sons,
New York, NY. 257pp.

HARNETT, D.L. 1982. Statistical methods, 3rd ed. Addison-
Wesley Publ. Co., Reading, MA. 730pp.

HUNTSBERGER, D.V. and P. BILLINGSLEY. 1977. Ele-
ments of statistical inference, 4th ed. Allyn and Bacon,
Inc., Boston, MA. 385pp.

JAMESON, D.A. 1985. Sampling intensity for monitoring of
environmental systems. Applied Mathematics and Com-
putation. Elsevier Sci. Publ. Co., Inc., New York, NY.

JOHNSON, D.H. 1981. The use and misuse of statistics

in wildlife habitat studies. Pages 11-19 in Capen, D.E.,
ed. The Use of Multivariate Statistics in Studies of
Wildlife Habitat. U.S. Dep. Agric, For. Serv. Gen. Tech.
Rep. RM-87. 249pp.

KREBS, C.J. 1978. Ecology — the experimental analysis of
distribution and abundance, 2nd ed. Harper & Row,
New York, NY. 678pp.

NEWNHAM, R.M. 1968. A classification of climate by prin-
cipal component analysis and its relationship to tree
species distribution. Forest Science 14:254-264.

SNEDECOR, G. and W.G. COCHRAN. 1967. Statistical
methods, 6th ed. Iowa State University Press, Ames.
593pp.

WARWICK, P.V. 1975. Canonical correlation analysis:

Subprogram CANCORR. Pages 515-527 in Nie, N.H.,
C.H. Hull, J.G. Jenkins, K. Steinbrenner, and D.H. Bent,
eds. Statistical Package for the Social Sciences (SPSS),
2nd ed. McGraw-Hill, New York, NY. 675pp.

WEBSTER, R. and P.A. BURROUGH. 1974. Multiple dis-
criminant analysis in soil survey. J. Soil Science
25:120-134.

756 Statistical Analysis

38

HABITAT

EVALUATION

SYSTEMS

Allen Y. Cooperrider

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

"What Is Game Range?' When the game manager
asks himself whether a given piece of land is suitable
for a given species of game, he must realize that he
is asking no simple question, but rather he is facing
one of the great enigmas of animate nature. An an-
swer good enough for practical purposes is usually
easy to get by the simple process of noting whether
the species is there and ready, or whether it occurs
as 'similar' range nearby. But let him not be cocksure
about what is 'similar,' for this involves the deeper
questions of why a species occurs in one place and
not in another, which is probably the same as why it
persists at all. No living man can answer that ques-
tion fully in even one single instance."

— Aldo Leopold, Game Management

Editor's Note: Habitat evaluation systems are the
tools for analyzing wildlife habitat, and the habi-
tat model is the basic building block for such sys-
tems. Therefore, use of such models and systems
is basic to the habitat inventory and monitoring
process. Biologists should be familiar with existing
techniques in this area Most of the major concept
and system developments have occurred in the past
15 years and the methodology is likely to continue
to develop rapidly. This chapter provides an intro-
duction to habitat evaluation systems and models,
their history, construction, and use in the inventory
and monitoring process, as well as an overview of
current models and systems.

INTRODUCTION

Unorganized data are of little value to the biolo-
gist or manager. Data need to be organized, inte-
grated, and summarized into useful information. The
biologist can then use such information to make
inferences about relative conditions of habitats and
predictions about future conditions under alternative
management practices. Statistical analysis, as dis-
cussed in the previous chapter, is one means of ana-
lyzing data.

In this chapter, I discuss methods specifically
designed for organizing and analyzing wildlife habitat
data. I do not discuss in detail models and systems
designed for analyzing a single species (e.g., a system
for analyzing mule deer habitat ) or a particular type
of habitat (e.g., a system for analyzing riparian habi-
tat ). Rather, I focus on models and systems that can
be used for many different species (e.g., habitat suit-
ability index [HSI] models) or many different types
of habitats or communities (the system to be used
determines the data to collect). The reverse process,
in which a biologist collects data and then searches
for a system that can use the data, is inefficient and
should be avoided.

The habitat model, a correlation between habi-
tat components and some attribute of animal popula-
tions), is central to any habitat evaluation system,
although it may not be labeled as such. Therefore,
much of this chapter focuses on habitat models
which are the basic building blocks of any such
system.

In this chapter, I —

( 1 ) Review the historic development of habitat
evaluation systems and models;

( 2 ) Discuss the elements and types of habitat
models;

( 3 ) Describe the procedure a biologist should
follow to develop or modify a model for use
in habitat evaluation;

Habitat Evaluation Systems

757

(4) Discuss the relationship between habitat eval-
uation systems and models and the habitat
inventory and monitoring cycle; and

(5) Review models and systems currently avail-
able or in use.

DEVELOPMENT OF HABITAT INVENTORY,
MONITORING, AND EVALUATION SYSTEMS

Inventory and monitoring of wildlife habitat has
been conducted in the U.S. for over 50 years, al-
though such efforts were rarely described by those
names. Aldo Leopold, in his classic work on Game
Management, published in 1933, described methods
for "game range evaluation" that are essentially habi-
tat inventory and monitoring techniques. Such evalu-
ations were mostly limited to practices such as
browse condition and trend surveys for white-tailed
deer (Odocoileus virgin ianus) winter ranges.

Although more and better techniques were de-
veloped, until about 1970 most habitat surveys con-
sisted of measuring a few habitat components for a
particular game species. Methods were not available
for either analyzing habitat for nongame species or
analyzing the capability of a habitat to support a
wildlife community (as opposed to a single species).

At the same time, zoologists, naturalists, and
other biologists were acquiring a large body of
knowledge on the distribution of vertebrates and the
habitats used by most vertebrates in North America.
Habitat observations were typically made inciden-
tally by biologists who were studying a particular
species or surveying a region for wildlife. Few biolo-
gists made any effort to systematically collect or
synthesize such information. Until about 1970, the
best summaries of such species-habitat relationships
were the popular field guides such as Peterson's
Field Guide to Western Birds.

These guides provided simple "verbal models"
of habitats used by individual species. For example,
the American dipper (Cinclus mexicanus) is charac-
terized as using habitat described as "Fast-flowing
streams in or near mountains. Lower levels in win-
ter." (Peterson 1969).

Beginning in the late 1960s, interest in habitat
evaluation increased dramatically in the U.S. The
public became ( 1 ) more aware of the value and im-
portance of all wildlife, not just the huntable or eco-
nomically important species, and (2) more
concerned about the impact of human actions on
wildlife habitat. This concern was eventually ex-
pressed in a series of federal legislative mandates
that were used by agencies to manage habitat for all

wildlife species and to publicly address and mitigate,
where possible, the impact of management actions
on wildlife resources.

This mandate required new systems and ap-
proaches to habitat inventory and monitoring. Wild-
life biologists were thus faced with new problems
and challenges.

The first major challenge for biologists was deal-
ing with all species or at least all vertebrate species.
Wildlife biologists were not trained or experienced
in working with most nongame species. On the
other hand, zoologists were not accustomed to col-
lecting the types of data that were relevant to man-
agement. Neither group was experienced in working
with an entire vertebrate community. Habitat re-
quirements for many species were not known, and
techniques for measuring habitats or populations of
many species groups were primitive. Costs for meas-
uring habitats and populations of all species were
prohibitive, yet systems for setting priorities for field
work did not exist.

The second major challenge was to make pre-
dictions about impacts from human activities. Biolo-
gists had always made such predictions, but they
tended to be informal and based on experience and
subjective interpretation of limited data. Legislation
such as the National Environmental Policy Act
(NEPA) required impact predictions to be based on
data that were systematically collected and analyzed,
and that such predictions be formally presented and
open to public scrutiny in environmental assessment
reports (EARs), environmental impact statements
(EISs), and other decision documents.

More recently the need for systems to valuate
wildlife resources has emerged. Economic analysis of
wildlife resources is still primitive, but economic
values can be quite important in influencing manage-
ment systems. Such economic analyses require input
on the wildlife population responses to management
actions.

At the same time new demands were being
made of biologists, new technology became avail-
able. Of particular importance were the new devel-
opments in remote sensing and computer
technology.

Remote sensing has been used by wildlife biolo-
gists for many years, primarily aerial photography
for cover mapping and radiotelemetry for tracking
animal movements. However, the development of
these technologies, particuarly in the past 1 5 years,
has allowed biologists to collect better data, more
efficiently than in the past.

758

Habitat Evaluation Systems

Newly developed computer hardware and soft-
ware have probably been the most important new
technologies for the biologist concerned with wild-
life habitat inventory and monitoring. Two factors of
computer technology have been particularly valu-
able: ( 1 ) the capability of efficiently storing and re-
trieving large amounts of data and (2) the ability
to rapidly manipulate numbers and images.

The storage capability of computers has allowed
large amounts of inventory and monitoring data to
be efficiently stored and retrieved. This has relieved
much of the burden of handling large volumes of
data and lessened the errors associated with tran-
scribing data by hand.

The data manipulation capability of computers
has allowed not only rapid generation of summary
statistics, but also more sophisticated approaches
such as multivariate analyses. Such analyses would
not be practical without the aid of modern com-
puters. Similarly, many of the models developed for
integrating data would not be practical without both
the storage and data manipulation capabilities.

HABITAT MODELS

The habitat model forms the basis for all habitat
inventory, management, and monitoring. It is thus
the general underlying principle of habitat
management.

A habitat model is a method of using a set of
habitat components or attributes to predict some at-
tribute of a wildlife population or populations (Fig-
ure 1 ). All habitat models are designed for this
purpose. Habitat models are complex because of the
almost limitless habitat components and attributes
that can be used (Table 1 ), the diversity of popula-
tion attributes that can be predicted (Table 2), and
the complicated relationships between them.

define "fast-flowing" and "near" in terms of meters
(feet)/seconds and kilometers (miles) respectively,
then we would have a simple, quantitative model.

To understand and categorize a habitat model,
one needs only to identify the habitat components
being used as predictors, the population attribute
being predicted, and the type of function being used
to relate one to the other. In the following discus-
sion I describe each of these.

Model Components — Habitat Attributes

The habitat attributes listed in Table 1 form the
basic predictors of terrestrial habitat models. They
are analogous to the independent variables in a
regression equation. They are discussed in more
detail below to help explain the concept of a habitat
model.

Geographic Location. The single most important
predictor of occurrence is probably geographic loca-
tion. Most wildlife species are quite restricted in
geographic distribution; therefore, geographic loca-
tion, together with knowledge of a species distribu-
tion, is adequate to predict species potentially
present in an area. However, wildlife species are
rarely present continuously within their geographic
ranges, and complete delineations of all sites used by
a species are usually not available. Furthermore, loca-
tion is of little help in predicting more sophisticated
attributes of a wildlife population, such as relative
abundance or density. Therefore, geographic location
in habitat models is usually used implicitly in models
as a boundary condition, i.e., animals outside their
known geographic range are assumed to be absent.
More accurate prediction of presence and more de-
tailed predictions about population attributes ob-
viously require more detailed information on habitat
components present.

Habitat models have always been used by biolo-
gists; the previous example of the American dipper
using habitat consisting of "fast-flowing streams in or
near mountains" is a simple, verbal model. If we

Habitat Components or
Attributes

- Cover Type

- Forage Supply

- Other

Predictive

Equation

Population
Characteristics

- Presence

- Abundance

- Density

- Other

Vegetation. Beyond geographic location, vegetation
is probably the next most important habitat compon-
ent. Almost all habitat models use vegetation in pre-
dicting animal population attributes. In the simplest
models, the presence and absence of animal species
are simply associated with a vegetation type such
as pinon-juniper (Pinus-Juniperus sp.) or sagebrush
(Artemisia sp. ) grassland. More complex models
consider structural components of the vegetation
(e.g., tree density, shrub density, etc.) or more detail
on plant species composition (e.g., percentage cover
of big sagebrush [A tridentata], density of antelope
bitterbrush [Purshia tridentata}).

Figure 1. Basic components of a habitat model.

The relative importance of vegetation structure
(physiognomy), as opposed to plant species compo-
sition (floristics) in determining suitability of habitat

Habitat Evaluation Systems

759

Table 1. Habitat components and attributes useful in predicting presence, abundance, or density of
vertebrates.*

Geographic location

Vegetation — Live

Vegetation type

Species composition (floristics)

Presence

Abundance

Density

Cover

Frequency

Biomass

Diversity
Vegetation structure (physiognomy)

Presence

Abundance

Density

Cover

Frequency

Biomass

Diversity

Vegetation — Dead

Litter or mulch

Dead and down woody material

Persistent

Non-persistent

Snags

Physical features

Landform types

Alluvial fan

Rock pediment

Other
Landform attributes

Slope

Aspect

Elevation
Soils

Edaphic habitats
Geomorphic habitat features

Cliffs

Caves

Talus

Lava flows

Sand dunes

Other

Animal-made habitat features

Beaver dams

Dens

Nest cavities

Man-made features

Roads
Bridges
Buildings
Other

Water

Presence
Attributes
Depth

Flow (velocity)
Temperature
Chemistry (DO,
Substrate

pH, turbidity, TDS, etc.)

Food supply

Vegetation

Animal prey base (vertebrates)

Macroinvertebrates

Other
Presence, absence, or abundance of competitors
Presence, absence, or abundance of predators
Presence, absence, or abundance of parasites or

diseases
Presence, absence, or degree of human disturb-
ance

Noise

Human activity

Traffic

Other
Presence, absence, or intensity of hunting or
harvesting

Weather and climate

Historical occurrence

The spatial and temporal arrangements of tne attributes and components are also an important determinant of animal abundance

for wildlife species, is a subject of continuing debate.
Biologists, such as Short (1983, 1984) and Short
and Burnham ( 1 982 ), have proposed using habitat
models primarily based on structural characteristics,
whereas some researchers provide evidence that
species composition is more important. The manage-
ment biologist does not need to be concerned with
the controversy. Some wildlife species greatly de-
pend on particular plant species (e.g., sage grouse
[Centrocercus wophasitmns] on sagebrush, ruffed
grouse [Bonasa umbel I us] on quaking aspen [Popu-
lus tremulokles]). The biologist working with such

species is well-advised to pay attention to species
composition. On the other hand, habitat layers or
other structural components, as proposed by Short
( 1984), can provide a good initial prediction of ani-
mal species presence or absence. Chapter 3 1 in this
book describes how both types of measurements can
be taken and used to make predictions about animal
populations.

Vegetation has been measured in many ways for
many purposes, resulting in literally thousands of

760

Habitat Evaluation Systems

measurement techniques being described in the
literature. For this reason, the chapter on vegetation
(Chapter 31 ) does not attempt to cover all possible
techniques. Rather, it presents an example of how
vegetation variables can be used to predict species
occurrence and abundance.

Wildlife biologists often collect large amounts of
vegetation data without a clear idea of how they are
going to use such data. Habitat models are the mech-
anism for using vegetation data to make predictions
about animal populations. The biologist should,
therefore, have a model in mind before collecting
vegetation data. The same caveat applies to all other
types of habitat variables, but is emphasized here
since so much time, effort, and money continue to
be expended in collecting vegetation data that are
not used.

Table 2. Animal population characteristics
predicted by habitat relationships models.

Single Species Characteristics

Presence or absence

Probability of occurrence

Abundance or density

Relative abundance or density index

Carrying-capacity index

Biomass

Geographic distribution

Mortality or natality

Animal condition

Population dynamics

Multiple-Species Characteristics

Species richness

Species diversity

Biomass

Guilds or life-forms

Community suitability index

Dead Vegetation. Various types of dead vegetation,
such as snags, litter, and downed logs, are becoming
increasingly more important to wildlife biologists.
As a result, measures of these components are being
used increasingly in habitat models. However, few
formal systems for quantifying such components are
available, and in many cases the biologist must de-
vise ad hoc measurement systems. These habitat
variables are described in more detail in Chapter 27,
Terrestrial Physical Features.

Physical Features. Physical features are important
habitat components in both aquatic and terrestrial
systems. Because of their importance, two chapters
in this book have been devoted to describing sys-
tems for measuring them (Chapter 27, Terrestrial
Physical Features, and Chapter 28, Aquatic Physical
Features). In addition, because of its central import-

ance, an entire chapter has been devoted to the sub-
ject of soils (Chapter 26).

Water. Many terrestrial wildlife species require free
water. Therefore, presence of free water is an impor-
tant habitat component for these species. This may
be expressed as a distance from free water, as a den-
sity of springs, or as many other factors. The proper-
ties of the water are generally not important. With
amphibious and aquatic species, however, properties
of water such as turbidity, temperature, and pH be-
come important. Therefore, measurements of theses
properties are described in a separate chapter (Chap-
ter 30, Water Quality ). In addition, two other chap-
ters deal with water-related variables. Chapter 29,
Hydrological Properties, deals with dynamic water
variables such as flow rates that are important for
many wildlife and fish species; Chapter 32, Macroin-
vertebrates, focuses on using measurements of ma-
croinvertebrates as indicators of water quality and
other habitat deterioration.

Food Supply. Food is such an important factor
determining presence, absence, or abundance of
wildlife species that it is usually incorporated into
habitat models, either implicitly or explicitly. When
animal species are simply associated with a vegeta-
tion or cover type, or with a habitat layer within
a vegetation type, the model assumes the vegetation
type provides an adequate food supply. At the other
extreme, food supply may be so important to some
species that actual measurements or estimates of the
food supply must be made. Schroeder ( 1984), for
example, describes a winter model for black brant
(Branta bernicla) that uses two habitat variables,
one of which is the percentage of cover of the for-
age eelgrass (Zostera marina). In the case of preda-
tory animals, the density or abundance of prey
species may be the important habitat variable. Use of
food supply as a habitat variable is probably best
developed for large herbivores such as elk (Cervus
elaphus) and deer. With these animals, some habitat
models (frequently called carrying-capacity models)
predict the number of animals that can be supported
on a range by using food supply as a factor. Some of
these models not only require data on weight of
individual plant species present by season, but also
data on the nutritional content of various plant spe-
cies or species groups. Such models are discussed in
more detail in Chapter 25, Ungulates.

Presence, Absence, or Abundance of Com-
petitors. The presence of competitors is rarely con-
sidered formally in habitat models. Yet, the
phenomena of competition is recognized by many
biologists as an important factor affecting distri-
bution or abundance of animal species. There is, of
course, much disagreement among biologists about
the importance of competition and the mechanism
by which competition occurs. Similarly, many biolo-

Habitat Evaluation Systems

761

gists do not consider presence of a competitor as a
"habitat component." However, in the context of
habitat models, predators may be treated as habitat
components just like any other. For example, many
species of cavity-nesting songbirds are adversely
affected by the presence of starlings (Sturnus vul-
garis). Removal of starlings can improve the habitat
for many of these species and allow them to become
more abundant. Therefore, habitat models for such
species must consider the influence of the presence
or abundance of starlings in the habitat.

Presence, Absence, or Abundance of Predators.

Like competitors, predators are not usually consid-
ered in formal quantitative habitat models, but are
often discussed in verbal or written models. As with
competition, biologists disagree about the role and
mechanism of predation in limiting animal popula-
tions. In any case, if a biologist believes that a preda-
tor is or could limit a population, presence or
abundance of the predator should be measured and
included as a habitat factor in the model. Human
hunting, which may be considered a special case of
predation, is discussed below.

Presence, Absence, or Abundance of Parasites
or Diseases. Like the previous two "habitat compo-
nents," parasites or diseases are rarely considered
in formal models. However, the role of parasites or
disease in limiting animal populations is well-docu-
mented in many specific cases. As with competition
and predation, an impact to a wildlife population
from parasites or disease is often correlated with, if
not caused by, a change in physical habitat condi-
tions. In the case of disease, however, very little
study has been done on correlating disease impacts
with changes in habitat.

Presence, Absence, or Degree of Human Pis-
turbance. Disturbance includes a wide variety of
factors other than physical or vegetative features of
the environment, such as noise, highway construc-
tion, drilling, etc. In many cases, these are very im-
portant factors affecting abundance of vertebrates. In
fact, they are frequently factors that biologists are
asked to quantify in analyses such as EISs and EARs.
Yet very few, formally published habitat models in-
clude disturbance as an explicit habitat component.
Some models do consider disturbance as an implicit
factor, however. For example, Schroeder ( 1984)
developed an HSI model for black brant in which
roosting cover was one of the habitat variables. Hu-
man disturbance has also been identified as an im-
portant factor limiting black brant populations.
Therefore, in the model, roosting cover is defined as
the "percent of shoreline . . . that contains sandy
areas that are isolated from human disturbance"
(Schroeder 1984). Thus disturbance is considered as
either present or absent.

A more challenging problem is determining the
effects of a given degree of disturbance on a wildlife
population. Few, if any, formal models have even
attempted to do this. Yet, such information is needed
for biologists to evaluate the impacts to wildlife from
the myriad human activities occurring in wildlife
habitats. I would expect that in future years biolo-
gists will be developing and using more models that
incorporate disturbance as a habitat factor.

Presence, Absence, or Intensity of Hunting or
Harvesting. Like other factors above, hunting is not
generally considered a habitat factor. However, in
the context of habitat models, it is useful to consider
it such, since it obviously can limit animal popula-
tions. Its impact is also greatly interdependent on
physical and vegetative factors, such as the amount
or quality of cover. If hunting pressure is limiting
wildlife populations, then habitat models that only
consider physical or vegetative habitat components
will be poor predictors of population responses to
habitat management or disturbance. Therefore, in
developing and using habitat models, biologists must
first consider if hunting is limiting the population. If
hunting is not a factor, which is the case with vir-
tually all nongame species and many game species in
this country, the biologist may safely omit it from
further consideration. However, if it is or appears to
be a factor, the biologist must take it into account.
Analyzing impacts of habitat alteration on wildlife
populations has little value when physical or vegeta-
tive habitat factors are not limiting.

Weather and Climate. Weather is the state of the
atmosphere at a given time; climate refers to the
characteristics of the atmospheric conditions of a re-
gion. Or, as one wag described it, "climate is what
you expect; weather is what you get." Whereas cli-
mate is relatively predictable, weather is not, and
weather strongly influences many wildlife popula-
tions. Furthermore, as Bailey (1984) pointed out,
even though the biologist or manager has no control
over the weather, when devasting impacts occur to
wildlife populations from severe weather conditions,
the manager may be blamed, "for there is no satisfac-
tion in blaming the weather." Therefore, in inventory
and monitoring, biologists must incorporate weather
measurements or considerations into their study
design whenever dealing with populations that are
heavily influenced by weather. Because of the impor-
tance of weather and climate in determining animal
abundance, an entire chapter in this book (Chapter
35, Weather and Climate) has been devoted to the
subject. Weather is not included explicitly in most
formal habitat models; however, it is generally recog-
nized as a concomitant influence that affects animal
abundance in many ways and, under extreme condi-
tions, may override the importance of all habitat
variables.

762

Habitat Evaluation Systems

Historical Occurrence. Historical occurrence is
not generally thought of as a habitat component, yet
it can be an important characteristic of habitat. If
an animal species has been recorded as occurring on
an area within historic times, then such information
is excellent evidence that the species can or does
occur there as long as the physical habitat has not
been drastically altered. Of course, the shorter the
time since that occurrence has been verified, the
more probable it is that the species persists there.
Even when a species is known to have been extir-
pated, historical occurrence is important information
when considering such things as reintroductions.
Introduction of animals, such as peregrine falcons
(Falco peregrinas) and bighorn sheep (Ovis cana-
densis) into areas that were not historic habitats, has
rarely been successful.

Model Components — Population Attributes

As with the habitat components of a habitat model,
many population attributes can be predicted (Table
2 ). These may be either characteristics of a single
species population, multiple species populations, or
community characteristics, such as species richness.
However, unlike with habitat components, most of
these models can only predict one population or
community attribute. In other words, several habitat
components may be used as predictors, but only one
population attribute is normally predicted.

Most models have been developed to predict
attributes of single species. In fact, most wildlife
research and management has been single-species
oriented. Furthermore, even many multiple species
models, such as those for species richness, are devel-
oped from a set of single-species models.

Presence or Absence. The presence or absence of
a species is the simplest sort of population attribute
that can be predicted. However, for many purposes,
such a determination is adequate information for the
majority of wildlife species.

Probability of Occurrence. Probability of occur-
rence considers presence or absence as a probability
rather than as an absolute. Since models produce
predictions rather than absolute truths, uncertainty is
always associated with their predictions; use of
models that predict occurrence probability are more
realistic because they attempt to quantify the uncer-
tainty associated with the predictions. For example, a
biologist may need to predict if an area is used by
prairie falcons for nesting. Yet time is not available
to make an absolute determination since the analysis
must be made in winter and the proposed impact
will occur during the next breeding season. Using a
model that has probability of occurrence as an out-
put, a biologist can predict, based on habitat compo-
nents, that "the probability of the site being

occupied by prairie falcons is 90% ." Models for
probability can be used for other purposes, such as
predicting the probability that a population of a cer-
tain density occurs. The most common use of proba-
bility measures of animal populations as predicted
outputs is in "pattern recognition (PATREC)
models," discussed in more detail later in this
chapter.

Abundance or Density. Abundance here refers to
the number of animals in a population, whereas den-
sity refers to the number of animals per unit area.
Abundance or density is quite difficult to predict.
Thus, very few models attempt to actually predict
such attributes. In most situations, biologists neither
can nor need to make such precise estimates.

Relative Abundance or Density Index. Abun-
dance or density indexes are values that correlate
with abundance or density, but for which the quanti-
tative relationship between the two is unknown.
For example, an area with a mule deer
(O. hemionus) index of 0.5 may be expected to
have twice as many deer as an area with an index of
0.25. However, the relationship between the index
and the actual number of deer is unknown. If, for
example, an index of 0.5 was known to correspond
with a density of 30 deer per square mile, then it
would not be a density index, but rather a predictor
of density. Abundance or density indexes in habitat
models (predicted density indexes), like the meas-
ured abundance or density indexes described in
Chapter 2, Data Types, are most useful for compari-
sons. A biologist can compare different areas or the
same area during different years or under varying
habitat impacts.

Carrying Capacity. Carrying capacity is considered
by range managers to be the number of animals an
area can support. It is, thus, a property of the land
and the condition of the habitat, but is expressed in
terms of numbers or density of animals. The actual
number of animals present may be lower than the
carrying capacity. Carrying capacity is unaffected, for
example, by weather, hunting, or other decimating
factors that are not habitat related. Although the
term has generated much controversy and discussion
(Bailey 1984:279-303), it remains a useful working
concept that allows a biologist to quantify the value
of habitat without considering short-term fluctua-
tions in animal numbers from weather, hunting, or
predation. Carrying capacity has been used as an
output in many models (models developed for ungu-
lates in which carrying capacity is assumed to be
primarily a function of forage supply).

Carrying-Capacity Index. Carrying capacity in-
dexes are similar to density indexes except, like with
carrying-capacity estimates, they are strictly a func-
tion of habitat. The most common carrying-capacity

Habitat Evaluation Systems

763

indexes are the habitat suitability indexes used in
the U.S. Fish and Wildlife Service HSI models
(Schamberger et al. 1982). These models use an
index that varies from 0 to 1 , with 1 being some
maximum carrying capacity, and 0 indicating unsuit-
able habitat.

Biomass. Biomass is rarely used in habitat models
since it is usually not of interest to managers, at least
with terrestrial species. However, it is often used in
ecological models designed for understanding eco-
system processes. Nonetheless, some management
models do exist, particularly in fisheries. For exam-
ple, Binns (1979) describes a model that uses nine
stream habitat variables to predict the standing crop
of trout.

Geographic Distribution. Geographic distribution
is basically an extension of presence or absence. In
presence and absence models, the biologist merely
predicts whether a species will be present on a
given site. By using habitat attributes at a series of
sites, one can predict the geographic area that a
wildlife species can be expected to occupy. One
may also note that "geographic location" and "his-
toric distribution" are also described as a model
input. Thus, according to the description of models
provided here, historic distribution can be used,
for example, to predict geographic distribution. This
may seem to be rather trivial. However, this is, in
fact, a rather commonly used model, even though it
is rarely identified as such. The implicit model, used
by most biologists as historical records of a species
in an area, is the best predictor of future occurrence,
and the more recent the historical observation, the
greater the predictive power. However, as discussed
under historical occurrence, past observation is not a
perfect predictor; conditions often change, resulting
in local extirpations, and the biologist needs to be
alert to such situations.

Mortality or Natality. Some models predict the
mortality or natality rather than a population density.
Mortality is usually expressed in terms like number,
density, or abundance of animals killed over winter.
Similarly, natality is usually expressed as some meas-
ure of productivity, such as young fledged per nest,
or as some ratio of young to old animals, such as
fawns per doe. When used, both types of measure-
ments are considered useful indicators of the
"health" of the species population.

Animal Condition. Animal condition, expressed in
such terms as weight, kidney fat indexes, etc., is
also used as a model output and, like productivity, is
considered a surrogate measure of the "health" of
the population.

Population Dynamics. In some cases, biologists
may be more interested in predicting the dynamics
of a population rather than a static or fixed popula-

tion level. All populations fluctuate in numbers and,
in some cases, the frequency and amplitude of such
fluctuations may be of more interest than the "aver-
age" population size. Although numerous population
dynamics models are available, most do not use any
physical or vegetative habitat features to predict
population responses. They use weather, hunting
pressure or, in some cases, predation. Thus they are
not commonly considered habitat models.

Very few models attempt to predict changes in
wildlife populations over time as a result of changes
in habitat factors, although a few are available
(Cooperrider and Bailey 1984; Cooperrider and Beh-
rend 1980). Models with dynamic outputs (ex-
pressed as changes over time ) require at least one
dynamic input (changes independently of the
model). Such inputs are said to "drive" the model. In
most dynamic or simulation models, weather is the
dynamic input; however, other inputs can be used.
The model of Cooperrider and Bailey ( 1984), for
example, can use forage production to drive the
model.

Model Components — Multiple Population
or Community Characteristics

Characteristics of several populations are the
plural counterparts of the population characteristics
just described. They are predicted from similar habi-
tat attributes. However, instead of making inferences
about only one species, some inference is made
about ( 1 ) several taxonomically related species in an
area (e.g., songbirds, ungulates); (2) several species
that use similar habitat components (e.g., cavity-
nesting birds); or (3) all species in a habitat.

Species Richness. Species richness is the number
of species present in an area. It is the plural counter-
part of presence or absence. Models that predict
species richness are usually composed of a series of
individual species presence and absence models, i.e.,
the models usually not only predict the number of
species, but also the specific species.

Species Diversity. Diversity or species diversity
refers to not only the number of species but also
their relative abundance in an area. Species diversity
is rarely used directly as an output from habitat
models, at least the type of habitat models used in
land management. This is not surprising, since a
model to predict species diversity must predict not
only presence or absence but also density or relative
abundance of each species present. Some models,
however, are based on the premise that diversity of
vegetation structure is correlated with species diver-
sity as measured by some species diversity index.
(See Chapter 2 for a discussion of diversity indexes.)
These models predict an index of species diversity

764

Habitat Evaluation Systems

without explicitly predicting the species and their
relative abundance.

Biomass. Biomass of all species, or all species
within a denned species group such as small mam-
mals, fishes, or ungulates, is sometimes used as a
model output. As with individual species biomass, it
is used most commonly with fish.

Guilds or Life-Forms. Guilds and life-forms are
both groups of species that use similar habitat
components for feeding, breeding, or other
biological functions. Several systems have been
devised that use guilds or life-forms as outputs, as
described in the next section. The advantage to the
biologist is that he or she can deal with 10 or 20
guilds, rather than 400 or 500 vertebrate species in
an area. Predictions can be made and verified in
terms of impacts on whole guilds rather than on
individual species. Life-forms, as discussed later in
the chapter, may be considered a special case of
guilds.

Community Suitability Index. In some cases, a
rating of the relative value of a community to sup-
port a wildlife community is the output from a
model. Such indexes are either implicitly or explic-
itly thought to be correlated with a population meas-
urement, such as species richness or species
diversity. For example, Short (1984) describes a
model for generating a suitability index for a habitat
based on the layers of habitat present. It is further
based on the assumption that structurally more com-
plex habitats have more wildlife guilds and greater
species richness.

Model Components — Correlation Functions

The precision, accuracy, and usefulness of habi-
tat models depend on the selection of the appropri-
ate habitat components for use. Equally important,
however, is the process of linking these variables
together in a meaningful way so the best prediction
can be made. A thorough description and set of
guidelines on model development and construction
is beyond the scope of this publication. However,
a few critical concepts and principles are discussed
under the Developing Habitat Models section, fol-
lowing. For a more thorough treatment of model
building, see the U.S. Fish and Wildlife Service Eco-
logical Services Manual on "Standards for the De-
velopment of Habitat Suitability Index Models ( HSI )"
(U.S. Department of the Interior, Fish and Wildlife
Service 1981 ), which contains a good introduction
to the subject. That discussion focuses on one partic-
ular type of model; however, the procedures can
generally be applied to other classes of models.

DEVELOPING HABITAT MODELS

The following basic steps can be used in devel-
oping a model:

( 1 ) Determine model objectives;

(2) Select and quantify habitat variables;

(3) Determine the correlation or prediction func-
tion;

(4) Document and verify the model.

These steps correspond roughly to the phases of
construction described in the U.S. Department of
the Interior, Fish and Wildlife Service Ecological
Services Manual ( 1981 ).

Determining Model Objectives

Model objectives should be based on the objec-
tives of the inventory and monitoring effort, as de-
scribed in more detail in Chapter 1, Inventory and
Monitoring Process. The process includes determin-
ing the desired output (e.g., species density, relative
density, species richness) as well as the geographic
area, seasons, and other factors the model should
predict.

Selecting and Quantifying Habitat Variables

Determining habitat variables that should be
incorporated into a model is critical. A biologist
should use the least number possible, since habitat
measurements are expensive to obtain. These varia-
bles should be the "limiting factors" as described
in basic ecology texts (Odum 1969) or surrogates
for them.

The concept of a surrogate variable is important
in model construction. A surrogate variable is one
that closely relates to a habitat variable of interest
but which, for one reason or another, is easier to
measure or verify. For example, a Lewis' woodpecker
(Melanerpes leivis) is reported to need forests hav-
ing tree canopy closures less than 75% and does
best in areas with 0 to 30% closure (Sousa 1983).
However, within the range of 0 to 75% canopy clo-
sures, the basal area of trees highly correlates with
canopy closure in many areas. Furthermore, basal
area is usually easier to measure than canopy closure
or may be available from sources such as timber
inventories. Therefore, the biologist may prefer to
substitute basal area as a surrogate for canopy
closure.

If a model is being used to predict the impacts
of a particular management action, then at least
some of the habitat variables selected should be

Habitat Evaluation Systems

765

those affected by the management action in some
predictable way.

The habitat variables then must be carefully
defined. To convert a conceptual or verbal model
into a written or even quantitative model generally
requires carefully defining the variables. For exam-
ple, basic habitat factors required by bighorn sheep
for survival have been described as food, water, and
escape terrain (Hansen 1980). Escape terrain has
been further described by various authors as open,
steep, rocky, and rugged terrain. The biologist must
define or quantify the terms if the model is to be
used by other biologists or if habitat measurements
are to be repeated. "Open" can be defined as the
percentage of canopy cover of trees and shrubs or as
some measure of visual obstruction. Similarly,
"steep" can be defined as percentage of slope. The
evolution of models from concepts to quantitative
descriptions requires carefully defined and eventually
quantified terms (Figure 2).

Determining the Correlation or Prediction
Function

Once habitat variables have been selected, de-
fined, and quantified, the biologist must determine
the relative importance of each to the others. In the
simplest case, a biologist can determine that all vari-
ables are equally important; these variables can then
be added to provide a simple additive model (Figure
3a). On the other hand, some habitat factors may
be considered more important than others. For ex-
ample, the biologist may decide that food supply
is twice as important as escape terrain. By attaching
a number to the variable, the biologist can weigh the
model so the index produced is influenced more by
food supply than other variables, resulting in a
weighted additive model (Figure 3b).

The relationship between variables may be
much more complicated, however, and the biologist
should try to model it. For example, the additive

American Dipper Habitat

a. Conceptual Model

b. Verbal Model

c. Quantitative Model

"Fast flowing streams in or
near mountains."

"Streams of over 5 cfs above
5,000-ft elevation within 6
miles of mountains."

Figure 2. Progression of models from conceptual to quantitative.
766 Habitat Evaluation Systems

relationship shown for desert bighorn sheep in Fig-
ure 3a may not be realistic. A habitat could have no
water at all and yet have other important compo-
nents. The index value derived would still be 0.66
for equation (a) or 0.75 for equation (b), which are
both fairly high ratings. Yet such habitats do not
support bighorn sheep. For these reasons, a multipli-
cative or limiting factor model (Figures 3c and 3d)
may be more appropriate. With such a model, if any
attribute has a suitability rating of 0, the overall in-
dex will become 0.

Habitat components can be combined in unlim-
ited ways using various weighting factors and mathe-
matical functions. Similarly, seasonal or functional
requirements can be combined to form compound
models. For example, summer habitat requirements
can be combined with winter habitat requirements,
or feeding habitat requirements with breeding habi-
tat requirements.

One of the most difficult problems of combining
habitat attributes is taking into account spatial ar-
rangement of habitat components. Biologists know
that interspersion and juxtaposition of habitat fea-
tures can be very important to animals. However,
quantifying these relationships in a meaningful way is
difficult. The use of geographic information systems
(GISs) to analyze these relationships may assist
greatly in the future.

Numerous statistical and numerical techniques
can assist in determining both the form of the func-
tion and the appropriate weighting. Techniques such
as linear regression, multivariate analysis, or least-
squares curve fitting may all be useful when data are
available. However, in many cases, the biologist must
rely on literature, expert opinion, and existing
sources of information, rather than raw data.

Documenting and Verifying the Model

Documenting a model is important. In many
cases, models are developed by using a group of spe-
cies experts. Even though every function or assump-
tion of the model is not backed up by extensive
research, a model that represents an expert concen-
sus is easy to defend. If the model is being con-
structed for some sort of impact analysis, then
getting interested parties to design or evaluate the
model(s) can be most helpful in obtaining project
impact concensus. Documentation is also important
when made available to other biologists with similar
problems.

A model, once constructed, is basically no more
than an untested or working hypothesis. Ideally, all
models would be thoroughly tested and refined be-
fore use. In practice, most decisions need to be
made with incomplete information. However, moni-
toring can be used to verify models, as discussed
in the next section.

Type of Model

a. Simple Additive Model--- -HSI = Vl + ^2 + Va

b. Weighted Additive HSI = 2Vl + ^ + V3

Model 4

c. Multiplicative Model HSI = (V, x V2 x V3)1

d. Limiting Factor Model HSI = Min (V,, V2, V3)

All four models calculate a habitat suitability index (HSI)
from 0 to 1 for desert bighorn sheep based on three
habitat factors — a forage index (V,), an escape terrain
index (V2), and a water availability index (V3); the latter
indexes also range from 0 to 1.

*

Without water, other important habitat components for
desert bighorn sheep are not sufficient.

Figure 3. Types of quantitative models.

Habitat Evaluation Systems

767

THE HABITAT INVENTORY AND
MONITORING CYCLE AND HABITAT
MODELS

The habitat inventory and monitoring cycle has
been explained in detail in Chapter 2. It is reviewed
here to clarify the central role of the habitat model
in both the cycle and the habitat evaluation systems.
For further discussion of the roles of models in mon-
itoring, see Salwasser et al. ( 1983), Cole and Smith
(1983), and O'Neil and Schamberger ( 1983).

The habitat inventory and monitoring cycle may
be viewed as consisting of five steps:

(1) Scoping,

(2) Data collection and analysis,

( 3 ) Prediction,

(4) Decision/action, and

(5) Monitoring (Figure 4).

These steps correspond roughly to the stages out-
lined in Chapter 1 , but emphasize model develop-
ment, use, or refinement. Habitat models and
evaluation systems are used in all steps, but most
importantly in step 3, where they are the basic tools
for predictions. Similarly, monitoring verifies predic-
tions and can also be used to refine model(s).

Scoping

The driving force for a monitoring program is a
"problem" or a proposed action of some sort. This
may be a very small action, such as a new fence, or a
major action, such as a power plant. It may be a
very loosely defined issue, such as a "need to reverse
habitat deterioration which is causing perceived
decline of deer in an area," or it may be a very spe-
cific proposal with specified alternatives. It may also
be a wildlife problem, such as the deer decline, or
an activity, such as a coal lease that is expected to
adversely affect the wildlife resource. A typical prob-
lem is generally the need to know the resources
present on an area for land-use planning, etc. This is
the driving force for a basic inventory. All these
problems, big and small, well-defined or vague,
should result in the same procedure. An important
principle is that the magnitude and intensity of the
data collection and analysis should correlate with the
perceived impact of the problem.

The next step is to clearly define the problem.
This will generally require some assessment of the
wildlife resources (animals or habitat) present,
which are likely to be affected. This assessment fre-
quently is based on some rapid examination or anal-
ysis of available data. A critical initial decision is to

Problem

Scoping

Data Collection
and Analysis

1 '

Monitoring

Prediction
of Impacts

n

Decision / Action

-*

Figure 4. The habitat inventory and monitoring
cycle.

determine the geographical area of concern and
"bound the study area." This process of taking a
vague problem or proposed action, clarifying it, and
making an initial assessment of potential resource
impacts is commonly termed "scoping."

Scoping may significantly alter the direction of a
project. By identifying actions particularly detrimen-
tal to wildlife at an early stage, a biologist may be
able to direct planners to less detrimental alterna-
tives. This effort may result in much greater benefits
to wildlife resources than a thorough analysis at a
later stage of an action.

At this stage, the biologist uses habitat models,
possibly very vaguely defined conceptually or quali-
tatively. For example, the biologist may know that
forage supply is a limiting habitat factor for a big-
horn sheep population. Therefore, he or she can
generally predict that a proposal to expand a grazing
lease into a critical summer bighorn range will be
detrimental to the bighorn. On the other hand, the
biologist may only have a very vague idea of how
detrimental the impact will be.

If the decision is made to further consider the
proposal, then the biologist must plan the data
needed. At this point, more quantitative habitat
models should be developed. In the bighorn exam-
ple discussed above, the biologist might need to
locate or develop a model that related forage supply
to numbers of bighorn. Forage supply would then
be defined more specifically and quantitatively as, for

768

Habitat Evaluation Systems

example, total pounds of grasses and perennial forbs
on summer range. This habitat variable ( and possibly
others identified in the model ) would then be the
ones for which the biologist would need to gather
data. Depending on the situation, the biologist might
also need to collect data on the bighorn sheep popu-
lation, such as numbers, movements, or lamb
survival.

Data Collection and Analysis

Data collection and analysis may suggest refine-
ment of a preliminary model. For example, after
intensive field work, the biologist may determine
that water supply is also limiting. The biologist then
may need to modify the model to incorporate water
as a habitat factor, and additional data on number
of waters, supplies, etc. may need to be collected.

Prediction

The next step is to use data to predict the effect
of alternative management actions. This step requires
using a model since the data collected must be used
to predict a future state. The models may be simple
or complex. Returning to the previous example,
the biologist might use the model to predict that
expanding the grazing lease would remove 50% of
the available forage, reducing the bighorn population
by 30% . In this case, the biologist has used a quanti-
tative carrying-capacity model. In other situations, a
biologists may be forced to use a simple conceptual
model. The trend, however, is to use more formal
and quantitative models, since these allow more
precision and can be clearly documented.

In the past, biologists have concentrated on
predicting biological consequences of actions. Re-
cently, biologists and economists have been asked to
go one step further and assess economic conse-
quences of these actions. This also requires using
models to attach values to wildlife resources with
and without the action.

Decision/ Action

The next step is the familiar one of choosing an
alternative and taking some action. Except in the
case of a wildlife-related action (such as a habitat im-
provement project), the biologist usually has a minor
role in this process. The manager must weigh poten-
tial impacts to other resources; economics; and other
social, political, and legal factors. However, the
model becomes the basic tool for summarizing the
collected data and making predictions. The model
becomes the tool of the manager, even though the
latter may be primarily concerned with the model
outputs. The model has, in effect, synthesized the
assumptions about limiting habitat factors and their
effects on wildlife populations.

Monitoring

Once an action has been taken, a biologist needs
to monitor the biological effects of the action. This
monitoring can serve two purposes. The biologist
can determine empirically whether the impact to
wildlife resources is as predicted or within tolerable
limits. If the action is not achieving the desired goal
or if the impacts are outside tolerable limits, then
the manager may be able to stop or modify the ac-
tion being taken. The biologist can also determine
the quality of the model in predicting impacts and
modify it as necessary.

Many biologists have advocated adding monitor-
ing as a key element in a resource management pro-
gram and termed it, among other things, "adaptive
environmental assessment" (Holling 1978), "adapt-
ive management" (Barrett and Salwasser 1982),
"cyclic incrementalism," or more simply "muddling
through" (Bailey 1982) or "common sense manage-
ment" (Barrett and Salwasser 1982). These ap-
proaches vary in details, but are all based on the
following assumptions:

( 1 ) Many features of biological systems ( such as
weather) are unpredictable;

(2) The tools for both measuring biological re-
sources and predicting future states are crude
and the time, money, and personnel for such
efforts are limited; and

(3) Continuing selective remeasurement (moni-
toring ) can be effective in both correcting or
improving management actions and also refin-
ing the predictive tools.

The need to monitor may seem obvious; however,
the list of projects and programs that have failed
because monitoring was not included (or was given
the lowest priority ) is long and growing fast.

Monitoring without an underlying model may
be useful in determining whether management ob-
jectives are being met, but it is unlikely to lead to a
better understanding of the system being managed.

Clearly, the five steps of the habitat and moni-
toring cycle are similar and sometimes overlap. Scop-
ing, for instance, is a form of low-level or "quick-
and-dirty" inventory and prediction phase. Although
monitoring may involve similar data collection, it
is done in the inventor}' phase and for different pur-
poses.

In the next sections, I describe the systems,
models, and other programs that can be used when
performing one or more of the tasks identified in this
section.

Habitat Evaluation Systems

769

CURRENT HABITAT MODELS AND
EVALUATION SYSTEMS

example, habitat suitability for the marten (Martes
americana) is a function of four variables:

Models, as considered here, are any formal
method for correlating habitat variables with popula-
tion attributes, whether they are for single species
or for multiple species or communities, regardless of
the purpose for which they were designed and used.

Simple Correlation Models

Single-species models range from very simple
correlations to rather elaborate correlations of spe-
cies presence with many habitat factors. In the sim-
plest case, presence of a species is simply associated
with presence of a vegetation type. More elaborate
variations take into consideration other habitat fea-
tures within that habitat type and other animal popu-
lation attributes. These include season of use;
successional stage of vegetation; special habitat re-
quirements; biological functions for which they are
used (breeding, feeding, resting [cover], special);
relative importance or value of the habitat feature
(critical, optimum, acceptable, marginal, etc. ); and,
in some cases, some qualitative measure of relative
density or abundance (abundant, common, rare,
etc.). The approach is exemplified in the models de-
veloped and reported by Verner and Boss ( 1980;
Figure 5).

Habitat Suitability Index (HSI) Models

HSI models have been developed by the U.S.
Fish and Wildlife Service and others for use in their
Habitat Evaluation Procedures (HEP; U.S. Department
of the Interior, Fish and Wildlife Service 1981 ). In
developing the HSI models, the biologist uses exist-
ing data, literature, and expert opinion to develop an
equation or algorithm to use a small number of se-
lected habitat variables in predicting the suitability
of habitat for a wildlife species. Suitability is indi-
cated by an index ranging from 0 to 1, with 0 indi-
cating unsuitable habitat and 1, optimal habitat. For

( 1 ) Percentage of tree canopy closure,

( 2 ) Percentage of overstory canopy containing fir
or spruce,

( 3 ) Successional stage of stand, and

(4) Percentage of ground surface covered by
downfall over 3 in. in diameter (Allen 1982;
Figure 6).

Since habitat suitability cannot be directly measured,
validation of HSI models is difficult if not impossible.
HSI models are usually developed for a single spe-
cies; however, sometimes they are developed for
a group of species or guilds, such as dabbling ducks,
which respond similarly to the same set of habitat
variables. Approximately 300 HSI models are in
some stage of development in North America, of
which, about half are for aquatic species.

Pattern Recognition (PATREC) Models

PATREC models are similar to HSI models in
that they use a set of habitat variables to predict the
expected capability of the habitat to support an ani-
mal species. They differ from HSI models in that they
use a formal statistical procedure to predict habitat
capability, expressed as a probability. Usually the
probability is of a certain population density or
classes of animals. For example, Williams et al.
( 1977) describe a PATREC model for pronghorn
antelope that predicts the probability that a habitat
with a given set of characteristics will support a high
or low density antelope population. By providing
definitions for high density and low density, the
model can also be used to calculate an expected
population density. PATREC models are generally
developed from field data; however, they can use
data from expert opinion.

§

Opti mum

Suitable

Marginal

Jl ac k-headed
Grosbeak R

12 3 3 4 4
A B A B

1 2 3 3 3 4 4 4
A B C A B C

rfrrHri Wr+rrr

rr

12 3 3 3 4 4 4

A BCABC

rrrrrrr

1 2 3 3 4 4
ARAB

Ett

H

12 3 3 3 4 4 4
A B C A B C

Wririhrt\

12 3 3 3 4 4 4

A B C A BC

12 3 3 3 4 4 4
A B C A B C

12 3 3 3 4 4 4
A BC A B C

FUNCTION

B-Breeding

F-Feed i ng

R-Resli ng

S-Season ot Occurrence

SUCCESSIONAL STAGES
1-Grass/Forb
2-Shrub/Seeding/Saplinc
3-Pole /Med i urn Tree
4-Large Tree

TREE CANOPY COVER CLASSES

A 0-39%

B 4 0 - 6 9 %

C 70%or Greater

Figure 5. An elaborate correlation model (from Verner and Boss 1980).
770 Habitat Evaluation Systems

r8

r|

rO

f8

y1

.to

■ o

-U1

r-~

i^

l^~

<*

<>e

<*

o

LO

in

-CD

. o

c

c\j

c

CM

» -<

in

' CM

D 00

ID

•<t

C\J

3 00

to

■<t

CM

O CO CD

1

CJ

O CO to

Tt CM

O

O

O

O

O

o

O

O

^ ci C)

d

o

-- o d

d d

xapu| Aim

qejjr

S

X3pu| Aii|iqeii

IS

xepu| Aimqe^n

S

X9pu|

Allliqe^ns

TJ

B

tO

^-^

u

to

XI

5

01

>>

U-l

5

M

a. *u

o

0

rn

3

o o

rl

0) r-l

^^

(0

a

01

60

U H

o

CO T3

60 60

10 tO

0

r-4

u 0>

to a

4-1 t-l 60

T>

<4-l IH

o

o

CO

r-l

s g

J»tH

CO r-l C

o

VO

>,

t-l rl

cu

T3 t-I

to O

•

(X

o a

u

H 01 H

rl

T>

r**

O

c

U 0
CO o

2

1 tl pi

a to to

o

•o

AI

rl

OI

tO

rl O

a

O 1 to

01

3 XI

4-1

o

a)

w

•rl XI

60 rl

O

(0

CU

01

> 01

o u

m

to 3 o>

CO El r-t

a 3

3 4J

U T)
60 01

•H

g

01

3

o

1) £ O

riH

rl

x:

•H

M
4J

M-l CO

o o

u

una
u

cm a

O r

5

TJ

l-H

•H

3 ^-N^

rr, /-v

o x;

5

CU

iM!

&•« y

M-l

co <: «

O Q

fr< y

3

r-l
XI

*~\

r"N

/^

>^v

nj

N

^w

>*

•H

t>"

>

>

h

•—•

N->

N-^

Ste*'

CO

>

a

CU

•^

^

^

4J
rl
CO

Q

0

,Si

4J

CO

0

V

^

4-1

a

■H

to

a

a

OI

o>

3 •

14

4J

r-l 9

o

(0

CO o

> r-l

*2

01

<u
a

OI

TS CO

>
o

0

•h

r-l

a o>

y

0-

i

CO
r-l

1-4

CO
<4-l

S

rl
01

u

■H 3

r-l
(►. CO
4J >

OI

3

01
4J

t-i

-o

CU

T-l

r-l

a

4)

CO

O

0

r-l CU

CO

•H

"8

•H

u

•o

CO

•H

■H 10
J3 01

>

a

S3

5

>.

•o

CO J3

M

)-■

0<

X>

4J 4J

a*

o

^•\

y, o>

a

■H

•a

<4-l

M

-a-

3 o

TS

TJ

•H

3 CD

a

CO

S2

a

SI

CU

CO 4J

M

x

II

tn

t>0

CO

t-l

U

^-\

■H

OI

a

0)

r-l P.

IJ

a

01

•

H 15

fr.

T3

*

u

u m

en

o

>

a

4J

(s..

(3 CM

M

3

u

o

•H

4-1

•H

Q»

t^

Q

•H ~^.

01

o

(0

>, u

M-l

60

y

TJ

I-l

^0

r-l

T3

r-4

o

CL O

o eo

co

*J C

T-l

^

,

>s

r-N

a

r-4

iH

o

a

TI

o>

^^*

J= CO

X)

%

•

"N

4-1

*4f

M

O

y

/■> c u

/~\

4) T-l 00

i— 1

y

00

<0

v.

•H

>

<4-t

O CO T-l

o

eor-i c

O

co

0

■n r-~

4J

r-l

S-.

>»0 CJ CM

o

10 t5 t-i

CM

y

r4 >

•H

■H

X

4-1

0>

CXi-l

r-l

U 0) r-l

U

u

r"N

3

XI

•H

-C

o

A| |"S

CO 01 o.

o

3

\D O

4J X!

CO

CO

c*>

I-l

4J

a

A|

CO CO

CO

• o

CO 60

4->

>

•H

CO

r-l | CO

01

1^ r-l

3

•H •

JO

rl

CJ

4-1 TJ

CO Xl 1

60

tl

■o

to o

3 co

X

to

O

i-l I) tl

CM

O 3 Oi

a

3

a

42 a|

r= r4

CO i-4

u

<4-<

«

> rl CO

>

O U 1-4

3

l-l

3

O. J3

4)

CM

•H

01

01 "H

■H J3 O

O

3

o

-*

CO 4J

4-1 T3

>

3

CO

u

> rl

CO CO CL

:►.

rl

-S t>,

u

CO O

CO

01

u

A ° g-

aI

co

60

4iA

60 r-4

4J Q

><

3

1 0

01 ^> /-s

^-s

>

•rl

u

r-l

<M

O <4-l o

o

y < co

u

O

M-l

^

0

X> 60

r-4

CO

(0

o

^ O CJ

v^

y

o

o o

co a

>

4J

>

M

M

3
CO

K

l-l 4-1
14-1

^S

s^

J3
CO

S

^

/-^

^v

^^

CU c

CO O
i— 1

n

n

o

r-l

CM

CI

«*

c o

01 i-H

M

T)

>

>

>

>

•H O.

1-4 CM

ro <f

4J 0

to

0>

> —

V"'

v-^

s-^

0 CO
Li 01

> >

> >

co u-i

r-l

OS

4J

01 U

3 01

T-l

4J U

y -c

W

01 O

rH 4J

*

Q y

CO

O 60

*-\

1-V

a

^ TH

rH

CM

n co

>-•

^ 3

X

c

V

■a

tC

o
S

C

u

c

'a,
t-
c2

"v
■o
o
E

53
X
c
<

U

3
6C

Habitat Evaluation Systems

771

Habitat Capability Models

Habitat capability models, as used here, are simi-
lar to HSI models, except the habitat variables are
used to predict an animal density. Thus, these
models could include the traditional carrying-capac-
ity models used for many ungulates. Clearly, some
PATREC models fall into this category when the
population variable is some measure of population
density. The U.S. Forest Service uses the term habitat
capability models to include PATREC models and
HSI models; however, the U.S. Forest Service con-
verts the models so the output is some measure of
population density.

Matrix Models

The simplest form of multi-species models is the
species-habitat matrix, in which the presence of each
species in a cover type is indicated by an X or other
similar notation. Many variations on this approach
exist. A relative density (e.g., H = high, M = me-
dium, L = low) or abundance (e.g., A = abundant,
C = common, U = unusual, R = rare, etc. ) may be
indicated. Similarly, the season of use (e.g., W =
winter, Sp = Spring, etc.) or breeding status (e.g., B
= breeds in habitat, N = does not breed in habitat )
may be indicated. Such information is commonly
provided by computer-based systems such as RUN-
WILD (Patton 1979) and "Procedures" (U.S. Depart-
ment of the Interior, Fish and Wildlife Service
1980a).

Such models have commonly been used in doc-
uments such as EISs to describe current wildlife
resources that may be affected by a land-use deci-
sion; however, they are not particularly useful for
predicting which species will actually be affected or
how it will be affected. Such prediction requires that
species presence or density be related to compo-
nents of the habitat, such as vegetation structure or
physical features. Several systems, including life-
forms and guilding, do this to some extent.

The Life-Form Approach

The life-form approach was developed in Ore-
gon by Thomas ( 1979), based on earlier work by
Haapanen (1965) on birds in Finnish forests. Thomas
and others expanded the concept to include all ter-
restrial vertebrates. A life-form is defined as "a group
of wildlife species whose requirements for habitat
are satisfied by similar successional stages within
given plant communities" (Thomas 1979:-i<S0)

For the Blue Mountains of Oregon, Thomas
(1979) aggregated over 300 vertebrate species into
16 life-forms (Figure 7). These life-forms were then
correlated with successional stages of each plant
community. Thus, if a biologist can predict the effect

of management on succession, the life-forms and
species that will be present can also be predicted.

The life-form approach was developed for forest
lands in response to timber management programs,
intended to provide a forest manager with informa-
tion on impacts to wildlife from management prac-
tices. A major impact of logging is to set back forest
succession, and succession in forest lands is rela-
tively predictable. This approach is more suited to
forest lands than rangelands, where livestock grazing
is often the principal commercial use and plant
succession is less predictable. Thomas (1979) also
recognized the need to supplement the life-form
approach with consideration of special habitats (e.g.,
riparian zones); unique habitats (e.g., cliffs, caves);
and separate models for featured species (e.g., mule
deer and elk).

Guilding Approach

Use of a community guild model or "guilding" is
similar to the life-form approach. A guild was origi-
nally defined as a "group of species that exploit the
same class of environmental resources in a similar
way" (Root 1967). Application of the guild concept
appears to have many potential uses in wildlife man-
agement; however, the way guilds are defined and
used varies greatly (Verner 1984). Short (1983, 1984)
and Short and Burnham (1982) have developed for-
mal procedures for applying the guild concept to
wildlife habitat evaluation and have termed it the
"community guild model." Development of the tech-
nique was supported in part by the U.S. Bureau of
Land Management. The most recent version of the
model is described here.

Guilds are defined according to the habitat lay-
ers a species uses for breeding and feeding. Species
are then assigned to appropriate cells based on exist-
ing data and literature. For example, a species, such
as the black-headed grosbeak that breeds in shrubs
and feeds in trees, shrubs, and terrestrial (ground)
surface (Short 1983:186), would be assigned to rows
8, 6, and 5 of column 7 in Figure 8. ( In a life-form
analysis, this bird would be assigned to life-form 1 1 ;
Thomas 1979:323)

Species occupying the same combination of
cells are then lumped into guilds much as done with
life-forms. Using such a model, a biologist can pre-
dict the guilds affecting various layers of habitat that
will be lost or adversely affected by management
actions.

Habitat Evaluation Procedures (HEP)

The HEP system was formally developed and
used by the U.S. Fish and Wildlife Service and some
other federal agencies, particularly the U.S. Bureau of
Reclamation and the U.S. Army Corps of Engineers.

772

Habitat Evaluation Systems

M / / 0 i

A*/ ■? / <? A/ ■?

/ \M> / Q / <Z> / / <&

1

in water

in water

1

bullfrog

2

in water

on the ground, in bushes,
and/or in trees

9

long-toed salamander, western
toad, Pacific treefrog

3

on the ground around water

on the ground, and in bushes,
trees, and water

45

common garter snake, killdeer,
western jumping mouse

4

in cliffs, caves, rimrock, and/or
talus

on the ground or in the air

32

side-blotched lizard, common raven,
pika

5

on the ground without specific
water, cliff, rimrock, or talus
association

on the ground

48

western fence lizard, dark-eyed
junco, elk

6

on the ground

in bushes, trees, or the air

7

common nighthawk, Lincoln's
sparrow, porcupine

7

in bushes

on the ground, in water, or the
air

30

American robin, Swainson's thrush,
chipping sparrow

8

in bushes

in trees, bushes, or the air

6

dusky flycatcher, yellow-breasted
chat, American goldfinch

9

primarily in deciduous trees

in trees, bushes, or the air

4

cedar waxwing, northern oriole,
house finch

10

primarily in conifers

in trees, bushes, or the air

14

golden-crowned kinglet, yellow-
rumped warbler, red squirrel

11

in conifers or deciduous trees

in trees, in bushes, on the
ground, or in the air

24

goshawk, evening grosbeak, hoary
bat

12

on very thick branches

on the ground or in water

7

great blue heron, red-tailed hawk,
great horned owl

13

in own hole excavated in tree

in trees, in bushes, on the
ground, or in the air

13

common flicker, pileated
woodpecker, red-breasted nuthatch

14

in a hole made by another
species or in a natural hole

on the ground, in water, or the
air

37

wood duck, American kestrel,
northern flying squirrel

15

in a burrow underground

on the ground or under it

40

rubber boa, burrowing owl,
Columbian ground squirrel

16

in a burrow underground

in the air or in the water

10

bank swallow, muskrat, river otter

Total:

327

'Species assignment to life form is based on predominant habitat-use patterns.

Figure 7. Life-forms for the Blue Mountains of Oregon (from Thomas et al. 1979).

Habitat Evaluation Systems

773

o
o

D5

C

01

10. Feeds Elsewhere

9. Air

8. Tree Canopy

•

7. Tree Bole

6. Shrub Layer

•

5. Terrestrial
Surface

•

4. Terrestrial
Subsurface

3. Water Surface

2. Water Column

1. Bottom Water Column

CO

CD
O

ZS

o

CO

B

CO

b

CL

E

CD

I—

c
E

o
O

§

"o

E
o

b

CD

c
E

o
O

CD

CD

o

m

ZJ
CO

b

CD

o

ZJ
CO

ID
Z!

CO

to

CD

b

I—

CD

o

ZJ
CO

TO
CD

b

I—

b
>-

CO

_l

JD

ZJ

_c

CO

CD
O

CD

CD
CD

h1

CL

o

c

CCS

o

CD
CD

H

CD

b

JZ

g

CD
CO

UJ
CO

"O
CD
CD

CO

1.

2.

3.

4.

5

6

7

8.

9.

10.

Breeding Loci

Figure 8. Habitat layers used to define guilds for use in the community guild model (from Short 1983).

774

Habitat Evaluation Systems

The basic building blocks of the system are the HSI
models previously described. One or more HSI
models are used to represent components of habitat
potentially affected by a project. The index values
are multiplied by the acres of habitat to produce
"habitat units" present before and after the project.
These habitat units are then used to determine equi-
table mitigation. An increase in habitat units from
mitigation is calculated in the same way to produce
the number of units before and after a mitigation
action.

HEP is used mostly by the U.S. Fish and Wildlife
Service and development agencies to analyze water
projects that have legal requirements for equitable
mitigation. It is used much less on projects that do
not have these requirements.

An important feature and one of the major
strengths of the HEP system is the emphasis on get-
ting all interested parties together early in the evalu-
ation to agree on the species and precise HSI models
to be used in the analysis. This helps prevent later
disagreement regarding the results of the evaluation.

In practice, HEP is used with very little follow-
up monitoring.

The HEP system is thoroughly described by the
U.S. Department of the Interior, Fish and Wildlife
Service ( 1980b) with various applications discussed
in Schamberger and Farmer ( 1978) and Urich and
Graham (1983).

Wildlife and Fish Habitat Relationships
(WFHR) System

The U.S. Forest Service has developed a system
of habitat evaluation called the Wildlife and Fish
Habitat Relationships (WFHR) system. This is not a
formal system like HEP, but rather a set of standards
and collection of evaluation tools to be used by
agency biologists (Nelson and Salwasser 1982). Biol-
ogists may do evaluations by using the life-form ap-
proach or HSI, PATREC, or habitat capability models
as appropriate (Sheppard et al. 1982). In addition, on
some projects, the U.S. Forest Service makes use of
dynamic simulation models such as DYNAST to pre-
dict forest succession after timber harvesting or
other disturbance. Static models such as HSI models
are then used to predict the wildlife species or abun-
dance associated with future successional stages
(Barrett and Salwasser 1982). The U.S. Forest Service
also emphasizes and, thus, has developed methodol-
ogy to analyze cumulative impacts (Salwasser et al.
1983) and viable populations (Salwasser and Samson
1985). Methodology for such analyses is still primi-
tive but an important subject that will likely be used
more and more in the future.

SUMMARY AND CONCLUSIONS

The field of habitat evaluation has advanced
rapidly in the past 1 5 years. The concept of a habitat
model as a basis for habitat evaluation has been de-
veloped and used in a variety of ways. Although such
models provide prototypes for biologists, the biolo-
gist will need to specifically develop models for the
area and project under consideration. Such model
building is basically nothing more than a more for-
mal and quantifiable approach to the conceptual and
qualitative models that biologists have used in the
past.

Habitat models are clearly an integral part of the
inventory and monitoring process, although they
may not be labeled as such. Although monitoring can
be done without any sort of an explicit model, such
monitoring is not likely to increase the understand-
ing of wildlife resources being managed, nor lead
in the long run to better management. Human dis-
turbance factors are incorporated into very few
models, yet they are among the most important im-
pacts to wildlife. Future habitat models will need
to be developed to incorporate these factors.

The U.S. Fish and Wildlife Service HEP proce-
dures, by emphasizing the process of gaining concen-
sus on models to be used in an analysis, have
demonstrated a useful way to alleviate confrontation
in the habitat evaluation process. Disagreements
focus on analysis tools that are easier to evaluate ob-
jectively, rather than on analysis results. This ap-
proach promises to be useful in many situations,
regardless of whether the biologist is using other for-
mal aspects of HEP.

The U.S. Forest Service has developed and ap-
plied many new approaches. The linking of dynamic
simulation models of succession with static habitat
models, which the U.S. Forest Service pioneered,
should be useful wherever succession is reasonably
predictable. Similarly, the efforts of U.S. Forest Ser-
vice personnel and collaborators to develop and use
cumulative effects models and viable population
analysis models are pioneering efforts that are both
necessary and promising. Future habitat modeling
will likely emphasize these two areas.

As with other areas of inventory and monitoring,
use of habitat models and evaluation systems re-
quires the biologist to use sound, professional judg-
ment and to keep aware of current developments.
The techniques will probably continue to develop
rapidly, and the biologist cannot afford to be content
with "cookbook" procedures or outdated
technology.

Habitat Evaluation Systems

775

LITERATURE CITED

ALLEN, AW. 1982. Habitat suitability index models: Mar-
ten. U.S. Dep. Inter., Fish and Wildl. Serv. FWS/OBS-
82-10.11. 9pp.
BAILEY, J.A. 1982. Implications of "muddling through" for
wildlife management. Wildl. Soc. Bull. 10:363-369.

. 1984. Principles of wildlife management. John

Wiley and Sons, New York, NY. 373pp
BARRETT, R.H. and H. SALWASSER. 1982. Adaptive man-
agement of timber and wildlife habitat using dynast
and wildlife-habitat relationship models. Paper pre-
sented at Western Assoc, of Fish and Wildlife Agen-
cies, Las Vegas, NV, July 21, 1982. 16pp. (mimeo)
BINNS, N.A. 1979. A habitat quality index for Wyoming
trout streams. Wyoming Game and Fish Dep., Fish.
Res. Rep. 2. 75pp.
COLE, C.A. and R.L. SMITH. 1983. Habitat suitability in-
dices for monitoring wildlife populations. Trans.
North Am. Wildl. and Nat. Resour. Conf. 48:367-375.
COOPERRIDER, AY. and J.A. BAILEY. 1984. A simulation
approach to forage allocation. Pages 525-559 in De-
veloping Strategies for Rangeland Management. Rep.
prepared by the Committee on Developing Strategies
for Rangeland Manage., Natl. Res. Council/Natl. Acad.
Sciences, Westview Press, Boulder, CO. 2022pp.
and D.F. BEHREND. 1980. Simulation of the inter-
action of deer with northern forest vegetation. New
York Fish and Game J. 27:142-155.
HAAPANEN, A. 1965. Bird fauna of the Finnish forests in
relation to forest succession. I. Ann. Zool. Fenn.
3(3):153-196.
HANSEN, C.G. 1980. Habitat. Pages 64-79 in Monson, G.
and L. Summer, eds. The Desert Bighorn. Univ. Ari-
zona Press, Tucson.
HOLLLNG, C.S., ed. 1978. Adaptive environmental assess-
ment and management. Vol. 3, International Series on
Applied Systems Analysis. John Wiley and Sons, New
York, NY. 377pp.
NELSON, R.D. and H. SALWASSER. 1982. The Forest Ser-
vice wildlife and fish habitat relationships program.
Trans. North Am. Wildl. and Nat. Resour. Conf.
47:174-183.
ODUM, E.P. 1969. Fundamentals of ecology. W.B. Saunders

Company, Philadelphia, PA. 546pp.
O'NEIL, L.J. and M.L. SCHAMBERGER. 1983. Habitat

models as a monitoring tool. Pages 424-427 in Bell,
J.F. and T. Atterbury, eds. Renewable Resource Inven-
tories for Monitoring Changes and Trends. SAF83-
14, Oregon State Univ., Corvallis.
PATTON, DR. 1979. RUNWILD II: A storage and retrieval
system for wildlife data. Trans. North Am. Wildl. and
Nat. Resour. Conf. 44:425-430.
PETERSON, R.T. 1969. A field guide to western birds:
Field marks of all species found in North America
west of the 100th Meridian, with a section on the
birds of the Hawaiian Islands. 2nd ed. (revised and
enlarged). Houghton Mifflin Co., Boston, MA. 336pp.
ROOT, R.B. 1967. The niche exploration pattern of the

blue-gray gnatcatcher. Ecol. Monogr. 37:317-350.
SALWASSER, II., C.K. HAMILTON, W.B. KROIIN, J.F. LIP-
SCOMB, and C.H. THOMAS. 198.3. Monitoring wildlife

and fish: Mandates and their implications. Trans.
North Am. Wildl. and Nat. Resour. Conf. 48:297-307.

and F.B. SAMSON. 1985. Cumulative effects analy-
sis: An advance in wildlife planning and management.
Trans. North Am. Wildl. and Nat. Resour. Conf.
50:313-321.

SCHAMBERGER, M.A. and AH. FARMER. 1978. The habi-
tat evaluation procedures: Their application in project
planning and impact evaluation. Trans. North Am.
Wildl. and Nat. Resour. Conf. 43:274-283-

, , and J.W. TERRELL. 1982. Habitat suitability

index models: Introduction. U.S. Dep. Inter., Fish and
Wildl. Serv. FWS/OBS-82- 10. 2pp.

SCHROEDER, R.L. 1984. Habitat suitability index models:
Black brant. U.S. Dep. Inter., Fish and Wildl. Serv.
FWS/OBS-82/ 10.63. 11pp.

SHEPPARD, J.L., D.L. WILLS, andJ.L SIMONSON. 1982.
Project applications of the Forest Service Rocky
Mountain Region wildlife and fish habitat relationships
system. Trans. North Am. Wildl. and Nat. Resour.
Conf. 47:128-141.

SHORT, H.L. 1983. Wildlife guilds in Arizona desert habi-
tats. U.S. Dep. Inter., Bur. Land Manage., Tech. Note
362. 258pp.

. 1984. Habitat suitability index models: The Ari-
zona guild and layers of habitat models. U.S. Dep.
Inter., Fish and Wildl. Serv. FWS/OBS-82/ 10.70. 37pp.

and K.P. BURNHAM. 1982. Techniques for structur-
ing wildlife guilds to evaluate impacts on wildlife
communities. U.S. Dep. Inter., Spec. Sci. Rep. — Wildl.
244. 34pp.

SOUSA, P.J. 1983 Habitat suitability index models: Lewis'
woodpecker. U.S. Dep. Inter., Fish and Wildl. Serv.
FWS/OBS-82/ 10.32. 14pp.

THOMAS, J.W., ed. 1979. Wildlife habitats in managed

forests — the Blue Mountains of Oregon and Washing-
ton. U.S. Dep. Agric, For. Serv., Agric. Handbook
553. 511pp.

URICH, D.L and J.P. GRAHAM. 1983. Applying habitat

evaluation procedures ( HEP ) to wildlife area planning
in Missouri. Wildl. Soc. Bull. 11:215-222.

U.S. DEPARTMENT OF THE INTERIOR, FISH and WILD
LIFE SERVICE. 1980a. An evaluation of "A procedure
for describing fish and wildlife in Pennsylvania." U.S.
Dep. Inter., Fish and Wildl. Serv. FWS/OBS-79/19-A.
15pp.

. 1980b. Habitat evaluation procedures (HEPs) 102

ESM. U.S. Dep. Inter., Fish and Wildl. Serv. (unnum-
bered)

. 1981.

Standards for the development of habitat
suitability index models. 103 ESM. U.S. Dep. Inter.,
Fish and Wildl. Serv. (unnumbered)

VERNER, J. 1984. The guild concept applied to manage-
ment of bird populations. Environmental Manage.
8(1):1-13.

and A.S. BOSS, eds. 1980. California wildlife and

their habitats: Western Sierra Nevada. U.S. Dep. Agric,
For. Serv., Gen. Tech. Rep. PSW-37. 439pp.

WILLIAMS, G.L., K.R. RUSSELL, and W K. SEITZ. 1978.

Pattern recognition as a tool in the ecological analysis
of habitat. Pages 521-531 in Marmelstein, A., ed.
Classification, Inventory' and Analysis of Fish and
Wildlife Habitat. U.S. Dep. Inter., Fish and Wildl. Serv.
FWS/OBS-78/76. 604pp.

776

Habitat Evaluation Systems

39

EVALUATION AND
INTERPRETATION

Allen Y. Cooperrider

U.S. Bureau of Land Management
Service Center
Denver, CO 80225

"Statistical analysis is not a substitute for thinking."
— Alan Speigel, Speigel's Laws of Management

Editor's Note: The best data sets and most elegant
quantitative analyses will not likely affect manage-
ment unless they are interpreted and evaluated for
those who make the decisions. This process requires
the biologist to creatively bring together informa-
tion from various sources. These sources include
not only data collection and statistical analyses of
them, but also field observations, experience from
other areas or projects, and opinions from others
knowledgeable about the subject or area. Few biol-
ogists are effective in influencing decisions, unless
they have mastered the art of evaluation and inter-
pretation. This chapter discusses the process and
some factors a biologist should consider at this
stage of an inventory or monitoring effort.

INTRODUCTION

Interpretation and evaluation of data are critical
steps in the inventory and monitoring process. At
this stage, the biologist must explain what all the
data, information, and analyses mean. This is not a
formal process; if it were, it would be included with
other analytical methods. Since the process is essen-
tially subjective, very little has been written about it.
Why, then, include a chapter on such a nebulous
subject? Because biologists often do a poor job of
interpretation and evaluation or tend to skip this
step.

ciAk..,

Evaluation and Interpretation 777

Biologists frequently present a manager with
lists of species present in habitats, correlation coeffi-
cients, t-tests, or other test statistics and expect a
manager to understand how these relate to the land
management decisions to be made. The manager
does not understand; therefore, the biologists are not
performing their jobs as thoroughly and profession-
ally as they could. A biologist is tasked with explain-
ing the meaning of data; suggesting implications from
the findings; cautioning about known or potential
weakness in the data; and explaining the significance
of the anticipated resource impacts. The biologist
who fails to do this has not completed the job and is
not fully using the information. Collecting and ana-
lyzing data are frequently expensive, and the biolo-
gist should not waste information derived from such
data.

The art of interpretation and evaluation cannot
be taught, any more than one can be taught to think
or to use good judgment. The aims of this chapter
are more modest. In it, I explain what interpretation
and evaluation are and list some of the many possi-
ble aspects that must be considered during the pro-
cess. Biologists hoping to improve their abilities to
interpret and evaluate may find some useful guid-
ance. However, the process cannot be reduced to a
cookbook procedure.

Interpretation and evaluation is thus one of the
most important and challenging tasks of the profes-
sional, yet one which is frequently done poorly and
with very little guidance. The reasons biologists
struggle with these tasks usually relate to their train-
ing and background. Many wildlife biologists are
thoroughly trained in the scientific methods and
have experience in research. This scientific training
is quite valuable and important. However, it tends to
make biologists cautious. Scientists are taught to —

( 1 ) Collect data carefully and systematically,

( 2 ) Concentrate on what the data demonstrates
conclusively,

(3) Avoid predicting trends, and

(4) Leave out value considerations.

A management biologist should also collect data
carefully and systematically. However, the manage-
ment biologist must operate quite differently with
the other three factors since these are opposite from
what a biologist is asked to do in interpretation and
evaluation. Typically a manager wants predictions
about the future. Rarely does a biologist have
enough data to predict in a scientific way what will
happen on an area; instead he or she has many di-
verse sources of data and information upon which to

base some future states. This is needed and wanted
by management to make an informed decision. A
scientist can afford to wait for conclusive results; a
resource manager cannot.

Similarly, biological scientists do not deal much
with values other than "scientific value" and that
elusive "ecological value;" most biologists are taught
that consideration of values is the domain of philoso-
phy, economics, sociology, psychology, or religion.
Yet, a management biologist must constantly deal
with values. Biologists are paid to manage wildlife
and wildlife habitat because people place some value
on those resources. But, those values are quite di-
verse. A professional biologist must be aware of
these values and explain them to the manager.

Ideally, a scientific manager would have "hard"
data and conclusive evidence on which to base all of
his or her decisions. In real life, in resource manage-
ment as well as in agriculture, business, economics,
politics, and other fields, decisions are made on
"soft" data. A decisionmaker may have good data on
one decision factor and only an opinion or limited
data on other factors. A manager would be foolish to
make decisions based only on factors for which good
data were available, if he or she knew other factors
were also important.

INTERPRETATION

To interpret has been defined as "to explain or
tell the meaning of: present in understandable terms"
(Merriam- Webster 1983)- Once data are collected,
organized, put on the computer, processed and mas-
saged in a variety of ways, and the results printed,
the information needs to be explained. Casual obser-
vations while out in the field and talks with other
people in the field; specialists in other disciplines;
and other biologists familiar with the area or the
problem being addressed, are other sources of infor-
mation. The experience of the biologist with similar
problems or areas is also a source of information.
All these sources can be used to explain to the pub-
lic, the manager, or the decisionmaker what the data
mean. They can also be used to predict a future with
or without a proposed action. Some of the following
questions should be considered.

What do the data mean? A biologist may have
been studying the behavior and foraging habits of elk
{Cervns elaphus) to determine if expansion of a
grazing lease into its summer range will adversely
affect the herd. The biologist has measured elk and
cattle use in a series of adjacent pastures. After ana-
lyzing the data, he or she may have concluded with
the aid of a statistician that "use of pastures by elk is
significantly and negatively correlated with use by
cattle (R" = 0.99)." More obscurely, "Spearman's

778

Evaluation and Interpretation

rank correlation coefficient between elk use days and
cattle use days is -0.94." These statements need to
be translated into a language understood by the nat-
ural resource manager, i.e., "elk are avoiding areas
used by cattle." (The pure logician, blissfully igno-
rant of western cattle operations, could of course
argue that cattle are avoiding elk, but managers are
not too interested in this type of argument. )

What is the cause? Biologists and managers
must be concerned with causes. Knowing that some-
thing happens is useful, but knowing why is even
more useful. Considering the previous example, the
manager may want to know why elk avoid areas
used by cattle. One possibility is that elk simply pre-
fer different types of pastures than those normally
used by livestock operators for cattle grazing. How-
ever, the biologist may have observed that elk use
pastures until cattle move onto them, after which
they avoid them. This is useful information in inter-
preting data, even though it has not been collected
with a well-designed study. The biologist may be-
lieve from his or her observations that elk avoid
pastures because forage conditions are much worse
after cattle have been "highgrading" current forage
production. Furthermore, the biologist may be able
to cite studies in similar situations that indicate such
avoidance is due to forage conditions. Thus informa-
tion from a variety of sources has been used to ex-
plain the observed phenomena.

A common problem in determining causality is
that the results observed are due to some unex-
pected but obvious cause, other than the one a biol-
ogist was trying to isolate. Weather is often such a
factor. For example, a biologist may have been moni-
toring changes in vegetation cover in two similar
streams: one grazed by cattle; the other ungrazed.
During the fifth year, a 100-year storm occurs in the
drainage of the ungrazed stream, causing severe
flooding and scouring of the entire streambed. The
data after 5 years would suggest that grazing im-
proves the condition of the streamside vegetation,
whereas just the opposite is true. A biologist would
be foolish, not to mention unprofessional, to me-
chanically interpret the data without recognizing the
more obvious, but largely undocumented, causes.
In Chapter 35, Weather and Climate, ways of dealing
with variations caused by weather are described.
These methods will help in most cases, but excep-
tions will occur. Record events, such as the 100-year
storm, the record low frost, the wettest spring on
record, are occurring somewhere in the world every
day.

What are the expected impacts of the pro-
posed management? This is, of course, the reason
data have been collected and the crucial question to
be answered. Consider the previous example again.
If elk avoid areas used by cattle because forage con-
ditions have been degraded, what are the expected

impacts of expanding grazing leases? Initially, one
may expect that the expansion will change the pat-
tern of elk use on the area. This may be a concern to
either livestock or wildlife management. Certain
types of livestock management, such as rest rotation
systems, may be doomed to fail because elk will
concentrate in the pastures that are supposed to be
"rested." Or, such management may force elk to
move to other areas where they may be more vul-
nerable to poaching or where they may damage agri-
cultural crops.

The biologist is thus interpreting and extrapolat-
ing the data that have been collected to make infer-
ences that are useful in making management
decisions. In following this process, he or she must
also be evaluating the adequacy of the information.
This leads to another series of questions.

How good are the data? A biologist rarely can
collect the quantity or quality of data required for
risk-free conclusions. For one thing, budgets for such
efforts are usually rather limited. Furthermore, well-
designed studies for evaluating year-to-year variation
typically take 3 to 5 years of data collection; yet,
managers cannot wait 5 years to make decisions.
Thus, the best study designs can rarely be used, even
if a biologist or agency could afford them.

Even the data collected are often inadequate or
questionable. Erroneous data are collected by inex-
perienced, trained, or unmotivated technicians.
Equipment fails or malfunctions occur resulting in
data gaps or data of questionable accuracy. Weather,
travel difficulties, or budget problems prevent cer-
tain data from being collected. Management actions
or "acts of God" destroy or invalidate study sites
or plots. The list goes on, but these are normal
working conditions for the field biologist and should
not be confused with problems caused by poor
study design. The latter can be prevented; the for-
mer cannot. An experienced biologist can lessen
these problems by planning for contingencies, train-
ing personnel thoroughly, carefully testing equip-
ment, and taking other precautions. The problems
can never entirely be prevented, however.

This explains why most data are "soft." Hope-
fully, the data obtained are better than having no
data at all. However, the biologist must determine
the quality of the data and convey this information
to others who use or make decisions based on such
data. Consider the previous elk and cattle example. If
the technician who collected the habitat-use data
could not distinguish elk from cattle tracks, one may
not want to stake his or her reputation on the study-
results.

Study design may also cause problems. Design-
ing studies requires making a series of reasonable
assumptions so certain statistical procedures can be

Evaluation and Interpretation

779

used. Typically, assumptions that seemed reasonable
at the beginning of a study may become very ques-
tionable after more is learned about the habitat or
population being studied. For example, one may
assume that a drainage contains a discrete herd of
mule deer (Odocoileus hemionns) but, after a field
study, discovers two separate subherds: one that is
resident year-round and the other migratory and
found on the area only in winter. A study designed
to determine the amount of forage to be allocated
for deer might produce misleading results if based
on the former assumption.

Thus, the biologist may have only poor data, for
reasons beyond his or her control. Nevertheless,
the manager should be informed of the reliability of
the data. Poor data can cause enough problems, and
problems should not be compounded by pretending
the data are good.

Are the data adequate? I have alluded to the
fact that biologists rarely consider data adequate; yet,
most of the time, they must live with such data.
Sometimes, however, collecting more data is both
necessary and feasible. The biologist is responsible
for advising the manager of such contingencies. As
with initial planning decisions, such mid-course deci-
sions must be made by weighing the time and costs
involved against other priority projects.

To this point, I have talked about the quality of
the data, the meaning of the data, and predicted
impacts. A closely related question is "What is the
significance of such impacts?" Put another way,
"What values will be enhanced or destroyed or are
at risk?" This leads to the subject of evaluation.

EVALUATION

Evaluation is the process of determining the
value of something. Used here, evaluation is the pro-
cess of determining the value of the wildlife
resources.

After data have been carefully collected, ana-
lyzed, and interpreted, the value of the wildlife re-
sources that may be affected needs to be explained.
When questions such as "Who cares if the elk herd
decreases by 50% ?" or "What is 300 acres of riparian
habitat worth?" are being asked, the biologist needs
to place a value on the resource or at least explain
the resource values.

Wildlife biologists have traditionally concen-
trated on enhancing or ensuring the production or
conservation of wildlife resources in the best manner
possible. The criteria they use are primarily biologi-
cal or ecological and include factors such as "ecolog-
ical diversity" or "ecosystem stability" (Krutilla

1974). Most wildlife biologists place high value on
wildlife and have not felt a need to seriously exam-
ine wildlife values (Bailey 1984). Furthermore, many
biologists dislike the concept of placing monetary
values on wildlife, believing the real values of wild-
life are aesthetic, biological and, most importantly,
unmeasurable. However, if biologists argue a case for
wildlife resources against actions that destroy wild-
life habitat, they must be thoroughly familiar with
all wildlife values and some valuation methods
(Steinhoff 1971).

Economists have traditionally placed monetary
values on economically commercial resources (those
resources that can be bought and sold) so the re-
sources can be analyzed. Amenity resources (aes-
thetic, ecological, and other non-commercial
resources), however, are more difficult to place val-
ues on and, therefore, more difficult to analyze for-
mally (Krutilla 1974). Since many wildlife values fall
in the amenity category, economic analysis of wild-
life values is still in its infancy. However, the field
is developing rapidly and is of such importance that
a chapter on the subject is included in this book.
Whereas in this chapter I discuss some types of wild-
life values for the biologist to consider subjectively,
Chapter 40, Economic Analysis, presents the state-of-
the-art for formal quantitative analysis methods. For
further reading on wildlife values, I recommend Bai-
ley (1984:34-50), King (1966), Steinhoff ( 1971 ),
Shaw and Zube ( 1980), or Kellert ( 1980a, b).

King ( 1 966 ) comprehensively categorized wild-
life values. He divided these values, which have sub-
sequently been used by several other authors, into
seven categories:

( 1 ) Commercial;

(2) Recreational;

(3) Biotic;

(4) Scientific, philosophical , and educational;
( 5 ) Aesthetic;

(6) Social; and

( 7 ) Negative.

Of these values, only the commercial value can be
easily translated into dollars for conventional eco-
nomic analyses. Techniques for valuation of recrea-
tion are available and more are under development
(Sorg and Loomis 1985). Numbers 3 through 6 are,
for the most part, unquantifiable and will remain that

780

Evaluation and Interpretation

way for a long time, at least in terms of placing a
dollar value on them. Negative values can occur in
all categories. Damage to orchards by elk is relatively
easy to translate into dollar values; on the other
hand, the negative value of Canada goose droppings
on putting greens is harder to quantify.

Values, however, can be looked at and measured
in different ways. Kellert ( 1980a) measured and
described wildlife values in terms of 10 attitudes
toward animals (Table 1 ). Some of these, such as the
ecologistic and scientistic attitudes, correspond
neatly with King's values, whereas others, moralistic
and dominionistic, do not have any obvious counter-
parts. Kellert ( 1980a) surveyed American public
attitudes toward animals and more specifically wild-
life (Table 2) as one measure of values. The attitudes
are, of course, much more complex than can be
represented in a few simple tables, but biologists
need to be aware of public attitudes (values). Many
biologists, who were trained and spent years in the
profession when wildlife management was primarily
game management, were slow to recognize changing
public attitudes toward wildlife and the changes

they wanted from wildlife managers. U.S. Bureau of
Land Management biologists, for example, differ sig-
nificantly from the general public in the same tests
(Peyton and Langenau 1985). They score signifi-
cantly lower on utilitarian and negativistic scales and
higher on ecological, scientistic, dominionistic, and
naturalistic scales. Biologists must be aware of not
only their own values but also those of the general
public and be able to articulate these values. This is
a difficult, sensitive, but necessary task, for which
there are no easy solutions or cookbook approaches.
Kellert (1980) summarized the importance of the
task:

"The wildlife management field appears to be
confronted by a major change in the public it
serves, with many new and atypical groups be-
coming appropriate recipients of professional
attention. This expanded constituency must
inevitably constitute a threat as much as a chal-
lenge to a field that has historically defined itself
in far narrower terms. Nevertheless, the chal-
lenge represents a rare change, and it would be
a disservice to the profession, let alone to an

Table 1. Classification of attitudes toward animals (from Kellert 1980a).

Attitude

Description

Naturalistic

Ecologistic

Humanistic

Moralistic •

Aesthetic •
Scientistic

Utilitarian •

Dominionistic

Negativistic
Neutralists •

The primary characteristic is a strong interest and affection for outdoors and wildlife. Obser-
vation and personal involvement with wildlife are the keys to the naturalistic interest in the
outdoors. Animals provide the context and meaning for active participation in natural
settings.

Directed at a conceptual understanding of the interrelationships of species in the context of
ecosystems, with major concern for dependencies between animals and their natural habi-
tats. Wild animals are valued less as sources of affection or amusement, and more as de-
vices for comprehending the broader functionings of natural systems.

Emphasizes feelings of strong affection and attachment to individual animals, typically pets.
The animal is the recipient of feelings and emotional projections somewhat analogous to
those expressed toward other people.

Concern for the ethically appropriate humane treatment of animals. A basic moralistic princi-
ple is the fundamental similarity of all animals, each endowed with equivalent rights to
existence.

Emphasizes the attractiveness or symbolic significance of animals.

Concern for the biological and physical characteristics of animals. Wildlife are primarily
interesting as problem-solving objects, not as sources of companionship or wilderness
recreation.

The relevance of animals is largely derived from their usefulness to people. Animals should
serve some human purpose and, thus, be functional as sources of personal gain. Animals
are desirable insofar as they facilitate some type of tangible advantage or reward.

Satisfactions derived from the mastery and control of animals, typically in a sporting context
(e.g., rodeos, trophy hunting, bullfighting). Animals provide opportunities for expressions of
prowess, skill, strength, and masculinity.

Active dislike or fear of animals. This domain includes neutralism — a more passively oriented
avoidance of animals for reasons of indifference.

Primary orientation is a passive avoidance of animals caused by indifference.

Evaluation and Interpretation

781

Table 2. American attitudes toward wildlife (from Kellert 1980b).

Attitude

Estimated % of
American Popula-
tion Strongly
Oriented Toward
the Attitude*

Common Behavioral Expressions

Most Related
Values/Benefits

Naturalistic

10

Outdoor wildlife-related recreation; back-
country use, nature birding, and nature

hunting.

Outdoor recreation

Ecologistic

7

Conservation support, activism and mem-
bership, ecological study.

Ecological

Humanistic

35

Pets, wildlife tourism, casual zoo visitation.

Companionship,
affection

Moralistic

20

Animal welfare support/membership, kind-
ness to animals.

Ethical, existence

Aesthetic

15

Nature appreciation, art, wildlife tourism.

Aesthetic

Scientistic

1

Scientific study/hobbies, collecting.

Scientific

Utilitarian

20

Consumption of furs, raising meat, boun-
ties, meat hunting.

Consumptive,
utilitarian

Dominionistic

3

Animal spectator sports, trophy hunting.

Sporting

Negativistic

2

Cruelty, overt fear behavior.

Little or negative

Neutralists

35

Avoidance of animal behavior.

Little or negative

'Totals more than 100% since persons can be strongly oriented toward more than one attitude.

American public and wildlife resource in need,
if the professional reaction was more to avoid
an alien reality than a creative and bold
response to an evolutionary opportunity."

Clearly the days of simply managing ducks for
duck hunters and deer for deer hunters are gone.
The biologist must consider and explain wildlife
values to managers and others when presenting in-
formation from inventory and monitoring studies.

SUMMARY

Interpretation and evaluation are an important
part of the inventory and monitoring process. Biolo-
gists must explain to the manager and others what
the information means. Explanation, at this stage,
should be based on all sources of information, not
merely on a narrow statistical conclusion. Biologists
must also explain the values of the wildlife resources
under consideration. To do this, biologists should
be aware of the values the public places on wildlife
resources, since these may differ significantly from
their personal values.

782

Evaluation and Interpretation

LITERATURE CITED

BAILEY, J. A. 1984. Principles of wildlife management. John
Wiley and Sons, New York, NY. 373pp.

KELLERT, S.R. 1980a. Americans' attitudes and knowledge
of animals. Trans. North Am. Wildl. Nat. Resour. Conf.
45:1 1 1-124.

. 1980b. Contemporary values of wildlife in Ameri-
can society. Pages 31-60 in Shaw, WW. and EH.
Zube, eds. Wildlife Values. Center for Assessment of
Noncommodity Natural Resource Values, School of
Renewable Natural Resources, Univ. Arizona, Institu-
tional Series Rep. 1. Tucson. 117pp.

KING, R.T. 1966. Wildlife and man. New York Conserva-
tionist 20:8-1 1.

KRUTILLA, J.V. 1974. Methods for estimating the value of
wildlife resources. Pages 125-136 in Bailey, J. A., W.
Elder, and T.D. McKinney, eds. Readings in Wildlife

Conservation. The Wildlife Society, Washington, DC.
722pp.

MERRIAM-WEBSTER, A. 1983. Webster's ninth new Col-
legiate dictionary. Merriam-Webster Inc., Publishers,
Springfield, MA.

PEYTON, R.B. and E.E. LANGENAU, Jr. 1985. A compari-
son of attitudes held by BLM biologists and the gen-
eral public towards animals. Wildl. Soc. Bull. 13:117-
120.

SHAW, WW. and EH. ZUBE, eds. 1980. Wildlife values.
Center for Assessment of Noncommodity Natural
Resource Values, School of Renewable Natural Re-
sources, Univ. Arizona, Institutional Series. Rep. 1.
Tucson. 1 17pp.

SORG, C.F. and J. LOOMIS. 1985. An introduction to wild-
life valuation techniques. Wildl. Soc. Bull. 1338-46.

STEINHOFF, H.W. 1971. Communicating complete wild-
life values of Kenai. Trans. North Am. Wildl. and Nat.
Resour. Conf. 36:428-438.

Evaluation and Interpretation

783

40

ECONOMIC
ANALYSIS

John B. 1 .00 ni is

U.S. Fish and Wildlife Service

Western Energy and land Use Team

Creekside One Building

2627 Redwing Road

Fort Collins, CO 80526-2899

Editor's Note: Many biologists are just beginning to
recognize the need for economic analyses of wild-
life values. In many cases the economic evaluation
of the effects of a project will affect decisions more
than the biological evaluation. Biologists need to
be aware of bio -economic techniques and theory so
that they can plan biological assessments to pro-
vide economists and managers with the informa-
tion they need on economic values of wildlife.

INTRODUCTION

Several new federal directives requiring evalua-
tion of the dollar benefits of wildlife have increased
the need for biologists and economists to understand
the available bio-economic techniques. Improved
techniques for valuation of amenity (non-marketed
natural resources), such as wildlife, have largely dis-
placed older, more subjective (and often erroneous)
approaches. Any reader unfamiliar with the major
advances in the economics of non-marketed natural
resources will be pleasantly surprised with improve-
ments made since the 1970s.

The purposes of this chapter are to describe
concepts and measurement of economic values of
wildlife, to provide an overview of bio-economic
analysis techniques, and to discuss some of the issues
and common misconceptions about economic anal-
yses of wildlife values. More detailed information on
actual bio-economic analysis techniques can be
found in the appendix.

Most federal agencies are required to translate
the biological effects of their actions into economic
values. For example, the U.S. Bureau of Land Manage-
ment's (BLM's) Final Rangeland Improvement Policy
(U.S. Department of the Interior, Bureau of Land
Management 1982) indicates that willingness to pay
values for wildlife need to be developed so they
can be used in rangeland investment (SageRam) pro-
grams that rank alternative livestock management
actions. The U.S. Forest Service planning procedures
require estimates of present net value of all re-
sources having an established market value or an
assigned value (U.S. Department of Agriculture, For-
est Service 1982). Within these resources, wildlife
recreation has a dollar value set by the Chief of the
U.S. Forest Service; this value is initially developed
from existing research on the economic value of
wildlife recreation. The present net value (in dollar
terms ) has become one of the U.S. Forest Service's
key criteria in comparing planning alternatives and
in determining what represents maximum net public
benefits (Peterson 1983).

Current Address: Division of Environmental Studies, University of
California at Davis, Davis, CA 95616.

Economic Analysis

785

WILDLIFE VALUES AND THE PUBLIC TRUST
DOCTRINE

Much confusion exists about the economic
value of wildlife. In this section, the types of wildlife
values and their relationship to the public trust doc-
trine are discussed.

Economic value is often confused with financial
value. For a good or service to have economic value,
it must meet the following two criteria:

( 1 ) It must provide some consumers satisfaction
or enjoyment, and

(2) The good or service must be scarce in the
sense that consumers want more than is avail-
able at no cost.

Wildlife certainly meets both of these criteria.

Economic values include commercial, recrea-
tional, option, existence, and bequest values (Figure
1 ). Besides the economic values of on-site recreation
(both consumptive and non-consumptive) and com-
mercial uses of wildlife, there are many off-site user
values. These include option, existence, and bequest
value. Option value refers to an individual's willing-
ness to pay to maintain wildlife recreation opportun-
ities. Option value can be thought of as an insurance
premium people would pay to ensure wildlife recre-
ation opportunities in the future. Existence value is
the economic benefit received from simply knowing
wildlife exists. Bequest value is the willingness to
pay for providing wildlife resources to future
generations.

Figure 1. Economic values of wildlife.
786 Economic Analysis

The first pitfall one faces in identifying and
quantifying wildlife values concerns the difference
between economic values and financial values. Finan-
cial values are defined as actual revenue or sales
received by firms or public agencies (i.e., cash
changing hands). Financial values ignore social bene-
fits or costs that cannot be captured as revenue by
firms. At best, financial values are a subset of eco-
nomic values (see Figure 1 ) and, at worst, may be a
serious distortion.

The disparity between economic and financial
values for wildlife is confusing to resource managers
and local officials accustomed to dealing only with
marketed resources such as coal or timber. For these
resources, it is customary to assume the economic
and financial values of additional units of outputs are
almost synonymous.

Financial values (sales revenue, profit) are useful
in answering questions about profitability of guide
services or retail outlets for recreational equipment.
Financial feasibility of business is often important
to county and state officials from the standpoint of
job creation or property taxes. These legitimate con-
cerns have little to do with the economic value of
wildlife, however.

In the U.S., laws assign ownership of the nation's
wildlife resources to the state or federal government.
The governments thus have the responsibility to
manage these resources as trustees for the benefit of
the public. Public ownership implicitly recognizes
that inefficient resource allocation would result with-
out cooperative intervention. The public agent is
expected to incorporate all benefits and costs into
decisions and not merely the financial costs and re-
turns seen by private firms. This is called the Public
Trust Doctrine.

The Public Trust Doctrine recognizes that all
benefits and costs of wildlife to people are not cap-
tured in the commodity values of private markets.
Therefore, the public agencies are responsible for
protecting the interests of birders, hunters, and citi-
zens who enjoy knowing wildlife exists.

Although public agencies are charged with pro-
tecting all wildlife values, current methodology for
quantifying such values is often limited to recrea-
tional and commercial values.

MEASUREMENT OF ECONOMIC VALUES OF
WILDLIFE

As discussed earlier, the economic values of
wildlife and wildlife recreation differ from financial
values. Even in determining economic values of wild-
life, some confusion still exists. Before techniques

for measuring economic value are discussed, it is
necessary to have an operational definition of just
what needs to be measured.

In the marketplace framework of supply and
demand, the supply curve defines the quantities of a
good that producers are willing to market in return
for various payments. A supply curve has a positive
slope because, at a higher price per unit, producers
are willing to supply more goods. The demand curve
is generated by the quantities of a good that con-
sumers are willing to buy when faced with various
prices. Higher prices will lead to a lower quantity
demand, causing a negatively sloped demand curve.
The intersection of the supply and demand curves
indicates equilibrium — which is reached at a price
where the quantity that producers are willing to
supply equals the quantity consumers are willing to
buy (Figure 2).

60t

QUANTITY OF X

Figure 2. Supply and demand for good X.

In the hypothetical example in Figure 2, equilib-
rium is reached at a price of $30 and supply of good
X equals demand. The demand curve shows that a
consumer is willing and able to pay $60 for the first
unit of good X. When faced with a price of $30, a
consumer surplus of $30 results on the first unit.
Alternatively, consider the situation where X is free.
In this case, eight units of X would be demanded
and the entire area under the demand curve would
be a measure of consumer surplus with no revenue
collected. From this hypothetical example, one can
see that the price charged is a measure of expendi-
tures or financial value while the difference between
what the individual is willing to pay and expendi-
tures is a measure of net benefits (consumer surplus
or net willingness to pay).

Although the same concepts of supply and de-
mand and consumer surplus apply to all goods
whether they are marketed or not (as wildlife), em-
pirical estimation may be difficult for unmarketed
goods. Supply of a species is regulated by wildlife
managers and the carrying capacity of the land.
When carrying capacity has not been reached, the
management costs of supplying various quantities of
a species can be used to generate a supply curve.
Estimating a demand curve is less straightforward. In
the case of a market good, a demand curve is gener-
ated by monitoring purchasing behavior of individu-
als at various prices and mapping a demand curve.
However, in the case of wildlife-related recreation, a
demand curve cannot be derived easily because
wildlife ordinarily is considered a publicly owned
good that one does not have to pay to hunt or view.
Hunting and fishing licenses are not market-deter-
mined prices but administratively determined fees.
As a result, prices cannot be associated with various
quantities, and a demand curve cannot be plotted
directly.

For wildlife values, the demand curve is used as
a basis for estimating consumers' (or in this situa-
tion, recreationists' ) willingness to pay for increases
in wildlife recreation opportunities. (Harberger
1971; Sassone and Schaffer 1978). A hypothetical
demand curve (Figure 3) shows the number of trips
a recreationist would take at different travel costs.
The total area under the curve is the gross willing-
ness to pay. The net willingness to pay is the area
above the actual cost but below the demand curve.

60 -

Eb 50 -

DC
1-

OC

UJ 40 -

a.

(/)
OC

< 30 -

_l

-1

o

Q

20 -

Actual cost @-

yyyyyyyyy/yyyyyfa^QmaV[^ curve

IN.

1 Nl

1 ^\

1 \

0 1 2 3 (V) 5 6

TRIPS PER YEAR

Figure 3. Hypothetical demand curve for wildlife
recreation.

Economic Analysis

787

If the actual cost was $10 then the shaded area in
Figure 3 is the net willingness to pay which is $105.
The gross willingness to pay would be $105 plus
the $40 (4 trips x $10 per trip) which comes to
$145.

Economists term consumers' net willingness to
pay as "consumer surplus." Consumer surplus repre-
sents the consumers' additional willingness to pay
for the opportunity to hunt or fish at some site. It is
a net or additional willingness to pay because it is
added to current expenditures.

Consumer surplus sometimes stands as a con-
ceptual "stumbling block" for those involved in eco-
nomic valuation. Consumer surplus is often not seen
as a real or tangible economic benefit because it
represents money that has not actually been col-
lected by a business or government agency. If indi-
viduals could be feasibly charged a price equal to
their maximum willingness to pay for each unit, the
estimate of consumer surplus could be verified. Such
pricing schemes exist, but the government does not
use them to capture the full willingness to pay for
each unit (the consumer surplus) as revenue.

Several methods are used to derive demand
curves indirectly and provide measures of net will-
ingness to pay. Those most often used include the
travel cost method and the contingent value method.

Travel Cost Method (TCM)

This approach was developed by Clawson
(1959) to empirically estimate recreation benefits.
The method is based on the belief that travel cost
can be used as a proxy for price in deriving a de-
mand curve for a recreation site. The first step of
zonal TCM analysis involves dividing the area around
the site into zones or counties of origin of recrea-
tionists. One assumes that the costs from a particular
zone to a recreation site are the same for all individ-
uals in that zone. Based on origin data, a visitation
rate (trips per capita) is calculated for each zone.
Regression analysis can be used to estimate a func-
tion for visitation rates based on travel cost and socio-
economic data. Such a regression line is shown in
Figure 4.

The different travel costs for making a trip from
each of four origins (counties) surrounding a site are
plotted against the number of trips per capita from
each county to the site. These different combinations
of travel cost and trips per capita represent price-
quantity points that trace a demand curve. From this
demand curve, the consumer surplus or net willing-
ness to pay for recreation at a particular site can
be calculated by taking the area under the demand
curve but above the travel cost for each zone of
origin. For details on use of the technique, see
Rosenthal et al. (1984).

NUMBER OF TRIPS PER CAPITA TO SITE 1

S1 represents supply or travel costs

Figure 4. Regression of travel costs versus number
of visits per capita.

The TCM requires data on wildlife recreation-
ists' travel cost to or distance from a recreation site.
If surveys providing travel cost (or distance) and
city or county of residence are available, they should
be used. However, one advantage of the TCM is that
existing information from hunting licenses, game
tags, or even license plates can be used to determine
the wildlife recreationist's city or county of resi-
dence. If one knows the county or city of residence,
the round-trip distance to the fishing or hunting
site can be calculated from maps. The distance can
be converted to a travel cost by using the govern-
ment publication "Cost of Owning and Operating a
Motor Vehicle" (U.S. Department of Transportation
1985).

The information on visits needs to be grouped
by county so the trips per capita can be calculated
by dividing trips by county population. This be-
comes a dependent variable in the regression analy-
sis used to statistically estimate the demand curve.
Thus, it is generally important to know the county
or city (or zip code) of wildlife users. Knowing the
number of wildlife recreationists per vehicle and
number of days per trip is also useful.

Data sources are often available for calculating
travel costs. For example, several data sources exist
for big game hunting. State wildlife management
agencies often record the zip code or city of resi-
dence from hunting licenses or post-season harvest
surveys for the general big game seasons. Wildlife
agencies also record the herd unit from which an an-
imal is harvested. These provide the two basic types
of data for TCM: location of residence and herd unit.

788

Economic Analysis

Matching these two together allows calculation of
round-trip distance and estimation of round-trip
travel costs. For details on these procedures see
Loomis(1982).

In many states, license plate numbers are keyed
to county of residence. By simply visiting the recrea-
tion site of interest and recording license plate num-
bers (in some States, the county name is even
printed on the plate), one can get an approximate
idea of the origin of wildlife recreationists. This is
useful in collecting data on nonconsumptive users of
wildlife, such as birding or photography.

The TCM can be used to measure the economic
value of nonconsumptive wildlife recreation such as
birdwatching or wildlife photography. Data on dis-
tance traveled can be obtained on site at areas
where a significant amount of nonconsumptive use
of wildlife takes place.

A useful characteristic of TCM is that it can be
used for valuation and use estimation simultaneously.
A simple single site TCM demand curve can be used
to estimate hunter demand over time. This is done
by multiplying each county's current trip per capita
rate times future county population.

Thus, if a demand estimate is needed for the
year 1990 for example, all one must do

( 1 ) Use the equation to estimate current trips per
capita at the current travel cost for each
county, and

(2) Multiply this current trip per capita by each
county's forecasted population in 1990.

When using a regional TCM that includes a qual-
ity variable such as harvest, one can even estimate a
future visit rate per capita that varies with future
harvest. To get a future demand estimate, the analyst
would multiply each county's future trip per capita
figure by each county's future population.

The strengths and weaknesses of TCM as a tool
for estimating benefits of recreation are described by
Dwyer et al. ( 1977) and Rosenthal et al. ( 1984).

Contingent Value Method (CVM)

The Contingent Value Method (CVM) is also
known as the "Direct Method" or "Bidding Method."
Unlike the familiar market situation where people
alter consumption in response to price changes,
bidding involves asking individuals to respond to
changes in hypothetical prices for an unmarketable
good. The term "contingent valuation" stems from
asking individuals how their behavior would change,

contingent on a new hypothetical situation. Thus
individual users are asked to provide data to estimate
their net willingness to pay (consumer surplus).

Estimation (distinguished from application) re-
quires administration of a carefully constructed sur-
vey. The U.S. Water Resources Council (1979) and
Dwyer et al. (1977) provide good reviews of the
steps in survey design and implementation. Although
wildlife biologists at the field level will rarely con-
struct such a survey, they should understand how
the method works and what the information means.

The CVM relies on mail and personal or tele-
phone interviews. The interview begins with an in-
troduction and a full explanation of the purpose. The
introduction is often aided by the use of maps,
graphs, or photographs in an attempt to define the
good being valued. Next, participants are asked ( 1 )
how much they would be willing to pay to achieve
an improved situation, (2) how much compensation
they would require to accept a reduction from their
current situation, or ( 3 ) how much they would be
willing to pay for current circumstances. Depending
on how the questions are phrased, the resulting bids
represent either a willingness-to-pay or willingness-
to-sell measure of consumer surplus (see Brookshire
et al. 1980).

The iterative bid procedure may be necessary to
obtain a complete measure of maximum willingness
to pay. In this process, a recreationist reacts to a
starting bid and is successively asked if he or she
would pay higher incremental amounts until a nega-
tive response is obtained. The last positive response
is considered the measure of maximum willingness
to pay. The bidding process forces individuals to
analyze their preferences more fully, providing a
more accurate measure of consumer surplus.

Values of trout fishing and deer hunting in 1 1
western states derived with CVM are shown in Table
1 . These dollar values per day resulted from an itera-
tive bidding sequence performed in a personal sur-
vey and represent net willingness to pay over and
above current costs (in 1980 dollars).

Other sources of CVM derived values include
articles appearing in such journals as Land Econom-
ics, American Journal of Agricultural Economics,
Western Journal of Agricultural Economics, and
Journal of Leisure Research. PhD dissertations and
Masters theses prepared in the economics depart-
ment of the University of Wyoming, University of
New Mexico, Utah State University, and University of
Washington often empirically estimate dollar values
for wildlife recreation in western states by using the
CVM. The Transactions of the North American
Wildlife Conference often contains economic values
derived by CVM.

Economic Analysis

789

Table 1. Value (net willingness to pay) for trout fishing and deer hunting (U.S. Department of the Interior,
Fish and Wildlife Service and U.S. Department of Commerce, Bureau of Census 1982).

Dollars per Day of Trout Fishing

Dollars per Day of Deer Hunting

Arizona

Value2

Mean3

Value2

Mean3

19.54

±4.19

32.50

±4.95

California

20.53

±2.08

37.35

±7.87

Colorado

16.16

±1.95

23.49

±4.18

Idaho

12.93

±0.93

28.77

±2.63

Montana

16.47

±1.88

25.42

±2.43

Nevada

12.35

±1.48

29.02

±3.92

New Mexico

15.70

±1.46

29.11

±2.60

Oregon

13.49

±1.82

21.44

±3.06

Utah

12.57

±1.17

25.72

±2.69

Washington

14.03

±2.54

24.18

±4.26

Wyoming

16.87

±1.54

36.26

±3.26

'Values derived using Contingent Value Method; iterative bidding was used.

2Values are expressed in 1980 dollars.

3Dollars are expressed as mean, plus or minus standard error.

Experience with CVM indicates that it can reli-
ably measure values if the survey is carefully con-
structed and carried out. However, studies generally
indicate that CVM estimates of net willingness to pay
are conservative estimates of value (Bishop and He-
berlein 1979; Brookshire et al. 1982). This conclu-
sion is based on comparisons of actual market data
to the Contingent Valuation bids.

The CVM is particularly well-suited for estimat-
ing the value of nonconsumptive wildlife recreation
as well as option, existence, and bequest values.

Persons having little or no experience with
CVM should seek assistance from experienced users
of the method. Data from studies that employed
the method may also be useful, since they are ex-
pressed in units, such as dollars per hunter day.

Advantages of Valuation Methodologies

Two techniques for measuring the net willing-
ness to pay for wildlife recreation have been dis-
cussed— Travel Cost Method and Contingent Value
Method. Each method has its strengths and weak-
nesses in valuing wildlife recreation. TCM has two
advantages:

( 1 ) It has the capability to use existing data from
hunter applications, licenses, or other sources

for statistically estimating a demand curve
for the activity at a particular site.

(2) It can be used for both a benefit estimate and
use projection (demand projection) over
time from the same model framework.

The advantage of CVM is its ability to look di-
rectly at the dollar value of improvements in harvest
or quality of recreation (for example, the increase
in value for higher success rates or for trophy
animals ).

There are several advantages to empirical tech-
niques such as the TCM and CVM. One advantage
is that these methods provide objective measures of
the benefits of wildlife recreation as revealed by
the recreationists themselves. This is a significant
improvement over the subjective methods whereby
specialists assign the value based on the benefits they
think the recreationists receive.

Another advantage of TCM or CVM is that confi-
dence intervals around the benefit estimates can be
displayed to give the decisionmaker some indication
of the precision of the benefit estimate. In this way,
sensitivity analysis can be performed and determina-
tions made as to whether more precise estimates
of wildlife values are critical to ranking of
alternatives.

790

Economic Analysis

BIOLOGICAL-RECREATIONAL-ECONOMIC
RELATIONSHIPS: OVERVIEW

Many methods of predicting bio-economic re-
sponse require estimating how a change in manage-
ment will cause change in population numbers and
thus in recreation (hunter or angler) days. Values
from CVM can then be used to attach a dollar
amount to a recreation day. Five procedures for bio-
economic analysis are listed below:

(4) "Short Form for Bio-economic Evaluations of
Wildlife in Washington State" (Oliver et al.
1975).

(5) "Human Use and Economic Evaluation"
(HUEE) system (U.S. Department of the Inte-
rior, Fish and Wildlife Service 1980a).

These procedures are described here briefly;
two, Hunter Day Short Form and HUEE, are docu-
mented in more detail in Appendixes I and II.

( 1 ) "Hunter Day Short Form" developed by the
U.S. Bureau of Land Management's Moab Dis-
trict Office (U.S. Department of the Interior,
Bureau of Land Management 1983). This
procedure has been developed for big game
hunting only.

(2) Suislaw National Forest's Salmonoid Fisheries
Model (Kunkel andjanik 1976).

(3) "Biological Response Approach" developed
by the U.S. Bureau of Land Management, Ore-
gon State Office (U.S. Department of the Inte-
rior, Bureau of Land Management 1979).

All of these procedures share certain similarities
(Figure 5). In particular, all five approaches require
that harvest (or population) figures be known or
estimable (Figure 5).

In the State of Washington's "Bio-Economic
Short Form" (Oliver et al. 1975) and the U.S. Forest
Service's anadromous fish model (Kunkel andjanik
1976), biological effects must be estimated in terms
of effects on populations. If population simulation
models are available, the linkage of resource deci-
sions to populations may be possible.

Although HUEE analysis (U.S. Department of the
Interior, Fish and Wildlife Service 1980a) and the

USFWS

HUEE

(5)

Recreation Total

days = recreation

Animal days

$
Day

Total

recreational

value

BLM's

(1) Moab District
Short Form

(2) Oregon State
Office

USFS

(3) Salmonid
Fisheries
Model

State of Washington

(4) Bio-Economic
Short Form

Figure 5. Overview of bio-economic approaches.

Economic Analysis

791

Oregon Biological Response approach (U.S. Depart-
ment of the Interior, Bureau of Land Management
1979) can model the biological impacts in terms of
population or harvest, both have the capability to
convert changes in habitat variables (e.g., food,
cover, reproduction) into changes in carrying capac-
ity. Changes in carrying capacity can often be con-
verted to changes in population or harvest. Since
changes in habitat variables due to management ac-
tions are often easier to predict, this habitat-based
evaluation may be useful in evaluating Habitat Man-
agement Plans (HMPs) or Allotment Management
Plans (AMPs).

The next step required in all five approaches is
the estimation of wildlife recreation use levels asso-
ciated with a given number of animals (or fish) avail-
able for harvest. Typically, this involves multiplying
hunter days or angler days per animal harvested
by the change in animals available for harvest.

Once the change in days of wildlife recreation is
known, these days are often multiplied by an average
dollar value per day. This yields an estimate of the
economic effects of the change in wildlife recreation
associated with some original biological change, not
the total economic effects to society as a whole. It
does not include option, bequest, or existence val-

ues. These values can be quantified, but at a cost that
often makes it impractical for routine analyses of
HMPs, AMPs, etc.

The five methods provide a range of analysis
techniques varying from simple to complex. Thus it
cannot be said that one approach is better than the
other, only that one approach may be more cost-
effective for screening possible allotment manage-
ment plans (AMPs) or habitat management plans
(HMPs), whereas another, better suited for in-depth
analysis of the remaining (after screening) candidate
AMPs or HMPs.

One bit of inventory data is common to all of
these basic approaches: determining the wildlife
recreationists' days dependent on or produced from
the wildlife unit under study. This concern is impor-
tant because very few animals may actually be har-
vested in the area under study. For migratory
animals, big game animals, and fish species, each
piece of habitat contributes something to producing
a harvestable animal. For such migratory animals,
determining the equivalent number of animals de-
pendent on a habitat area can be fairly complex.
A step-by-step procedure for doing this is shown in
Table 2.

Table 2. A step-by-step procedure for calculating equivalent number of animals harvested that are dependent
on a habitat area.

Section 1— Animals harvested in the study area.

Enter animals harvested in study area

Multiply line 1 by the percentage that are year-round residents

Number of animals harvested on site that are migrants; subtract line 3
from line 1

Multiply line 4 by the dependency of these migrant animals on the study
area

Subtotal of animals harvested on study area; add lines 3 and 6

line 1

line 2

line 3 (answer)

line 4

line 5

line 6 (answer)

line 7

Section II — Animals harvested outside the study area but dependent on the study area.

Enter animals harvested in state wildlife management area or herd unit
adjacent to or surrounding study area

Enter percentage of these animals that spend time on the study area but
are harvested elsewhere

Multiply the percentage in line 2 by line 1 and enter on line 3

Multiply line 3 by the importance of the study area's habitat to the species
(can use time in study area)

line 1
line 2

line 3

line 4

line 5 (answer)

Section III— The total equivalent number of animals harvested that are dependent on the study area is
found by adding line 7 from Section I and line 5 from Section II. Total

792

Economic Analysis

The first step is to determine the number of
animals of a particular species that are harvested in
the study area. This can often be determined by
looking at state wildlife management agencies' har-
vest data. The population or harvest in the wildlife
management unit may need to be prorated to one or
more specific allotments. Although an acre-per-acre
proration is the simplest, use of information on pop-
ulation concentration will allow a more accurate
proration. Details of the Moab District Short Form
and HUEE procedures are described in Appendixes I
and II, with additional references provided there.

DISCUSSION

Wildlife biologists and managers are often con-
fused about the economic value of wildlife because
they generally do not have enough knowledge of
basic economic concepts. This is aggravated by a
common failure to identify and separate different
policy questions that require different kinds of dollar
value-related answers. The information required by
questions of national economic efficiency, for exam-
ple, are different from that needed to address con-
cerns about local economic impact.

Three issues cause confusion:

means the efficiency of resource use has been in-
creased by reallocating resources from lower value
uses to higher value uses, which is economically
efficient.

The Resources Planning Act ( RPA ), National
Forest Management Act (NFMA), BLM's Rangeland
Investment Policy, and the U.S. Water Resources
Council Principles and Guidelines all require re-
source valuation in terms of net economic surplus to
the consumers or producers. This net surplus is the
value remaining after all costs have been paid, and is
the net willingness to pay. This net economic value
is not measured by the actual expenditures or costs
of the consumer. Expenditure information is useful
for certain kinds of policy decisions requiring knowl-
edge of community dependency, as opposed to net
benefit and economic efficiency. It is also useful for
analyzing economic impacts where the purpose is to
expose the distribution of costs and benefits.

Use of actual expenditure information to meas-
ure wildlife benefits for efficiency-related decisions is
incorrect and misleading. The price actually paid is
an expenditure or cost incurred by the purchaser.
This price or expenditure is received as sales reve-
nue or wages by the provider of the good to the
purchaser.

( 1 ) The distinction between economic efficiency
values and expenditures;

(2) The relationship of price to consumer sur-
plus; and

(3) The difference between net benefits and re-
source quality.

These are discussed here; more detail on the subject
can be found in Loomis et al. (1984).

Economic Efficiency Values Versus
Expenditures

Many of the questions posed by federal and
state wildlife programs involve determining whether
the economic gain from some investment, such as
fish ladders or habitat developments, exceeds the
costs of such developments. A similar question is
asked when evaluating National Forest plans or the
cost-effectiveness of mitigation plans. To determine
whether the benefits exceed the cost of some re-
source action, the willingness to pay (consumer and
producer surplus ) of project gainers needs to be
compared with the willingness to pay of the losers
(Dwyer et al. 1977; U.S. Water Resources Council
1979, 1983; Walsh 1986 ). When the willingness-to-
pay values of project gainers exceed willingness-
to-pay values of losers, the present net value is posi-
tive or the benefit-cost ratio is greater than 1. This

Costs are defined as benefits foregone. The more
it costs society to harvest a certain number of trees,
the less the net gain to society. That is, the more
one gives up to get something, the less net benefit
there is to having it. A timber sale will often require
building several miles of expensive road which will
result in thousands of dollars of expenditures.
However, if the value of trees is less than the ex-
penditures, there has been a net loss to society. Ex-
penditures in excess of economic benefits means the
cost of what was given up exceeded the benefits of
what was received.

An example of beneficial treatment wildlife gets
when the net benefits (gross benefits minus expendi-
tures) of agricultural development are compared
with habitat preservation can be seen in a paper on
the Birds of Prey Conservation Area in southern
Idaho (Hyde et al. 1982). Hyde et al. evaluated a
trade-off between agricultural development and pres-
ervation of the prey base. The net benefits of agricul-
tural development were very low because of the
high costs needed for farmers to cultivate the land
and pump water from the Snake River. If one judged
the economic benefits on an expenditures basis,
the inefficient agricultural development would look
valuable. However, given high costs, there are the
low net benefits of agriculture. Hence, the opportu-
nity cost of maintaining the prey base for the Birds
of Prey Conservation Area was low and the net bene-
fits were greater for the habitat preservation option.

Economic Analysis

793

There are legitimate uses of expenditure infor-
mation, particularly measuring the local income and
employment effects of a management action. Like the
process of economic growth itself, changes in re-
source allocations will displace some job skills (and
workers having only those job skills) and create
demands for workers with different job skills. The
National Environmental Policy Act and local public
officials often require federal agencies to display the
types of workers and industries that will be posi-
tively or negatively affected by resource actions
taken on public lands.

Price Versus Consumer Surplus

Another obstacle to correct wildlife valuation is
the common allegation that market prices and con-
sumers' surplus cannot be compared. This problem
stems from confusion about underlying concepts and
failure to separate different questions requiring dif-
ferent answers. The correct measure of gross value
to the consumer is the amount of monies the con-
sumer is willing to pay for the goods in question.
Net value to the consumer is this sum minus the ex-
penditures required to obtain it, i.e., consumers'
surplus.

The technical relationship between price and
consumers' surplus is shown in Figure 6. Again, the
demand curve is the measurement of consumers'
valuation (Harberger 1971 ). The actual shape of the
demand function will depend on the nature of the
market and the context of the question. In Figure 6,
a downward-sloping function represents the entire
market or industry. For the good in question, let the
current quantity being exchanged be 100 units at a
price of $10. The gross value of one additional unit
in this market area is $10. This is the consumers'
gross willingness to pay for that unit. The price or
cost to the consumer is also $10, so the net willing-
ness to pay is zero. The marginal unit has no con-
sumer surplus.

If the quantity supplied is increased from 100
units to 150 units, the old price of $10 cannot be
used as the consumers' gross willingness to pay for
the 50 additional units. The gross willingness to pay
for these 50 additional units is the area under the
demand curve between 100 and 150 units, in this in-
stance, about $375. At the new level of consump-
tion, the 1 50 units can all be purchased for $5 each.
The gross willingness to pay for the last unit con-
sumed is thus the market price of $5 and the net
willingness to pay for this last unit is zero.

The willingness to pay for the other 49 added
units is greater than $5, however, as shown by the
downward slope of the demand curve, and thus gen-
erates consumers' surplus. For example, the 100th
unit consumed could have been sold for $10, but it
has been sold for $5. The net willingness to pay

50 100 101 150

QUANTITY/YEAR

Figure 6. Relationship between price and
consumer surplus.

enjoyed by the consumer for this unit is thus $5 in
the form of consumers' surplus. The additional con-
sumer surplus associated with the increase of 50
units is the triangular area below the demand curve
between 100 and 150 units and above the price of
$5, or about $125. When the added consumer sur-
plus from the decline in price from $10 to $5 on the
original 100 units is added to the $125, the total
increase in consumer surplus is $625. This is the net
economic benefit of the change whether the units
are bicycles, trees, or hunter days. The differences
arise because some questions ask for net willingness
to pay at the margin (e.g., a small change from 100
to 101 ), and some questions ask for net willingness
to pay for non-marginal changes (e.g., a very large
change from 100 to 150). In any case, if there is
surplus created by a quantity change, it should be
measured and counted.

Hunting and fishing often take place in small or
localized markets. Most hunters and anglers visit
areas within 200 miles of their homes. Removing
fishing opportunities at one major cold-water stream
or lake could make a substantial change in the price
(i.e., travel cost) of fishing opportunities. In such a
situation, there will be significant consumer surplus
associated with eliminating or adding an opportunity.
On the other hand, imagine a place in Minnesota
where there are hundreds of identical uncongested
lakes, each no more than a mile from another. The
loss of one lake would have no measurable effect on
the price faced by an angler and hence no loss in
consumer surplus.

794

Economic Analysis

Net Benefits Versus Resource Quality

Another source of confusion relates to differ-
ences in the value of resources of different quality.
The gross willingness to pay for steelhead fishing, for
example, is generally higher than for trout fishing.
However, because steelhead fishing opportunities are
often available only at a few remote rivers, travel
costs for steelhead fishing tend to be higher than for
trout fishing. The combined effect is that net willing-
ness to pay for steelhead fishing may be lower than
that for trout fishing. In these situations, lower net
benefit flows from the higher quality resource be-
cause the higher price offsets the quality difference.
On the other hand, introduction of a new steelhead
site that requires travel costs similar to those at trout
fishing sites will produce more net benefit than a
new or existing trout site.

A parallel situation exists for timber. For exam-
ple, walnut trees are much more valuable than pine
trees, but a remote stand of walnut trees in an inac-
cessible location will have a lower bid price than a
stand of pine trees near a mill. In this example, the
higher quality trees are worth less because of higher
harvest costs.

SUMMARY AND CONCLUSIONS

As is evident from this chapter, several bio-eco-
nomic analysis systems provide the capability to
translate the benefits of habitat improvements into
dollar terms. Besides fulfilling legal requirements to

make such conversions, the benefits can be com-
pared to the costs of making a given habitat
improvement.

Because the traditional valuation techniques
only measure recreation, recreation benefits cannot
appropriately be interpreted as the total value of
animals or their habitat. Rather, net benefits (present
net worth or internal rate of return) provide infor-
mation about the willingness to pay of recreationists
for the proposed habitat improvements. If the recrea-
tion benefits alone exceed the cost, the project will
usually increase national economic well-being. If the
benefits are less than the cost, the decisionmaker
must ask whether the project substantially increases
intangible economic efficiency benefits (option, ex-
istence, bequest values for a wildlife species of high
public interest) or substantially improves equity.
Thus, the internal rate of return or benefit-cost ratio
does not by itself make the decision to implement or
not implement a management action. The internal
rate of return or benefit-cost ratios tell the decision-
maker and the public taxpayer what the economic
efficiency of such investments are.

There may be other legitimate social objectives
that outweigh economic efficiency in determining
whether the investment is to be made. These other
objectives should be documented in the decision
process regardless of what the economic efficiency
analysis shows. In this way, bio-economic analysis
and economic analysis in general contribute to im-
proved and informed decisions rather than binding
the decisionmaker's hands.

Economic Analysis

795

LITERATURE CITED

ALLEN, A. and M. ARMBRUSTER. 1982. Preliminary evalu-
ation of a habitat suitability for the pronghorn in
McKenzie, J. ed., Proc. 10th Pronghorn Antelope
Workshop.

BISHOP, R. and T. HEBERLEIN. 1979. Measuring values of
extra market goods: Are indirect measures biased?
Am. J. Agric. Economics 6l(5):926-930.

BROOKSHIRE, D., A. RANDALL, and J. STOLE 1980. Valu-
ing increments and decrements in natural resource
service flows. Am. J. Agric. Economics 63(3)165-177.

, M. THAYER, W. SCHULZE, and R. D'ARGE. 1982.

Valuing public goods: A comparison of survey and
hedonic approaches. Am. Economic Review
72:165-177.

BROWN, W., C. SORHUS, B. CHOU-YANG, and J. RICH-
ARDS. 1983. Using individual observations to estimate
recreation demand functions: A caution. Am. J. Agric.
Economics 65( 1 ): 1 55- 1 57.

CLAWSON, M. 1959- Methods of measuring the demand
for and value of outdoor recreation. Resources for the
Future. Washington, DC.

DWYER, J.J. KELLY, and M. BOWES. 1977. Improved

procedures for valuation of the contribution of recre-
ation to national economic development. Res. Rep.
1 28, Water Resources Center, Univ. Illinois at Urbana.

HARBERGER, A. 1971. Three basic postulates for applied
welfare economics: An interpretive essay. J. Economic
Literature 9(3):785-797.

HYDE, W., A. DICKERMAN, and D. STONE. 1982. Devel-
opment versus preservation in the Snake River Birds
of Prey Conservation Area. Am. J. Agric. Economics
64:756-760.

JUST, R„ D. HUETH, and A. SCHMITZ. 1982. Applied wel-
fare economics and public policy. Prentice Hall, En-
glewood Cliffs, NJ.

KRUTILLA, J.V. 1967. Conservation reconsidered. Am.
Economic Review 47:777-786.

KUNKEL, C. and PJANIK. 1976. An economic evaluation
of salmonid fisheries attributable to Suislaw National
Forest. Suislaw National Forest, Corvallis, OR.

LOOMIS, J. 1982. Use of travel cost models for evaluating
lottery-rationed recreation: Application to big game
hunting. J. Leisure Res. 1 4( 2 ):1 17-124.

and R. OLSON. 1981. Recreational and commercial

activities involving wildlife potentially affected by
the Lower Gunnison River Project. U.S. Dep. Inter.,
Fish and Wildl. Serv., Fort Collins, CO. (unpubl. rep.)
-, G. PETERSON, and S. SORG 1984. A field guide

to wildlife economic analysis. Trans. North Am. Wildl.
Nat. Resour. Conf. 49:315-324.
MATULICK, S., J. HANSON, I. LINES, and A. FARMER.

1982. HEP as a planning tool: An application to wa-
terfowl enhancement. Sci. Pap. 6163, Agric. Res. Cen-
ter, College of Agriculture, Washington State
I Diversity. Pullman.

MISHAN, E. 1976. Cost-benefit analysis. Praeger Publishers,
New York, NY.

OLIVER, W., C. YOUNG, and D. ELDRED. 1975. A short
form for bio-economic evaluations of wildlife in
Washington State. Bull. 7. Washington Game Dep.
Olympia.

PETERSON, M. 1983. Response to Assistant Secretary
Crowel regarding land and resource management
planning. U.S. Dep. Agric, For. Serv. Memorandum
dated May 17, 1983.

ROSENTHAL, D.J. LOOMIS, and G. PETERSON. 1984.

The travel cost model: Concepts and applications. U.S.
Dep. Agric, For. Serv. Gen Tech. Rep. RM-109- Fort
Collins, CO.

SASSONE, P. and W. SCHAFFER. 1978. Cost-benefit analy-
sis: A handbook. Academic Press, New York, NY.

U.S. DEPARTMENT OF AGRICULTURE, FOREST SER

VICE. 1982. National Forest System land and resource
management planning in Federal Register, September
30, 1982, Vol. 47(190):43,026-43,052.

U.S. DEPARTMENT OF THE INTERIOR, BUREAU OF
LAND MANAGEMENT. 1979. Biological response ap-
proach. U.S. Dep. Inter., Bur. Land Manage., Oregon
State Office, (unpubl. rep.).

. 1982. Final rangeland improvement policy. In-
struction Memorandum 83-27. October 15, 1982.
1983 Rangeland investment analysis Bull. UT-060-

83-B6, Moab District Office, UT.

U.S. DEPARTMENT OF THE INTERIOR, FISH and WILD-
LIFE SERVICE. 1980a. Human use and economic
evaluation (HUEE). U.S. Dep. Inter., Fish and Wildl.
Serv., Div. Ecol. Serv. Manual 104: Washington, DC.

. 1980b. Habitat evaluation procedures (HEP). U.S.

Dep. Inter., Fish and Wildl. Serv., Div. Ecol. Serv.
Manual 102. Washington, DC.

. 1981. Standards for the development of habitat

suitability index models. U.S. Dep. Inter., Fish and
Wildl. Serv., Div. Ecol. Serv. Manual 103. Washington,
DC.

. and U.S. DEPARTMENT OF COMMERCE, BUREAU

OF CENSUS. 1982. 1980 National survey of fishing,
hunting, and wildlife associated recreation. U.S. Govt.
Printing Office. Washington, DC. 1 56pp.

U.S. DEPARTMENT OF TRANSPORTATION. 1985. Cost
of owning and operating a motor vehicle. Federal
Highway Administration. Washington, DC.

U.S. WATER RESOURCES COUNCIL. 1979. Procedures
for evaluation of national economic development
(NED) benefits and costs in water resources planning
in Federal Register, Vol. 44 (242) Final Rule. Decem-
ber 14, 1979.

. 1983. Economic and environmental principles and

guidelines for water and related land resources imple-
mentation studies. March 10, 1983. U.S. Govt. Printing
Office. Washington, DC.

WALSH. 1986. Recreation economic decisions. Dep. Agric.
and Resour. Economics. Colo. State Univ.,
Fort Collins.

796

Economic Analysis

APPENDIX I. BLM Moab District Hunter Day Short Form

The BLM Moab District Hunter Day Short Form
(U.S. Department of the Interior, Bureau of Land
Management 1983) was developed to allow for a sys-
tematic but rapid evaluation of the change in big
game hunter days associated with changes in live-
stock use levels. The first page of the form provides
a simple word model of the changes in life requisites
or habitat components in relation to habitat suitabil-
ity for a particular big game species (Figure A-I-l ).
The word model keys in on the critical factors likely
to be affected by changes in livestock uses.

The current model structure provides some
information on limiting habitat factors but the impor-
tance of that limiting factor is not explicitly recog-
nized. That is, the model structure assumes that lack
of water can be offset by greater cover, space, or
forage. This assumption may be tenuous at times,
and the information on the limiting factor can be
helpful to biologists in designing habitat improve-
ments. If water is the limiting factor, then guzzlers
would provide the biggest payoff, not more cover.
This basic model could easily be modified to make
the limiting factor concept explicit. Techniques for
incorporating limiting factor concepts into indexes
are described in "Standards for the Development
of Habitat Suitability Index Models" (U.S. Depart-
ment of the Interior, Fish and Wildlife Service 1981).

The second page of the Moab District Short
Form requires calculating hunter days from popula-
tion and application of the index factor (Figure A-
1-1). The hunter day /population estimates provided
at the top of the table are for southeastern Utah.
These numbers may seem counter-intuitive because
they combine percentage of harvestable population
and days needed to harvest one animal.

The biologist must estimate the existing popula-
tion level. In addition, for habitat improvement proj-
ects, the prior stable or potential population must
be estimated. If such estimates are available, this
approach is probably preferable. However, existing
population data and potential population estimates
(prior stable) may not be available. In these cases,
harvest estimates must be obtained. State wildlife
management agencies generally have existing harvest
data. Estimating potential harvest is no more difficult
than estimating potential population, given optimum
habitat conditions. In addition, using harvest allows
use of hunter days per animal harvested, a number
that is also more readily available. A Hunter Day
Short Form for use with harvest data is also available
(Figure A-I-2). Page 1 of the form is identical to that
shown on Figure A-I-l, but the calculations on page
2 are based on hunter days per harvest rather than
hunter days per population. Detailed instructions for
using the Moab District Hunter Day Short Form can
be found in BLM's Moab District Bulletin UT-060-83-
86 (U.S. Department of the Interior, Bureau of Land
Management 1983).

In summary, the Hunter Day Short Form pro-
vides a very easy to use and yet systematic approach
for evaluating biological-economic effects for big
game animals. By using some of the modifications
suggested in this paper, the short form analysis can
be upgraded when sufficient data exist on the impor-
tance of one habitat type as a limiting factor. The
form can be modified to use harvest data when pop-
ulation data are unavailable. The basic advantages
of the Moab District Hunter Day Short Form ap-
proach are minimal data requirements and speedy
analysis. It is a very useful approach when one has
dozens of areas to evaluate in a short time.

Economic Analysis

797

Figure A-I-l. Rangeland investment analysis — based on population.

HUMAN USE AND ECONOMIC EVALUATION

104 ESM A

Appendix A. Bio-Economic Analysis of W

ildlife

Table A-2. Rangel<

Hu

X Hu
of

Allotment 4"|
Wildlife Spcoes

Cateoory Rank

Forage 3
Compe t i t ion

1.2

0

-1,-2

-3

Forage 3

1,2

0
-1.-2

-3

Cover 1,2,3

0

-1.-2.-3

Water 1,2,3
0
-1.-2.-3

Space 1,2,3

0

-1.-2.-3

Tot al Points / (numt

10 I ( 3 »
/ ( >

Release No. 1-85

ind investment analysis - based on population.

nter Day Estimates without Investment package
nter Day Estimates with investment package
Habitat changes through time

Alternative 1

TifER Wildlife Biologist

Criteria C

Forage consumption conflicts totally or
nearly el iminated.

Forage consumption conflicts reduced.

Ho change.

Forage consumption conflicts increased.

Forage consumption conflicts become a
major problem.

Key browse, forb, and grass species
Increase in vigor and trend by more than
30 percent.

Either/or browse, forbs, and grass increase
In vigor and trend by less than 30 percent.

No change.

Either/or browse, forbs, and grass decrease
In vigor and trend by less than 30 percent.

Key browse, forb, and grass species decrease
1n vigor and trend by more than 30 percent.

Cover availability Increases.

No change.

Cover availability decreases.

Water availability increases.

No change.

Water availability decreases.

Spatial conflicts decreases.

No Change.

Spatial conflicts Increases.

Total Points

>er of categories considered « 3) ■ Adjustment

CJTM 3/W IIP,

'oints Points
'lanae 1 Chanae 2

2~

3

1

2-

2-

10

Factor

August 1985

. 3) " | .fo7 | Adjustment Factor Change 1

3) • | | Adjustment Factor Change 2

104-ESM-A-ll

798

Economic Analysis

Figure A-I-l. Rangeland investment analysis — based on population (concluded).

HUMAN USE AND ECONOMIC EVALUATION

104 ESM A

Appendix A

. Bio-Economic Analysis of Wildlife

Table A-2. Continued.

Hunter Day / population (HD/pop) estimates

Deer 2.0 HD/pop
Elk 4.0 HD/pop
Antelope 0.2 HD/pop
Desert Bighorn Sheep 0.2 HD/pop

Base Year:

months o

n allotme

X

nt f 12 - x

IOO
Z-o

ZOO

■ ZS~

Existing P
HD/pop
HD
Length of

opul at ion

stay adjustment

Habitat Worsening |

X

♦

1
1

Habitat Improvement

Z.OO Prior Stable for
IOO Existinq Pop ■

X

X

_ Months on

Al lotment .- 12 ■ x

5o

Base year HD input

'hange 1

•a

SM

Change

X

♦

2

Base Year HO
Adjustment factor
HD Loss
Base Year HO

1

1

1

HO Input

1

1

1

Years to Change

General An

IOO
• fc7

k>l

2.

I34HD

.2S

33HD
,"50 HD

X
X

X

♦

Potential Change
Adjustment factor
Popul at ion Change
HD / pop
HD

Time Adjustment
HD Change
Base Year HD

| HD Input

| Yearj to Change

1

66 1

1

1

104-E

1

Release No. 1-85

-A-12

August 1985

Economic Analysis

799

Figure A-I-2. Rangeland investment analysis — based on harvest.

HUMAN USE AND ECONOMIC EVALUATION

104 ESM A

Appendix A. Bio-Economic Analysis of Wildlife

Table A-3. Range'

Hu

of

Allotment j
Wildlife Species

Cateoory Rank

rorage 3
Compet i t ion

1.2

0

-1.-2

-3

Forage 3

1.2

0
-1.-2

-3

Cover 1,2,3

0

-1.-2.-3

Water 1,2,3
0
-1,-2,-3

Space 1,2,3

0

-1.-2.-3

Total Points / (numb
10 1 ( 5 x

1 ( x

Release No. 1-85

and investment analysis - based on harvest.

nter Day Estimates without investment package
nter Oay Estimates with investment package
Habitat changes through time

Alternative /

OtrBR Wildlife Biologist

CTIM .^ITU

Points Points
lanqe 1 Chanqe 2

Criteria C

Forage consumption conflicts totally or
nearly el lminated.

Forage consumption conflicts reduced.

No change.

Forage consumption conflicts increased.

Forage consumption conflicts become a
major problem.

Key browse, forb, and grass species
increase in vigor and trend by more than
30 percent.

Either/or browse, forbs, and grass increase
In vigor and trend by less than 30 percent.

No change.

Either/or browse, forbs, and grass decrease
In vigor and trend by less than 30 percent.

Key browse, forb, and grass species decrease
in vigor and trend by more than 30 percent.

Cover availability increases.
No change.

2.

3

1

Cover availability decreases.

Water availability increases.
No change.

z

Water availability decreases.

Spatial conflicts decreases.
No Change.

z

Spatial conflicts increases.

Total Points
er of categories considered x 3) » Adjustment

10

Factor

August 1985

3) • | .lol | Adjustment Factor Change 1

3) • | | Adjustment Factor Change 2

104-ESM-A-14

HOO

Economic Analysis

Figure A-I-2. Rangeland investment analysis — based on harvest (concluded).

HUMAN USE AND ECONOMIC EVALUATION

104 ESM A

Appendix A. Bio-Economic Analysis of Wildlife

Table A-3. Continued.

Hun

ter Days/harvest (HD/h

arves

t) estimates

for

Utah

Deer

8.0

HDAiarvest

Elk

11.3

HD/harvest

Antelope

3.0

HDAiarvest

Desert Bi

ghorn

Sheep

13.0

HD,"harvest

from

Utah Big

Game

Harvest

Book

Base Year:

£5" Existing Harvest (or avg. of

a . 3 ye"*)

O HD/harvest

2O0 HD

months on allotment f 12 ■ x

.2.5 Length of stay adjustment

50 Base year HD input

Habitat Worsening

Change 1

Change 2

Base rear HD
Adjustment factor
HD Loss
Base Year HD

~| HD Input

| Years to Change

Habitat Improvement

50 Prior Stable

or

Potential

Long

Run

Harvest

25^ ExistinqHarvest

■ zf

Potential Change

X

.(of
/4>.75

X

Adjustment Factor
Population Cringe

X

&

X

HD /harvest

3 Months on

\y\

HD

Al lotment .- 12 «

X

.15

33

X

Time Adjustment
HD Change

♦

1

1

h

♦

Base Year HD

8&.I

1

1

HD Input

10 1

1

1

Years to Change

Release No. 1-85

104-ESM-A-15

August 1985

Economic Analysis

801

APPENDIX II. Human Use and Economic Evaluation (HUEE) System

The Human Use and Economic Evaluation
( HUEE ) system was developed by the U.S. Fish and
Wildlife Service Western Energy and Land Use Team
(WELUT). Human Use and Economic Evaluation
(HUEE) procedures provide means for determining
both the extent of human uses of wildlife and the
dollar values of these uses. These procedures were
developed and are intended for use with the Habitat
Evaluation Procedures (HEP). The HEP and HUEE
together with the Habitat Suitability Index (HSI)
models (U.S. Department of the Interior, Fish and
Wildlife Service 1980a, 1980b, 1981 ) provide a com-
plete set of procedures for field staff to use in mak-
ing both biological and economic assessments of
wildlife resources.

HUEE procedures can be used in conjunction
with HEP or with just population or harvest data. In
addition, HUEE is compatible with the Travel Cost
Method (TCM) and Contingent Value Method
(CVM) valuation. HUEE can also use unit day values
when the added costs of TCM or CVM cannot be
justified.

The HUEE procedures are designed for use by
field staff, principally biologists, assigned to evaluate
the impacts of resource development projects. These
procedures may be applied in field studies without
the assistance of economists or recreation planners.
However, to apply advanced methods such as the
TCM or CVM, the assistance of a specialist, such as
an economist or recreation planner is recommended
(U.S. Department of the Interior, Fish and Wildlife
Service 1980a).

HUEE provides a way to evaluate both the ef-
fects on supply of hunter days and effects on the
demand for hunter days (Figure A-II-1 ). The term
"potential use" in Figure All- 1 refers to the amount
of use people wish to make of the wildlife resource.
The term "sustainable use" is the amount of use days
that can be provided by the habitat. "Planned use"
is the lesser of the sustainable and potential use.
Thus, HUEE is general enough to account for both
situations where demand exceeds supply at the cur-
rent license cost and also where supply of animals
exceeds the demand for hunting. Figure A-II-2 shows
how supply and demand over time are related to
HUEE's concepts of sustainable use, potential use,
and planned use. In this example, the initial factor
determining hunter days is human demand. By year
20, demand has risen and supply fallen such that
hunter days demanded and supplied arc equal. After
year 20, demand (at current permit fees) exceeds
supply so that lottery rationing or shorter seasons
are likely to keep hunter pressure in line with avail-
able populations.

HEP

\

Change in

Habitat Units

(HUs)

ir

Future
Population
Estimates

ir

Sustainable
Use

Potential
Use

Travel Cost
Method
(TCM)

\

Si

/

/

/

f

Planned
Use

Economic

Value of

Use

<r jt

%

\

Economic Value

of Change

in Use

Contingent
Value
Method
(CVM)

Figure A-II-1. Human use and economic
evaluation (HUEE) system.

Figure A-II-3 illustrates a negative impact to a
species' population from a project. Loss of habitat
generally does not change the demand or potential
use by humans but rather adversely affects the eco-
system's capability to support days of hunting or
fishing. Thus, the effects are modeled as a reduction
in sustainable use. Alternatively, enhancement of
habitat suitability by management can increase sus-
tainable use.

This conceptual framework is fairly complete,
but to be useful, it must be easy to translate into
practice. The two factors influencing the actual
amount of wildlife recreation that can be realized are
potential use and sustainable use.

Potential use can be estimated in several ways.
A biologist can do a time series regression of past
hunter/angler day levels in that management unit.
This equation can then be used to forecast future
wildlife recreation demand, assuming current trends
continue. It is important to remember that these
forecasted levels of demand will not be translated
into actual use unless the wildlife habitat can sup-
port such levels of use.

S02

Economic Analysis

400 T SUSTAINABLE USE

POTENTIAL USE

10 20 30 40

TIME(YEARS)

Potential use

Sustainable use
^without project

IMPACT OF PROJECT „

TIME

Figure A-II-2. Relationships between sustainable
use, potential use, and planned use.

Figure A-II-3. Negative economic impact of a
project from reduction in sustainable use.

An alternative and better way to forecast future
use is to use TCM. This technique estimates the eco-
nomic benefits per day and allows use of forecasting
methods in valuing wildlife use.

Once the number of wildlife recreation days
demanded (potential use) are known, one needs to
develop the amount of use the wildlife habitat can
support or supply on a sustainable basis (sustainable
use). The steps in calculating sustainable use are
shown in Figure A-II-4. The end result of these calcu-
lations are ( 1 ) sustainable use days of fishing or
hunting or (2) nonconsumptive use days associated
with a given alternative plan.

Each species or group of species (e.g., water-
fowl, upland game birds, etc. ) must be calculated
separately. The calculation can begin at Point 1
(Habitat Units [HUs]), Point 3 (animal population
size), or Point 5 (harvest). The starting point is the
biologist's decision and is generally constrained by
the data available on HUs, population size, or har-
vest. If an HSI model for the species being evaluated
is available along with either current population or
current harvest, it is fairly easy to estimate future use
days.

If suitable HSI models are not available, then the
future population size or harvest rate with and with-
out the project must be estimated by some other
method in order to estimate future use davs.

HUs are the basic units of analysis in the U.S.
Fish and Wildlife Service HEP. HEP and the use of
HSI models are described in other chapters in this
book and in U.S. Department of the Interior, Fish and
Wildlife Service ( 1980b, 1981 ). An HSI is a number
from 0 to 1 that serves as an index to the quality
of the habitat in an area for a given species. An index
of 1 represents "optimal" habitat and an index of 0
represents totally unsuitable habitat. For terrestrial
species, HUs are the product of the HSI and the
acreage for a given site.

HSI x Acreage = Habitat Units

example: 0.60 x 1,000 acres = 600 habitat units.

Animal numbers are calculated as the product of
HUs and number of animals per habitat unit (produc-
tivity per HU in Figure A-II-4). This calculation as-
sumes that animal population size is linearly
correlated with both habitat suitability (as measured
by the HSI) and acreage. If both values (HUs and
animals per HU) are available then the calculation is
straightforward.

Typically the latter figure (animals per HU)
must be estimated or calculated from other data. If
current populations are known, the animals per HU
can be calculated by dividing current population (or
the average population for the last 3-5 years) by
current HUs. This yields animals per HU that can

Economic Analysis

803

Percent Catchable

Size

(Fish Only)

Catchable

Crop
(Fish Only)

Habitat
Units
(HUs)

Annual

Harvest

(Fish)

Annual

Harvest

(Wildlife)

— ►

Population
Size

V""

Productivity
per HU

n

Harvest
Rate

A

-fe

= Annual

Sustainable

Use Days

Use Days
per Animal
Harvested

Figure A-II-4. Steps for calculating sustainable use.

then be multiplied by HUs in future target years to
calculate estimated animal numbers under alternative
management.

Next, harvest must be calculated. For terrestrial
species, the population size is multiplied by a sustain-
able or allowable harvest rate. This is a harvest rate
that can be sustained without detriment to the popu-
lation. For fish, the calculation involves the addi-
tional steps of calculating percentage of population
of catchable size from which a catchable crop is
calculated.

The next step is to estimate hunter or angler
days per harvest so that sustainable use days can be
calculated. Hunter days per harvest can be calculated
from data on success rates and number of days in
the field. Otherwise, it must be estimated or extrapo-
lated from other areas. A decrease in the number of
animals available for harvest could cause success
rates to drop, but use days would remain the same.
That is, the same number of hunters continue to
hunt even though they must hunt more days for
each animal harvested. If this is an accurate descrip-

tion of hunter behavior in the study area, then HUEE
can reflect this. This constant level of lower and
lower quality hunter days will be reflected as a drop
in economic benefits because as harvest per day
decreases, benefits per day generally decrease. Thus,
one advantage of HUEE over other systems is its
ability to estimate changes in economic values even
if the number of days hunted or fished remains un-
changed.

The sustainable use days are then compared to
potential use and the lessor of potential use and
sustainable use is recorded as "Planned Use" (Figure
A-II-2). The planned use is then multiplied by value
per day to obtain the economic value of such recrea-
tion. Beyond this point, HUEE provides a format and
step-by-step procedure and accounting system for
projecting these values into the future and making
various types of corrections and adjustments.

More information on the details of HUEE can be
found in the U.S. Fish and Wildlife Service Manual
(U.S. Department of the Interior, Fish and Wildlife
Service 1980a).

804

Economic Analysis

41

WRITTEN
COMMUNICATIONS

Donald Zimmerman

Department of Technical Journalism
Colorado State University
Ft. Collins, CO 80521

"The greatest problem in communication is the
illusion that it has been achieved."

-George Bernard Shaw

Editor's Note: No inventory or monitoring study is
complete until results are communicated to the
people who need to see them. Many of us allot little
time to uniting reports or making verbal presenta-
tions— but these activities are vital if management
and the public are to understand our work and
make better natural resource decisions.

This chapter provides methods to identify and ac-
complish your objectives for written communica-
tions. The chapter on verbal presentations does the
same for oral communications, because speaking
to groups is just as important in getting your mes-
sage across.

INTRODUCTION

Once you finish your data analysis and interpre-
tation, you will write the needed memos, letters, re-
ports, and other documents. You must write. It is
part of your job.

\ MUST WRITS ?

Written Communication

805

Good writing becomes important for two rea-
sons. First, writing that communicates a project's re-
sults may be as important as the study itself. If you
do not communicate your study's results in writing
so others understand the findings and know the
needed actions and management decisions, the re-
source may not be managed properly. By writing,
you document and explain the action needed.

Second, good writing becomes important for
you because your career depends on it. If you do
not communicate with your peers, supervisors, and
administrators, they will not know your abilities and
skills. Successful resource managers communicate
through writing and they become known for making
themselves understood. Good writing advances your
career.

When you think about writing, what ideas
emerge? Why do you write? What do you want to
accomplish? Do you think about your readers?
Do you plan ahead or do you wait until the last
minute to write?

There is more to consider: What constitutes
writing? How do you organize your thinking? Do
you outline? Do you rewrite? Do you find writing
difficult or easy? Do you avoid writing?

In this chapter I present a 1 2-step writing pro-
cess for preparing longer reports:

1 . Thinking and planning

2. Organizing your writing

3. Preparing illustrations

4. Drafting the manuscript

5. Writing abstracts and executive summaries

6. Revising and self-editing

7. Editing for style and grammar errors

8. Correcting spelling errors

9. Editing illustrations — tables and figures

10. Having the manuscript retyped

1 1 . Copyediting the manuscript

12. Seeking peer reviews.

As you read through the chapter, remember that
some documents do not have illustrations and others
do not require peer reviews. Furthermore, no two
professionals write in exactly the same way. Some in-
dividuals may combine steps whereas others may
break the identified steps into several activities. From
the 1 2-step writing process, adopt and adapt tech-
niques to improve your writing process.

I close the chapter with a brief discussion of
printing, the Information Age, and computers' im-
pacts on writing and communication.

A WRITING STRATEGY

Before I discuss writing further, consider how
others judge your writing. They judge your writing
on its content, its appearance, and its communica-
tion effectiveness. The content centers on how well
you carried out your professional problem solving.
This book will help you improve your content. Ap-
pearance centers on style, mechanics, spelling, typ-
ing, layout, and printing. Later, I suggest style man-
uals and guides and techniques for checking your
writing's appearance. Communication effectiveness
centers on your objectives — do you want to inform,
instruct, or persuade your readers? Will your readers
understand the points you are making? Will your
readers agree with you? Will your readers accurately
perceive your content?

Consciously or not, many professionals ask
themselves such questions as they write. By asking
yourself such questions, you will begin thinking
about your writing and answering the questions that
will help you communicate more effectively.

To achieve your objectives, your writing must
be clear and concise. Therefore, I suggest a writing
strategy and include ways to improve your writing's
objectives.

Thinking and Planning

Writing begins when you start thinking about
what you will write, your illustrations, and your con-
tent. Effective writing requires thinking ahead —
weeks, months, or years for larger projects. When
you begin thinking about writing, ask: Who are my
readers? What do I want to communicate to my
readers? What is the appropriate written format?
How long will my writing take?

"Who Are My Readers?"

As a biologist, you have diverse readers. Think
about who reads your writing. Reader-oriented writ-
ing may change information levels, behaviors, and
possibly attitudes. But too often we fail to keep our
readers clearly in mind. Within the Bureau of Land
Management and other government agencies you
have two major groups of internal readers: ( 1 ) tech-
nical readers — i.e., fellow scientists, and (2) man-
agers— i.e., the decisionmakers.

Depending on the content of your writing, you
will have different external readers: ( 1 ) biologists
and managers from other agencies; (2) judges, law-
yers, congressmen, and senators; and (3) the general
public — including special interest and advocacy
groups.

806

Written Communication

Consider the functions of your writing. Most
technical and scientific writing informs, instructs,
persuades, and documents. Informing requires at-
tracting the reader's attention, having the message
accepted, having it interpreted, and having it stored
for later use (Schramm and Roberts 1972). Instruct-
ing requires the same steps as informing, plus the
reader must do something with the message — i.e.,
practice or use the information provided (Schramm
and Roberts 1972). Persuading requires convincing
the readers of your position; it requires that readers
yield to your viewpoint. Documenting provides a
permanent record. Most technical reports, notes,
journal articles, memos, and letters document a pro-
ject, a finding, or a decision.

Think about the outcomes you want from your
writing. Do you only want readers to understand? Or
do you want readers to agree with you? Do you
want the readers to accurately perceive your facts?
Communication may improve understanding, but de-
crease agreement. Your readers may accurately per-
ceive your facts and your recommendations, but
they may not agree with you.

When you write for managers, keep in mind that
they like to have several options and the implica-
tions of each option clearly spelled out. If a manager
decides on option C, what are the likely conse-
quences of that decision? What impact will option C
have on the resources? What impact will option C
have on the agency?

WRITING "SHOULD" BEGltf
W'TH THINKING.

'•• •

Consider wild burros. From a management view-
point, too many burros may destroy a range or re-
duce its carrying capacity for livestock and wildlife.
So you recommend reducing burro numbers. From a
budgetary standpoint, shooting them may be the
least expensive control technique. But such a recom-
mendation would soon cause an uproar. Groups will
clearly understand your recommendation, but many
will not agree with you.

When you think of your readers, think about
how they will use the information you provide. Will
the information be for decisionmaking? If so, what
kinds of decisions will your readers be making? How
will your readers use the information? What informa-
tion will you need to provide so your readers make
a decision based on sound management principles?
What facts must you provide? How should you or-
ganize those facts?

How do your readers read what you write?
Communication research consistently shows that
most people read only what interests them or they
read only what they think will be of value to them.
Readers are selective. And your readers will not read
everything that you write. To find out how people
read what you write, produce a document and dis-
tribute it. Then ask the people what they read. Ask
them what they learned. Their answers will help you
tailor your subsequent writing to their needs.

When people read a document, they do not re-
tain all the information. Too often what they retain
and recall are only the points and information that
reinforce their viewpoints.

Beyond how readers read, consider their frames
of reference — the sum of their background, educa-
tion, and experience. Too often we assume our read-
ers see the world as we do. We assume our readers
understand the terms we use. We assume they com-
prehend what we are saying. Often they do not. At
best, the commonality of the frames of reference of a
writer and reader remain imperfect, incomplete, and
partial.

"What Information Do You Communicate?"

With your readers in mind, what content will
you present? How much information should you pre-
sent? What information will you leave out? Will leav-
ing out some information create problems later?
What information can you include? Will including
some information create problems?

"What is the Appropriate Format?"

Keeping selectivity in mind, choose the appro-
priate format to quickly give readers the key infor-
mation. Table 1 identifies readers and selects formats
and then suggests lengths and purposes. Whenever
possible, be succinct. Use Table 1 as a general guide.

Written Communication

807

Table 1. Summary of target audiences, type documents, suggested maximum lengths, and purpose.

Target
audience

Type
document

Maximum
length (pages)

Purpose

Subordinate
biologist

Memo

Letter

Manual

Guidelines

Procedures

1

1

Varies

Varies

Varies

Inform/Instruct
Inform/Instruct
Inform/Instruct
Inform/Instruct
Inform/Instruct

Other

biologists

(General)

Technical notes
Technical reports
Journal articles

5

5

Varies

Inform/Document
Inform/Document
Inform/Document

General
public

Letters
Short report

2
5

Inform
Inform

Advocacy
groups or
individuals

Letter

Short report
Technical report
EIS

Executive summary
of above

2

5

5

100

5

Document/Inform
Document/Inform
Document/Inform
Document/Inform
Document/Inform

Supervisory
biologists
Supervisory
Staff (i.e., Chief
of Resources)

Short report
Technical report

Memo

Short report

Technical report

EIS

EAR

5
5

1-2

5

Varies

100

100

Inform
Inform

Inform/Persuade

Document

Document

Document

Document

Line manager

Memo

Short report or
Executive summary
of EIS or
Technical report

1-2 with back-
up reports
5
5

Inform/Persuade

Inform/Persuade
Inform/Persuade

Congressional
inquiry
(Usually
public through
Congressman)

Memo

with accompanying

reports, etc. as

documented

2

Inform/Document

Court Judge

Affidavit
Deposition

Varies

Inform/Document
(Persuade)

808

Written Communication

The more you know about your readers, the more
you can target and tailor your writing and communi-
cate more effectively.

"How Long Will Writing Take?"

When you begin writing, you need to know
how many hours, days, or weeks it will take you to
write memos, letters, reports, and other documents.
So keep records. And when you plan a project, bud-
get enough time for completing the project's writing
tasks.

You may find incremental writing useful. Incre-
mental writing involves writing parts of a manuscript
as a project progresses. Often, you can write sections
of a report early in a project and use them in later
manuscripts. Such an approach allows you to pro-
duce a final report in less time at the project's end.
Avoid, if possible, putting off writing until the pro-
ject's end. By beginning your writing early, you will
have time for rewriting which will produce a better
manuscript.

Organizing Your Writing

Too often writers turn to an old report, an old
memo, or other manuscripts as a guide for organiz-
ing their writing. Before you do, ask yourself if regu-
lations, policies, or publication guidances require
that you use a specific organization. If so, follow it. If
allowed, add a memo, letter, abstract, or executive

summary giving the key points your readers should
know. By doing so, you will help busy readers recall
the essential points if they do not have time to read
the entire report.

Make sure your manuscripts communicate. Too
often manuscripts tell about rather than tell. Con-
sider the memo and comments in Figure 1.

For longer papers, you will find outlining makes
your writing easier. Outlining serves as a map to
guide your writing. Outlining saves you time — mak-
ing changes on an outline takes less time than re-
writing a manuscript. An outline helps you see your
organization more clearly than a manuscript's
narrative.

Professionals use different approaches to outlin-
ing. In one of the few studies of outlining practices,
McKee (1975) studied technical writers in the Soci-
ety for Technical Communication. Of the 80 writers
who responded to McKee's survey, 60% used out-
lines consisting of words or words and phrases. An-
other 30% used outlines consisting of words,
phrases, and sentences. Only 5% used the traditional
sentence outline, and 5% used no outline.

Chandler ( 1978) wrote that ready-made outlines
cheat you and your readers. He stated that ready-
made outlines hamper communication because the
authors who use them bury their main points and
force readers to read hundreds of words to learn the
main points. Instead of the ready-made outline,
Chandler recommended letting the contents and the
points you want your readers to know dictate how
you organize your manuscripts.

Some professionals summarize their main points
before they begin their outlines. Professor James A.
Bailey of Colorado State University says he writes his
conclusions in one to five short sentences. Then he
lets those sentences guide his outlining.

Bailey also cautions against letting the order of
data analysis dictate your organization. How you ana-
lyze data may have no relation to the data's order of
importance and the conclusions you draw from the
data. So let the importance of the points dictate the
data's organization and order of presentation.

Other writers begin by making a rough outline
of their major points and repeatedly expand the out-
line. Gradually, an outline emerges that the authors
use as a guide for drafting their manuscripts.

I used such an approach in outlining this chap-
ter. I prepared seven versions before a 23-page typed
working outline emerged. After letting the outline
rest, I discussed it with an editor and other profes-
sionals. Next I incorporated selected points from the
first 13 pages into the last 10 pages. These changes

Written Communication

809

Original Memo

Comments/Questions

FROM: District Biologist

SUBJECT: Interim Preliminary Report

Barstow-Las Vegas Motorcycle
Race on wildlife habitats

The initial investigation was completed
in October 1974 and the follow-up
investigation was completed in October
1975. The investigation was the most
interesting I have undertaken and it
should provide us with practical
information for assessing potential
damages from future motorcycle races
and related off-terrain vehicle use.

There were two study sites chosen in
the vicinity of the start of the
Barstow-Las Vegas Race. Study Sites
I and II contained two different
habitat types. Both pre and post race
sites were evaluated.

On Study Sites I and II the pre-race
capture success was low. The trapping
success rates after the race were even
lower. A marked reduction in trapping
success occurred on the area.

The effects of the race on the habitat
types varied and the agency can draw
specific conclusions from the pre and
post-race data for subsequent races.

Most data were of great practical use
and some were quite interesting. . .
and show how motorcycles and off-terrain
vehicles may have an impact on vegeta-
tion types in the deserts.

It was good that we were able to
complete the project as it provided
excellent information for management
decisions and assessing impacts on
the desert environment.

I think "atta bodys" should go to Dave
Smith and Kim Vandermiden, area
biologists who helped with the project
on their own time and at their own
expense. They helped collect the very
valuable and important information for
us.

Why not say interim report or preliminary report?
Do you really need both words?

Why take the whole sentence to add the dates?
Include them with the subject line.

Why was the investigation the most interesting that

you undertook?

What's the point of the memo? The State Biologist

doesn't have time to wade through a lengthy, general

narrative. Of what practical value should the data be?

How will the data help the agency assess future data?

Provide specifics.

What were the sites? What specific vegetation

sites/habitat types were studied? Why?

Do you mean that you had two or four study sites?

You're unclear. What's your experimental design?

What's the methodology? Provide specifics.

What do you mean low? Provide specifics. What does
even lower mean? Provide specifics. What's a marked
reduction? What are the limitations on the data?

How did the effects vary? Be specific. What
conclusions should you draw?

What are the practical values of the data? What
implications do the data and your interpretations
have for subsequent management? Should you hedge
with "may have an impact" or should you be specific
and identify the impact?

Why good? What does good mean?

What does "excellent information" mean?

Should the State Biologist send a letter of thanks? Or
take other action? Be specific.

Reorganize the memo so you tell the State Biologist
in the first paragraph the impacts of the race on the
habitat types and the management implications.
Provide alternatives; be specific.

Figure 1. Memo that "tells about" more than it "tells.'

810

Written Communication

gave a tighter focus on developing a writing strategy
and minimizing the discussion of research on profes-
sional writing. I then added more marginal notes.
Figure 2 shows one page that I used to guide my
drafting the manuscript.

So, what guidelines emerge from the foregoing
discussion? First, use some form of an outline. The
exact form does not matter as long as it helps you
write. You may find a detailed approach helpful or
you may find a few key words and phrases enough to
guide your writing. Second, let your outline rest be-
fore revising it. Time helps give you a critical per-
spective and will help you improve the manuscript's
organization. Third, when you revise, evaluate your
logic. Do you need to delete points, add points, or
rearrange points to build your rationale and logic?

Preparing Illustrations

Ideally, plan your illustrations — tables, photo-
graphs, graphs, and line art — when you begin the
project. You may encounter problems if you wait un-
til the project's end to shoot photographs and have
your line art prepared. Most illustrations benefit from
early planning in the project. At the latest, prepare
the tables and graphs after you have outlined the
manuscript.

Most style manuals and guides divide illustra-
tions into tables and figures. Any illustration that is
not a table is a figure. Although tables normally pre-
sent numbers, they can present short narratives. Fig-
ures include graphs, line art, photographs, and
listings.

Good illustrations save writing, support the nar-
rative, and show essential points that are difficult to
present in the narrative. Illustrations help you com-
municate your key points clearly and succinctly. And
a good illustration may eliminate hundreds or thou-
sands of words.

What kinds of illustrations should you use? Ta-
bles work best for presenting statistics, data, and
complex information. Graphs — line graphs, bar
graphs, and circle graphs — may give specific informa-
tion, but more often they give a general impression
and show trends. Newspapers and popular magazines
use greatly simplified graphs and tables. Line art
(drawings) serve well for many readers, but line art
must be oriented for readers. Will your readers un-
derstand the line art? Maps, for example, provide
useful, detailed information, but they need a north
arrow, legend, and scale. Without such information,
readers unfamiliar with an area will have trouble lo-
cating the site. Photographs provide an image of real
animals, objects, vegetation, or an area, and often
convey concepts more effectively than tables (Figure
3 ). Photographs work well for most readers, but

some readers may not see what you want them to
see. So provide information in the legend or caption.

Before preparing any illustrations, study the
publication's style for illustrations. If you have no
style manual or guide, follow the current Journal of
Wildlife Management style and CBE Style Manual
(CBE Style Manual Committee 1983).

Tips on Preparing Tables. A standard table format
produces uniformity and makes data presentation
more efficient and effective. Readers need not search
for missing information and they will find table inter-
pretation easier. Figure 4 shows the standard parts of
the table as adapted from the CBE Style Manual.
These parts also conform to the Journal of Wildlife
Management's style.

A cursory review I did of BLM publications
showed inconsistencies such as lack of a title above
tables and inconsistent use of divider lines. Other
problems involved cramming too much information
into a table and reducing tables to a size that many
readers could not easily read the numbers and
words. When reducing, some photocopying ma-
chines blur letters and numbers.

Another common problem involves incorrectly
using significant digits. Some authors use digits after
the decimal when whole numbers would do. Thus,
in a food habits study, reporting a particular food in
5.1% rather than 5% of the samples illustrates un-
necessary detail and a level of precision that does
not exist.

When tables will be printed from typed tables,
use a sans-serif type face such as Letter Gothic or
Orator (Figure 5). Sans serif faces have no cross
strokes at the letter's end. Serif type such as Prestige
Elite, Courier, and Modern has cross strokes at the
letter's end. Such type faces make for easier reading
of lengthy narratives but blur when reduced too
small, and reading becomes difficult.

If possible, use upright or vertical format tables
as Figure 4 illustrates. Avoid using tables that force
the reader to turn the manuscript sideways to read
the table and avoid reducing tables so small that
readers have trouble deciphering the numbers, as in
Figure 6. Many readers will not take the time and ef-
fort to turn a manuscript or decipher small printing.

Avoid multiple-page tables and tables with ex-
cessive columns. In most instances you can avoid the
above-mentioned problems by breaking complex ta-
bles into two or more tables.

Tips on Figures. When preparing figures, put the ti-
tles below, not above the figures. Use a consistent
format and style for titles and do not mix typed,
typeset, and hand-lettered titles. For typed

Written Communication

811

VI. A suggested approach to writing
A. Thinking and Planning

Provide adequate time for thinking and planning your communications.

Di?4*****!^' u Who are my readers? ^It^rt f frA+»t ft- ri4<+w*>

0^ A** - 9 ( — t—t

xLY* Jt7 2. What is my content? [Ft*»»€T'*r» ' ',t%Ti

k r/*" tPk 3. What do I want to communicate to my readers?*^ rf"r '

»<«*

V

*♦**

^EmmmumT*

4. What is the appropriate format and organization for presenting
my information? ^;cf0A^^ u/rf+l U*eU* t*»ft*v\iry$.'

Developing a working outline y"

1. Your outline may be a mixture of sentences, phrases, clauses
and words. It's a working document and seldom will others see it.

2. Outlining helps you see your thinking and putting your thinking
on paper helps most writers. kwW"-tffy/e -rt&At*±

3. Letting an outline rest a few days or more helps you reconsider
your approach.

4. Remember it's easier to revise an outline than it is to revise
an entire document.

5. When you begin writing, keep your outline in mind, but don't
feel constrained by it. After having thought about it, an improved
organization may emerge.

Prepare your tables and figures - - JUut+ra+i Vt- M*T**''al'»

1. Tables and figures should be based on your message and
reader's abilities to understand and interpret them. Anything
not a table is a figure. Figure XX provides guidelines for selecting
different kinds of tables and figures.

INSERT FIGURE XX ABOUT HERE

and figures. , - • ..-..«*• /Aa+a
E rtdLvCH VUY*ril4C

2. By planning illustrations early, illustrators, artists and
photographers can meet the publication's deadlines.

Figure 2. Sample page of working outline developed for this chapter.

812

Written Communication

Figure X. Recovery of Ramsey Creek following stream corridor fencing in 1974.

Table XX . Comparison of some vegetation and physical parameters of Ramsey
Creek from 1974 to 1980 following stream corridor fencing and natural
revegetation.

1974

1976

1978

1980

Vegetation

Percent Cover
Bare Soil
Litter
Grass
Forbs
Shrubs
Trees

8

2

5

12

26
4
23
15
30
2

17
14
9
22
35
23

5

12

9

8

30

36

Stems/Acre
over 1" dbh

-

-

354

986

Physical Parameters

Average stream
width (ra)

3.1

2.7

2.5

2.1

Average depth (m)

.09

.08

.21

.33

Strearabank stability
(7, vegetation)

Poor

(0-25)

Fair

(26-50)

Good
(51-751

Excellent
(75-100)

Streamside cover
(rating) (dominant
vegetation type)

Poor
(Forbs)

Fair
(Grass/Forbs)

Good
(Shrubs)

Excellent
(Shrubs/Trees)

jy Simulated Data

Figure 3. Comparison of tables and photographs illustrating the habitat changes from 1976 to 1980

Written Communication

813

©

0

©

Table XX. Average total mule deer per census on 4 burned
and seeded and adjacent unburned control quadrats in 1984
and 1985. Each quadrat was censused 3 times each January,

©►

(V)»-Mule deer per census

©1 ©•

1984

1985

(jT)»Quadrat

Burn

Control Burn

Control

©-

Apr 1980 burn

©'

Sagebrush

Pinyon/Juniper

Grassland

Oct 1982 burn

Grassland

10. 3a
15. 8b
18.1^

14. 4a

5.4
11.8
10.9

5.6

9.4

7.7

11.1

6.3

8.4

16. la

7.3

5.0

13V

©

ra Greater ( P < 0.05 ) than other treatment within year,
I Mann-Whitney U test.

Lb P <0.1, Mann-Whitney l) test.

Figure 4. Parts of a typical table.

©table number (?) table title or legend ©rule ©spanner head

(?) rules (?) sub-spanner heads with rules © stub head © spanner

©rule ©field ©subheadings ©stub ©rule ©footnotes

814

Written Communication

Serif Type Faces

Prestige Elite

Courier

Sans-serif Type Faces

Letter Gothic

Presentor Orator

Figure 5. Typical type faces illustrating serif and sans-serif type faces.

publications, type all titles and insert them after the
figures have been reduced to fit the page. Make sure
the type remains legible and easy to read. Make the
titles complete so the figure can stand alone by in-
cluding such information as the location of the study
and the plant or animal species studied.

Have graphs and artwork drawn larger than, but
proportional to the final page size. When drawing fig-
ures, do not use too much information or too many
lines and clutter the figure. But provide enough in-
formation so the figure is not confusing. For exam-
ple, adding a regression formula helps clarify your
data in a line graph.

When you submit figures for a publication, mark
the top. When you check the galleys or proofs of the
article, make sure the figure was not inserted upside
down or backwards. (Don't laugh — it happens much
to the embarrassment of authors and editors. ) If
someone else inserts the figures, make sure they
have inserted them correctly.

If you need photographs to illustrate a manu-
script, shoot black and white negatives during the
project. Even if you foresee no need for photo-
graphs, shoot them anyway. And when you have the
prints prepared, request glossy black and white
prints. For a detailed discussion of illustrations, see
Allen (1977) and CBE Style Manual Committee
(1983).

Drafting the Manuscript

To draft your manuscript, use your outline and
illustrations as guides and put your thoughts on pa-
per quickly. Do not become a slave to your out-
line— you may develop ideas that will improve your
organization, your content, or your points. If so, use
them. When you write your first draft, do not worry
about spelling, style, and grammar — correct them
later.

As you work through your outline, use tables,
figures, and other illustrations to reduce the length
of your narrative. Discuss their important points.
Point out what you think the data and information
mean. Place illustrations in the text right after their
first mention in the narrative. So provide instructions
for inserting the illustrations into the text. Insert a
line such as v

Insert Figure

XX about here

to tell the word processing operator, typist, or
printer where to insert the illustration in the copy to
be printed or reproduced.

Today, many professional writers compose di-
rectly on typewriters or word processors. Recent ad-
vances in computers suggest that more and more
professionals from all fields will soon be using com-
puters with word-processing capabilities. But for
now, many biologists do not have access to a com-
puter with word-processing capabilities or even a
typewriter for their personal use. So they must write
longhand with pencil and paper.

Draft your manuscript and then have it typed
double spaced with 1- or 1-1/2-inch margins on all
sides. Type on only one side of the paper. Once you
have a typed copy, make a photocopy for revising
and editing. Store the original in a safe place as you
may need it later.

Writing Abstracts and Executive Summaries

When preparing lengthy documents, include
either an abstract or an executive summary. Often an
abstract is the only part of your writing that busy
managers, supervisors, and biologists will read.
Professionals seldom read an entire report, article, or
document, but they do read the abstracts and execu-
tive summaries. Furthermore, the abstract or execu-
tive summary will often be separated from the report
and circulated to different readers. Therefore, your
abstract or executive summary' must stand alone and

Written Communication

815

I . !

oc

c

1

V

oc

u
3
O
u

u
oc
OJ
a

u

s:

51 3

55

3^ 5

S S £

M £

i! ?. a .? .« « f ° $ t* •

' > c ■ S "5.

i S»h|| 11

C
O

t

■o

«5

u

s

03
OC

c

3
H

OS

e
•s

u

e

■o

c

oi

OC

u

M

03

OC

c
'0

3

■a
u

SO

u

I-

816 Written Communication

contain the essential information your readers need
to know.

avoid sentences with greatly separated subjects and
verbs.

Abstracts and executive summaries should tell
rather than tell about (see Figure 1 ). Simply provide
the necessary specifics and come to your points
quickly. Usually, abstracts are for narrowly-focused,
specific research projects; executive summaries are
for more broad-based, diverse investigations.

Your abstracts should contain brief statements
on the research question or purpose, methodology,
and conclusions. With tight writing, you can con-
dense an abstract to one, two, or three sentences.
But normally your abstracts will run between 25 and
200 words, depending upon the agency's or publica-
tion's guidelines. Do not cite literature, refer to ta-
bles or figures in the report, or include tables or fig-
ures in the abstract.

Executive summaries may run up to 1,000
words, and they explain the project in more depth
than an abstract. Furthermore, executive summaries
are usually prepared for administrators and less tech-
nical readers. Therefore, translate jargon and define
technical terms to be sure your readers understand
your points.

Keep your writing succinct and provide only
enough detail to support your points. Keep in mind
that most readers will not read your full report.
Thus, the executive summary must be a tight presen-
tation of the key points. And, like writing the ab-
stract, do not include citations, illustrations, figures,
or tables.

Finally, most professionals find writing the ab-
stracts and executive summaries easy after they have
written the complete report. Conclusions, headings,
and major titles often provide guidelines for develop-
ing abstracts and executive summaries.

Revising and Self-editing

Clear writing comes from clear thinking and re-
writing. Rewriting entails self-editing — a way of revis-
ing your writing. By rewriting you can sharpen your
narrative and make your points succinctly. Some
professionals rewrite their manuscripts two, three,
four, or more times before they have a copy for oth-
ers to critique.

Self-editing helps you tighten your writing. Too
often we write long, complex, involved sentences.
Most readers find short sentences easier to under-
stand than long sentences. As a rule sentences
should average about 20 words, but avoid short,
choppy sentences. Vary sentence length. Avoid sen-
tences with excessive prepositional phrases; edit sen-
tences with long, complex, involved clauses; and

To practice self-editing, plan to make several
trips through your typed manuscript to achieve the
following goals:

1. Correct obvious typing and spelling errors.

2. Correct errors in content.

3. Correct organizational and logic problems.

4. Eliminate wordiness.

To illustrate self-editing, assume you have drafted the
introduction to a BLM Technical Note as in Figure 7.
How can it be improved? I pose questions about its
content and readers, and then I suggest how you
might tighten the copy. By applying self-editing to
your writing, you will clarify your points.

Correct Typing and Spelling Errors. Begin by
quickly reading through your typed manuscript and
correcting obvious typographical and spelling errors.
Correct the errors by using standard copyediting
symbols as illustrated in Figure 8. The copyediting
symbols show standard changes for retyping or type-
setting the manuscript as galleys — the printed copy.
When copyediting, use a soft-lead pencil. Changes
made with hard-lead pencils are hard to read and dif-
ficult to erase.

Check for Content Errors. Read through your
manuscript again looking for errors in content. Have
you accurately presented the data? Have any typo-
graphical errors changed any meanings? Are you us-
ing the correct terms? Have you used the terms in
the correct context? Have you presented anything or
used an approach that will raise questions about
your problem-solving skills? If so, have you explained
why? Critically review your manuscript as if it were
written by another author.

Errors easily creep in if you are not extremely
cautious and careful. Thus, the author who writes

"During the study we collected 1 2 fish species
in I960, 15 in 1965, 17 in 1970, 18 in 1975,
and 20 in 1980. The 116% decrease can be at-
tributed to. . ."

Here the author should have written, "The
116% increase in species can be attributed
to . . ." Although strikingly apparent, such errors
can creep in when writers become too close to
their writing. To reduce the chances of such
errors slipping into your writing, put it aside
for days or better yet a week or two before
you review it. Then carefully check for errors.
And then have another professional check your
manuscript.

Written Communication

817

Original

Comments

Suggested rewrite

In compliance with the National

Is the first paragraph needed?

To comply with the National

Environmental Policy Act (NEPA)

Don't the readers know about

Environmental Policy Act (NEPA)

of 1970, the Bureau of Land

the NEPA and EISs? If so,

of 1970, the Bureau of Land

Management (BLM) is required

delete the paragraph. If not,

Management (BLM) must file

by law to address the resources

tighten it.

Environmental Impact Statements

and impacts of land use on

(EISs) for uses of national public

national public lands in the form

lands.

of Environmental Impact

Statements (EISs).

Extensive wildlife inventories

Is it necessary to say the BLM

The BLM implemented (began?)

were first implemented on public

implemented the inventories?

extensive inventories . . .

lands following the October 1973

What did the suit change?

lawsuit filed in the Federal

Should it be that specific?

District Court of the District of

Should other conservation

The plaintiffs successfully argued

Columbia by the Natural

groups be mentioned? Who

that the initial BLM inventories

Resources Defense Council and

successfully argued? The

did not adequately address the

other conservation groups. It was

Defense Council? Other

impacts of local grazing and did

successfully argued that

groups? Can the sentence be

not comply with NEPA.

evaluation of livestock grazing on

recast in active voice?

public lands, first filed by the

BLM with the Council on

Environmental Quality was too

general in content to properly

address localized impacts, and

therefore, did not comply with

NEPA.

It is estimated that 310 stream

Why be so wordy? Recast into

The BLM administers about 310

miles and over 1 1 ,800 acres of

active voice for a tight

stream miles and over 1 1 ,800

riparian woodland are

sentence.

acres of riparian woodland in

administered by the BLM in

Arizona. To comply with the

Arizona. In an effort to comply

Here's another wordy

court, this study provided

with the court, it became the

sentence. Can it be tightened?

aquatic habitat and water quality

purpose of this study to provide

Can the "Bureau of Land

information to the BLM's Phoenix

aquatic habitat and water quality

Management" be abbreviated?

District Office.

information to the Phoenix District

Office of the Bureau of Land

Management. These baseline

Again, can the point be made

The baseline data provide

data are to serve as an

more quickly? Is "on public

information for developing

information source to be utilized

lands" necessary? BLM doesn't

management plans for terrestrial

in the management plans for

handle other lands, does it?

and aquatic resources.

terrestrial and aquatic resources

Did the suit focus on only

on public lands.

public lands?

Figure 7. Introduction to a Technical Note with comments and suggestions for rewriting.

818

Written Communication

Meaning

Delete

Circle means to spell out

Circle means to abbreviate

For capitals, underline three times

Draw a slanted line through a letter to change to
lower case

Delete letter and close-up using normal spacing -

Delete letters and use normal spacing between
words

Underline to have words set in italics

Begin a new paragraph

Center copy

Set or type flush left -

Set or type flush right
Insert missing words -

Insert missing letter —

Transpose words

Transpose letters

Separate run-on words
Insert period

Insert comma
Insert hyphen

Let original copy stand
Close up copy

For grossly misspelled words, mark out and write
above

Example

It has zones of high and

26/ trees were sampled
The [Bureau of Land Management) . . .
The department of the interior . . .

The /gency policy includes . . .
Its numbers were redu^ced.

. . . butiti? occurs^ in fewer numbers.

Dipodomys merriami
J Eight species were snap trapped.

"} Introduction [_.
References "C

""JTotals

The highest forj^rd's kangaroo rat . .

. . . northern grasshopper mouse . . .
Eight species were jtrappedj snap!.

Eight spcfkes . . .

Eightspecieawe re/trapped.

Eight species were trapped £V

Additionally.the white-footed mouse .

. . . the white-footed mouse . . .

Simple -mono t ypjrc. vegetation types . •

Arizona Upland was further divided into,
frhrco otondard habitat -oitoo based ony
gradient. They includG(f^desertscrub

/Ttoku-i-^

Its numbers were

crJVje&d .

Figure 8. Selected copyediting marks for revised typed manuscripts.

Written Communication

819

On the basis of your expertise in wildlife, do
you see any errors of logic in your approach?
Did you follow acceptable and logical ways of
solving the problem? Do your data support
your conclusions? Do you limit your conclu-
sions? In what ways might your content be crit-
icized? Have you adequately supported your
position? What weak points in logic, problem
solving, and methodologies become apparent
upon reading? Can you correct the weaknesses?
As weaknesses become apparent, consider the
potential criticism. Should you revise? You may
find yourself revising sentences, paragraphs, or
whole sections.

Check Your Organization. Read through your
manuscript again. Have you adequately considered
the reading habits of your readers? Do you give your
readers an overview of what is to come? Will your
organization help your readers understand your con-
tent? Do you progress logically from point one to
point two to point three? Should you make some
points later in your manuscript? Should some points
be made earlier? Have you adequately denned and
explained points when you introduce them? Does
your organization raise any questions about your
logic?

Eliminate Wordiness. Your writing will benefit by
cutting wordiness. With a heavy edit, you will im-
prove your writing's clarity and conciseness. When
you write tightly, your thinking comes through
clearly and you will increase your reader's
understanding.

To cut wordiness, eliminate unneeded words,
phrases, clauses, sentences, and paragraphs. When
possible, replace a clause with a phrase, and a phrase
with a word. Question your words. Can you say
what you want more effectively? Can you make your
points in fewer words? Do you need that word? That
phrase? That clause? What can you eliminate?

Developing your self-editing skills requires rec-
ognizing wordy constructions and knowing how to
improve them. To improve your writing:

1. Use active voice.

2. Use pronouns.

3. Replace small, weak verbs.

4. Eliminate unnecessary clauses.

5. Eliminate unnecessary phrases.

6. Omit needless words.

7. Replace vague, general, and abstract terms.

8. Avoid jargon and acronyms.

Use Active Voice Active voice makes clear who
did what. It tells the reader who carried out the ac-
tion. Compare the active voice with passive voice
constructions:

Active: The division manager cancelled the project.

Passive: The project was cancelled.

Passive: The project was cancelled by the division
manager.

In the first sentence we know who did what. In the
second sentence, we do not know who cancelled the
project. In the third sentence, we finally learn who
did what, but late in the sentence. The active voice
quickly tells your reader who carried out the action.
Use the active voice for shorter sentences.

You can recognize whether a sentence is active
or passive by asking, "Did the subject carry out the
action?" If not, you have a passive construction. You
can recognize passive constructions by looking for is,
was, were, and other to be verb forms. In the two
passive constructions above, was signals the passive
voice. In some sentences, a prepositional phrase be-
ginning with by signals the passive construction.

When should you use the passive construction?
The CBE Style Manual (CBE Style Manual Committee
1983) suggests the use of passive voice when the
agent of the action is irrelevant and when something
or someone other than the agent is more important.
Consider:

The National Environmental Policy Act was
passed in 1970.

Here the passage of the act was more important than
who passed it. In other words, as illustrated below,
what you want to emphasize dictates whether you
use the active or passive voice:

Many technical reports are prepared by con-
sultants.

Consultants prepare many technical reports.

Use Pronouns. Correctly used, pronouns pro-
vide a succinct way of making your points. The CBE
Style Manual (CBE Style Manual Committee 1983)
endorses "I discovered . . ." rather than "It was dis-
covered . . ." and calls the second example the "pas-
sive of modesty." Avoid such passive voice construc-
tions when your name or the agency's name goes on
the manuscript. By using personal pronouns you
write in active voice and you make it clear who did
what or who recommended an action.

820

Written Communication

Replace Small, Weak Verbs. Many writers
build their sentences around small verbs, is, are, was,
were, and related forms. Sometimes writers use trap-
ped verbs as shown in Figure 9.

Noun

Verb Form

refusal

refuse

promulgation

promulgate

assignment

assign

reliance

rely

verification

verify

happening

happen

occurrence

occur

participation

participate

disclosure

disclose

Figure 9. Common nouns representing trapped
verbs. Adapted from and used by permission of
Felker et al. (1982).

Eliminate Unneeded Clauses. Long, complex
sentences often reflect loose, rambling writing. Most
readers find long sentences difficult to follow. Can
the following sentences be written more succinctly?

It is the opinion of field personnel in the South
Dakota Department of Game, Fish, and Parks
that one of the primary factors influencing pop-
ulation levels of big game and turkeys in the
southern Black Hills is a limited water supply.

How can we tighten the sentence? Consider the first
clause, "It is the opinion of . . ." Why not replace the
clause with "believe." Thus

"Field personnel in the South Dakota Depart-
ment of Game, Fish, and Parks believe . . ."

Looking further, let's replace ". . . that one of the pri-
mary factors influencing . . ." with ". . . restricts . . ."
Recasting the sentence gives

"Field personnel in the South Dakota Depart-
ment of Game, Fish, and Parks believe that a
limited water supply restricts the population
levels of big game animals and turkeys in the
southern Black Hills."

I have trimmed 10 words, nearly 25%, from the orig-
inal 41 -word sentence. Later, when I discuss vague,
general, and abstract words, I show how to tighten
the passage by being specific.

To find trapped verbs, look for words ending in:
-al, -tion, -ence, and -ure (Felker et al. 1982) or look
for nouns near such verb forms as effected, obtained,
performed, produced, required, carried out, imple-
mented, accomplished, occurred (Brandner et al.
1974).

Consider "The area is a typical representation of
. . ." versus "The area represents . . ." or "Excessive
eutrophication of aquatic ecosystems by livestock
defecation results in pollution and bacterial contami-
nation . . ." versus "Livestock defecation pollutes and
produces excessive eutrophication of aquatic ecosys-
tems and contaminates them with bacteria . . ."

Compare "I have made general recommenda-
tions for maintaining and improving each group of
reptile and amphibian" versus "I recommend ways of
maintaining and improving each reptile group." And
consider "It was the purpose of this study to supple-
ment previous work completed for the Bill Williams
drainage (Kepner 1979) in addition to describing
the aquatic resources of the Hassayampa River drain-
age." versus "Our study supplements the work of
Kepner (1979) on the Bill Williams drainage and de-
scribes the aquatic resources of the Hassayampa
River drainage."

Eliminate Unneeded Phrases. Too many
prepositional phrases in a sentence clutter your writ-
ing. Often you can strike out unneeded prepositional
phrases. Look with suspicion on prepositional
phrases ending sentences. Are they really needed?
Do they repeat a point made earlier? If so, eliminate
them. Consider

"We tagged 25 mule deer in this study." versus
"We tagged 25 mule deer."

Change prepositional phrases to modifiers. Consider

"The evaluation report of the BLM . . ." versus
"The BLM evaluation report . . ."

or

"The population level of big game and tur-
keys ..." versus "The big game and turkey pop-
ulation levels ..."

Or replace prepositional phrases with possessives.
Consider "The report of the division biologist . . ."
versus "The division biologist's report . . ."

Written Communication

821

Or change prepositional phrases to adverbs.
Consider

"The rattlesnake crawled in a slow manner
across . . ." versus "The rattlesnake crawled
slowly across ..."

Omit Needless Words. Look for words you can
omit. Crossing them out often improves sentences.
Consider

in order to

in the same general area

brown in color

to

in the same area

brown

Figure 10, from the CBE Style Manual (CBE Style
Manual Committee 1983), lists succinct words and
phrases to replace wordy constructions.

Replace Vague, General, and Abstract Words.

The example of eliminating unnecessary clauses re-
ferred to "field personnel." What did the writer
mean? Biologists? Technicians? Game protectors?
Conservation officers? Replacing "field personnel"
with "biologists" leaves no doubt who believed what:

"Biologists in the South Dakota Department of
Game, Fish and Parks believe that a limited
water supply restricts the levels of big game
animals and turkeys in the southern Black
Hills."

Specifying the big game animals — deer and bear —
would further clarify the point.

If you use vague, general, and abstract terms,
you leave the interpretation to the reader, and the
reader usually has a different frame of reference.
Therefore, replace abstract words with concrete
words. Replace vague words with specific terms. Re-
place general words with precise words.

Figure 1 1 lists examples. If you do not agree,
your interpretation provides even stronger evidence
for using specific, precise, concrete terms.

Avoid Jargon and Acronyms. Although you
understand the jargon of your speciality, your read-
ers may not. Thus, "Palustrine wetlands" may convey
a precise meaning to a waterfowl ecologist, but
other biologists may call such areas "marshes" or
"swamps." Furthermore, your readers may not define
the terms as you do. If you use jargon, define your
terms. As you reread your manuscript, ask yourself if
your readers know the terms you use. If they know
the terms, do they define the terms as you do?

Acronyms create problems. One natural re-
source agency hires CIA agents — not the Central In-
telligence Agency type, but conservation information

assistants. Your varied readers may not know or un-
derstand acronyms you use. If you use acronyms, de-
fine them following their first mention such as
". . . Habitat Management Plan (HMP) . . ."

Some readers will not be familiar with the BLM
or other Government agencies. In addition, many
readers will not know your specialized field and the
acronyms you know so well.

Editing for Style and Grammar Errors

Read through your document checking for
grammar and style errors. Keep in mind that differ-
ent agencies and publications have different style
rules. For most technical manuscripts, follow the
current Journal of Wildlife Management style. To an-
swer specific questions, see the current CBE Style
Manual (CBE Style Manual Committee 1983), the
Chicago Manual of Style (University of Chicago Press
1983), or the U.S. Government Printing Office Style
Manual (U.S. Government Printing Office 1984).

If you are preparing publications for general au-
diences, check with the publication's editor to learn
the publication's style. Many newspapers and maga-
zines follow the Associated Press Stylebook (Angione
1977) or their own style manuals.

Periodically, style rules change and agencies and
publications issue new style manuals. If your style
manuals are more than 5 years old, check to see if
new editions have been issued.

Correcting Spelling Errors

Read through your manuscript checking for
spelling errors. When in doubt, look up the word. If
you cannot find the spelling in a desk dictionary, use
an unabridged dictionary. The Journal of Wildlife
Management and CBE Style Manual instruct authors
to follow Webster's Third International Dictionary,
unabridged (Gove 1966). Other style manuals and
publications may follow other unabridged dictionar-
ies. As a matter of style, dictionaries do spell some
words differently. Be especially careful with hyphen-
ated words. Because the language changes, publish-
ers issue new editions of unabridged dictionaries, so
your office may need to buy a new unabridged
dictionary.

Editing Illustrations — Tables and Figures

Make one trip through your manuscript check-
ing tables and figures. Make sure they are numbered
sequentially through the manuscript. Tables have
one sequential numbering and the figures have an-
other sequential numbering. Make sure your narra-
tive's table and figure references are correct and

822

Written Communication

Wordy

Concise

in the vicinity of near

in view of the fact that because

it is often the case often

it is possible that the cause is the cause may be

it is this that this

it would thus appear that apparently

large numbers of many

lenticular in character lenticular

masses are of large sizes masses are large

necessitates the inclusion of needs, requires

of such hardness that so hard that

on the basis of from, by, because

oval in shape, oval-shaped oval

plants exhibited good growth plants grew well

prior to (in time) before

serves the function of being is

subsequent to after

the fish in question this fish

the tests have not as yet the tests have not

the treatment having been performed after treatment

there can be little doubt that this is this probably is

throughout the entire area throughout the area

throughout the whole of the experiment throughout the experiment

two equal halves halves

a number of few, many, several

an innumerable number of tiny veins innumerable tiny veins

as far as our own observations are concerned,

they show we observed

ascertain the location of find

at the present moment, at this point in time now

bright green in color bright green

by means of by, with

(We) conducted inoculation experiments on inoculated

due to the fact that because

during the time that while

fewer in number fewer

for the purpose of examining to examine

for the reason that because, since

from the standpoint of according to

goes under the name of is called

if conditions are such that if

in all cases always, invariably

in order to to

in the course of during

in the event that if

in the near future soon

Figure 10. Replace wordy constructions with concise terms. From CBE Style Manual (CBE Style Manual
Committee 1983) with permission.

Written Communication

823

Vague, General, Abstract Terms

Specific, Concrete, Precise Terms

small pond a 4-hectare pond

a few hens in the flock five hens in the flock

many grass species 10 grass species

steep terrain 60° slopes

low biomass 10 kilograms

a hot day 40° C with 95% humidity

strong winds 55-miles-per-hour winds

several days six days

considerable disparity precipitation ranged

in monthly precipitation from 0 to 20 cm monthly

diminutive in size less than 8 mm

significant P <0.05

often used daily, weekly, monthly

mildly acidic report pH level

Figure 11. Suggestions for replacing vague, general, and abstract terms with specific, concrete, and precise
terms.

where you want them inserted is clearly marked. Be
sure the insertions follow their first mention in the
manuscript.

Carefully check the tables and figures for con-
tent, style, spelling, and typographical errors. Num-
bers are easily transposed, so recheck every number
in every table and figure in the final draft. Make sure
your calculations are correct. Check row and col-
umn entries because they are easily inverted and
transposed. Make sure no information has been
deleted.

Having the Manuscript Retyped

Once you have made the changes you want,
have your manuscript retyped. When you work with
typists and word processing operators, provide the
publication's layout and format guidelines. They give
instructions for setting margins, beginning the first
page, numbering pages, and placing information on
later pages. Some publication guidelines may call for
not breaking sentences and paragraphs between
pages.

If your manuscript was originally entered on a
word processor, check with the operator on how to
mark changes. Some operators want changes marked
in brightly colored felt-tip pens to speed up their
work. If you delete or add matter, find out how the
operator wants you to do so. Follow the operator's
instructions or you will slow their work of correct-
ing the manuscript.

Copyediting the Manuscript

After retyping the manuscript or revising it on
the word processor, carefully copyedit it. Look for
spelling, grammar, and style errors. Check to make
sure that nothing has been deleted or added. Be-
cause you have worked closely with your manu-
script, you may have difficulty spotting errors. You
will read things in and you will read things out. By
being familiar with the manuscript, you may over-
look errors and details. To overcome that problem,
have a competent peer, secretary, or editor check
your manuscript. But also read the manuscript your-
self.

Reading a manuscript aloud and slowly checking
works well for some authors. Others read their man-
uscripts into a tape recorder and play the tape back.
As the tape plays, they check spelling, grammar, and
style. If you find errors, correct them with the stan-
dard copyediting symbols.

Depending on the editor's guidelines, you may
submit a manuscript with a few copyediting marks,
or you may be required to submit a clean copy.
Manuscripts on word processors are easily changed.
But carefully copyedit the revised pages again; errors
may creep in. Likewise, when you have a page re-
typed, check for new errors that may have been
made in retyping.

Seeking Peer Reviews

Most manuscripts benefit from competent peer

824

Written Communication

reviews. The number of reviews you seek depends
upon your manuscript's length, the content, and the
manuscript's importance to the resource and the
agency. The more important the content, the more
reviews your manuscript will need. Sometimes you
will need only three or four reviews and in other in-
stances you may need more.

When you seek peer reviews, you are asking for
other professionals to view your manuscript from
their perspectives. No two reviewers will have ex-
actly the same comments. Some comments will be
superficial; others will provide detailed, constructive
criticism. Some reviewers will agree with you; others
will not.

When you read the reviews, keep in mind that
some reviewers' comments will be extremely helpful
and fair; others will not. When you receive a superfi-
cial, uncritical review, seek another review. When
you receive a highly critical, harsh review, consider
the reviewer's comments carefully. Some reviewers
may disagree with you or be threatened by your
conclusions. Such reviewers may be unfair and bring
irrelevant objections. So weigh each reviewer's com-
ments carefully, but do not be intimidated by their
comments.

Keep in mind your agency's internal review pol-
icy. Such policies prevent erroneous statements oc-
curring in agency documents. If your agency has a
review policy, follow it.

In some instances, you will find yourself having
to make changes reviewers suggest before your man-
uscript is approved for publication. Your editor, and
perhaps your supervisor, will provide helpful
guidelines.

As you rewrite and revise, work carefully. Incor-
porate the needed revisions and carefully copyedit
your manuscript for errors. Check every number in
every table again. Remember changes made now are
easier to make and less costly and time-consuming

than changes made after the final typing or typeset-
ting the manuscript.

PRINTING, COMPUTERS, AND
MANUSCRIPTS

Much of the time, your manuscripts will be
printed in their typewritten form. But occasionally, a
manuscript may be typeset and you may see it
through the printing process. Before you do, learn
the basic steps in the printing process and when and
where to make changes. The CBE Style Manual (CBE
Style Manual Committee 1983) contains three excel-
lent chapters on printing and publishing. Other help-
ful references include Turnbill and Baird (1980) and
Hill and Cochran (1979).

The rapid advances of computers in the printing
industry in the 1970s and 1980s and personal com-
puters in the 1980s ushered in the Information Age.
Software packages enable word processing on most
personal computers. Other software packages check
spelling, grammar, and style. And computers can be
linked to phototypesetters.

You may find yourself more involved with print-
ing than you envision. Preparing manuscripts for
typesetting will soon require entering specific com-
mands on your personal computer. You will need to
enter such information as the type face, size, spacing,
column width, and other typesetting specifications.
At this point you can call your printer and transmit
the manuscript by telephone to the phototypesetter.

The Information Age will change how we pro-
duce documents. The equipment on which we work
may change, but many of the steps will remain the
same. Equipment and software will make some tasks
easier, but you will still be writing and checking
your work. The Information Age will encourage the
proliferation of manuscripts, reports, and other docu-
ments. Therefore, it will become more and more im-
portant to target your communications to specific
readers and to write clearly and concisely.

Written Communication

825

LITERATURE CITED

ALLEN, A. 1977. Steps toward better scientific illustrations.
Allen Press. Lawrence, KS. 64pp.

ANGIONE, E., ed. 1977. The Associated Press stylebook

and libel manual. The Associated Press. 50 Rockefeller
Plaza. New York, NY 10020. 276pp.

BRANDNER, L.O., A. BIDWELL, and I. TEARE. 1974. Let
logic guide your writing. Contribution 1 36, Vice-
President for Agriculture Office, and 1431, Depart-
ment of Agronomy, Agricultural Experiment Station,
Kansas State University. Also published in September-
October 1974, J. of Soil and Water Cons. 29(5):235-
238.

CBE STYLE MANUAL COMMITTEE. 1983. CBE style

manual: a guide for authors, editors, and publishers in
the biological sciences. 5th edition, revised and ex-
panded. Council of Biology Editors. Bethesda, MD.
324pp.

CHANDLER, HE. 1978. The "how to write what" book.
American Society for Metals. Metals Park, OH. 92pp.

FELKER, D.B., F. PICKERING, V.R. CHARROW, V.M. HOL-
LAND, and J.C. REDISH. 1982. Guidelines for docu-
ment designers. American Institutes for Research.

1055 Thomas Jefferson St. NW, Washington, DC
20007. 117pp.

GOVE, P.B., ed. 1966. Webster's third new international
dictionary, unabridged. G & C Merriam Co., Spring-
field, MA. 2662pp.

HILL, M. and W. COCHRAN. 1979. Into print: a practical
guide to writing, illustrating and publishing. William
Kaufman, Inc. One First Street, Los Altos, CA 94022.
175pp.

McKEE, B. 1975. Types of outlines used by technical writ-
ers. J. Eng. Teaching Techniques. 17:30-36.

SCHRAMM, W. and D.F. ROBERTS. 1972. The process and
effects of mass communication. Revised edition. Univ.
of Illinois Press. Urbana. 997pp.

TURNBILL, AT. and R.N. BAIRD. 1980. The graphics of
communication. 4th ed. Holt, Rinehart, Winston. New
York, NY. 398pp.

UNIVERSITY OF CHICAGO PRESS. 1983. The Chicago
style manual. 1 3th edition. Univ. of Chicago Press.
738pp.

U.S. GOVERNMENT PRINTING OFFICE. 1984. The U.S.
Government Printing Office style manual. Revised
edition. Superintendent of Documents. U.S. Govern-
ment Printing Office, Washington, DC. 479pp.

826

Written Communication

APPENDIX.

Suggestions for a Professional Library

From time to time, you will have questions about
writing, mechanics, spelling, and style. Professional
writers and editors regularly have questions too. To
answer their questions, they turn to their office li-
brary of books on writing, data presentation, illustra-
tions, printing, grammar, style, and usage. Below
are suggested publications for your office library.

Books on Writing

CHANDLER, H. 1978. The "how to write what" book.

American Society of Metals, Metals Park, OH 44073-

92pp.
ROMAN, K. and J. RAPHAELSON. 1981. Writing that

works. Harper and Row. New York, NY. 105pp.
ZINSSER, W. 1983. Writing with a word processor. Harper

and Row. New York, NY. 1 1 7pp.
. 1980. On writing well. Harper and Row. New

York, NY. 187pp.

U.S. GOVERNMENT PRINTING OFFICE. 1984. GPO style
manual. Revised edition. U.S. Government Printing
Office. Washington, DC. 479pp.

Data Presentation and Illustrations

ALLEN, A. 1977. Steps toward better scientific illustrations.
Allen Press, Inc. Lawrence, KS. 64pp.

ENRICK, N.L. 1980. Handbook of effective graphic and
tabular communication. Robert E. Krieger Publishing
Company. Huntington, NY. 21 4pp.

REYNOLDS, L. and D. SIMMONDS. 1982. Presentation of
data in science. Martinus Nijhoff Publishers. The Ha-
gue and Boston. 209pp.

Printing

HILL, M. and W. COCHRAN. 1979. Into print: a practical
guide to writing, illustrating, and publishing. William
Kaufman, Inc. One First Street, Los Altos, CA. 94022.
175pp.

TURNBILL, AT. and R.N. BAIRD. 1980. The graphics of
communication. Fourth Edition. Holt, Rinehart, and
Winston. New York, NY. 398pp.

Style Guides

ANGIONE, H., ed. 1977. The Associated Press style book
and libel manual. The Associated Press. 50 Rockefeller
Plaza. New York, NY 10020. 276pp.

CBE STYLE MANUAL COMMITTEE. 1983 CBE style

manual: a guide for authors, editors and publishers in
the biological sciences. 5th edition, revised and ex-
panded. Council of Biology Editors, Inc. Bethesda,
MD. 324pp.

UNIVERSITY OF CHICAGO PRESS. 1983. The Chicago
style manual. 1 3th edition. University of Chicago
Press. Chicago. 738pp.

Grammar, Style, and Usage References

BERNSTEIN, T.M. 1978. The careful writer: a guide to
modern usage. Atheneum. New York, NY. 48"?pp.

. 1971. Miss Thistlebottom's hobgoblins. Farrar,

Straus, and Giroux. New York, NY. 260pp.

CREWS, F. 1979. The Random House guide. Random
House. New York, NY. 434pp.

O'HAYRE, JO. No Date. Gobbledygook has gotta go. Su-
perintendent of Documents. U.S. Government Printing
Office. Washington, DC. 20402. 112pp.

STRUNK, W. and E.B. WHITE. 1979. The elements of style.
Third edition. MacMillan Publ. Co. Inc. New York,
NY. 85pp.

Written Communication

827

42

VERBAL
PRESENTATIONS

Eugene Decker

Department of Fishery and Wildlife Biology
Colorado State University
Fort Collins, CO 80521

'Silence is the virtue of fools.

-Francis Bacon

Editor's Note: Communication should not end with
a written report. Unfortunately, many of us avoid
speaking to groups of people. We must overcome
this tendency if wildlife is to be adequately repre-
sented in multiple-use decisions and allocations.

Proper planning, practice, and advance work will
greatly improve the presentations you give, while
increasing understanding and acceptance of your
results.

INTRODUCTION

Communication has been defined as "the pro-
cess of informing and being informed; it implies un-
derstanding." In other words, if there is no under-
standing, there is no communication. Personal
presentations are often the best means to inform an
audience, as they provide a better opportunity to
communicate than do written communications.

A speaker usually receives immediate reaction
from the audience and can appropriately respond. In
addition to directly answering questions, the speaker
can respond to comments, facial expressions, atten-
tion levels, gestures, and other signals from the audi-
ence during the presentation. Such responses can
quickly clarify points not understood by the
audience.

Another advantage of personal presentations is
that a good speaker can convey credibility, sincerity,
concern, and emotion. This personal factor can be
most effective in getting a message across to a gen-
eral public audience.

A verbal presentation of a management plan by
a member or members of the resource staff may
often determine the acceptance of the plan by other
staff, administrators, or special interest groups. In
fact, the success or failure of many months of effort
by team members may well focus on a single
presentation.

Unfortunately, I have seen such team efforts
frustrated by presentations that were poorly pre-
pared and communicated. It is no wonder that dull,
highly technical, jargon -filled presentations illustrated
with unreadable visuals result in nonacceptance of
management plans. This chapter is intended to assist
personnel in preparing and delivering effective per-
sonal presentations which I hope will result in sound
resource management.

SPEAKER CREDIBILITY

Several factors affect your image as a speaker to

Verbal Presentation

829

the audience, some even before the talk begins. Your
appearance and behavior at a meeting before the
talk, support and endorsement by members of the
group, and a concise introduction by the host stating
your qualifications, all help establish you as a
professional.

Several factors during your talk can also affect
your credibility. First, knowledge of the subject mat-
ter is of utmost importance. Due to constant expo-
sure to media, Americans are a sophisticated audi-
ence and can easily spot a phoney. A speaker should
not try to discuss subjects or answer questions be-
yond his or her expertise.

The second factor is trust, especially with spe-
cial interest groups. An introduction by a host who
endorses you will help establish confidence in the
audience. Also, you should not avoid a controversial
issue, especially if you know your audience is al-
ready aware of it. If there are other sides to an issue,
be sure to acknowledge them before individuals in
the audience bring them up; admit earlier mistakes,
never try to hide them.

Dynamism is the third factor. Expressing inter-
est, concern, and enthusiasm in a well-delivered
presentation can greatly increase your effectiveness.
However, if the first two factors, knowledge and
trust, are weak, then it will make little difference
how dynamic you are. But, if your qualifications and
trust are sound, an energetic, interesting delivery
will greatly enhance your acceptance by the
audience.

3. Another problem is use of words that convey
different meanings to different people. Such
words as ecosystem, carrying capacity, environ-
ment, ecology, conservation, habitat, and wilder-
ness have broad meanings. Another confusing
term often used is resources. I heard an agency
person tell an audience that ". . . we have diffi-
culties managing your resources as Congress did
not give us the financial resources for the re-
sources necessary to do the job." You can imag-
ine the audience reaction!

4. The environment of the meeting site can cause
problems. The audience will not give you its full
attention if the room temperature is uncomfort-
able, the air stuffy due to poor circulation, if
there are distracting materials on the walls or
blackboard, disturbing noises, or the chairs are
positioned awkwardly in reference to the
speaker or screen. You should correct as many
of these problems as possible before the audi-
ence arrives.

5. The audience condition is also an important fac-
tor. You will have problems maintaining audi-
ence attention if people have been seated too
long without a break, if you have run into the
coffee break or lunch hour, if you are the last
speaker of a 4-hour series, or if it is very late in
the afternoon or evening. Be aware of these situ-
ations and try to correct and adjust by shorten-
ing your talk, giving them a break, or trying to
be more dynamic and interesting.

BARRIERS TO COMMUNICATION

A number of problems can interfere with your
audience's understanding of your messages. Below
are several common ones:

1. Use of jargon is a major problem with many
professionals. Jargon, bureaucratic buzz words,
in-terms, and acronyms (HEP, HMP, NEPA, AUM,
etc.) are used to save time when communicat-
ing with colleagues, but are usually unknown to
the general public. Use of such terms causes
confusion and often loss of attention by an
audience.

2. Technical terms, scientific names, and other for-
mal terminology are part of a professional's vo-
cabulary, but these terms are not commonly
used by the general public. Avoid these terms if
possible, but if you have to use one, be sure to
define it. As an example, the term "forb" is regu-
larly used by agency personnel, but it should be
defined for the general public.

6. The cultural, social, and educational aspects of
your audience must also be considered. You
should be aware of their history, experiences,
education, and values in order for you to effec-
tively convey your message to them. It is impor-
tant that your presentation relate to some aspect
of their lives to be clearly understood. An urban
audience (service club, naturalist group, etc.)
may have problems relating to issues of eco-
nomic survival of ranchers whereas a rural audi-
ence may not understand an integrated manage-
ment plan to meet increasing outdoor
recreation needs by city residents.

PREPARING FOR THE PRESENTATION

I offer the following suggestions to assist you in
preparing for personal presentations, particularly
slide talks, to special interest groups and the general
public. They will be covered in four stages: plan-
ning, preparation, presentation, and evaluation.
Presentations to agency staff and management level
personnel are discussed later.

830

Verbal Presentation

Planning

Once you have been asked or assigned to give a
presentation you should consider the following:

1 . Decide on the purpose of your presentation and
write it in a clear, concise statement (for exam-
ple, "to inform the Red Rock Livestock Associa-
tion of wildlife improvements," or "to explain
results of management alternatives on the ante-
lope herd," etc. ). This will help give direction to
your subsequent efforts. In general, your pur-
pose will be to produce a change in your audi-
ence. This may be one, or a combination, of the
following: ( 1 ) awareness, ( 2 ) knowledge, ( 3 ) at-
titude about a situation, or (4) behavior, i.e.,
support or action related to a situation.

2. Prepare an audience analysis. Determine as
much as possible about your audience before
the meeting. Such things as education levels, ex-
posure to previous speakers, age and sex com-
position, previous experiences, stands on similar
issues, their leaders, numbers of persons usually
present, and contact with other staff persons.
Your analysis should give a good picture of your
audience from which you can tailor your
presentation.

3. Complete a preliminary rough outline as soon as
possible. This is helpful as a reminder of needed
facts, graphics, and other visual aids that can be
useful to your talk.

4. Make a list of the visual aids and equipment you
will need. Color slides are especially useful in
describing field situations (before/after treat-
ments, contrasts, impacts, etc.), but may not be
appropriate for other occasions where factual in-
formation only is to be offered. For these pres-
entations, overhead transparencies or a flip chart
alone may be appropriate.

Handouts are useful for many occasions. A con-
cise outline of information may be helpful to
guide an audience through a complex process.
Do not provide lengthy written materials during
a presentation, as some people will read it and
not listen to you. Such materials should usually
be given at the end of the program.

5. Be sure to put the date of the presentation on
your calendar and allow time for preparation
and travel. If it is a major presentation, do not
schedule too much for that date, as a speaker
who has worked all day, driven 2 hours, and
rushed to a meeting will be tired, harried, and
usually a less effective speaker. Your presenta-
tion should be the top priority item for that day.

Refer to Appendix I, Checklist for slide show presen-
tations, to help you with planning for your programs.

Preparation

Now that your planning is underway, you are
ready to move on.

1 . Develop a final outline. I suggest an outline that
includes major items for your introduction,
main content, and conclusion. The introduction
to your talk should relate the audience to your
topic. The introduction should also tell the audi-
ence what you will be covering and why (this
relates to your purpose statement in the plan-
ning phase). You may also wish to tell the audi-
ence more about yourself, your qualifications,
and experience related to your topic. A good in-
troduction in your presentation will increase the
audience reception to you and your ideas.

Some speakers use 5- x 7-inch cards for their
outlines with just one or two brief statements
on each. (Be sure to number the cards to keep
them in order. ) These cards can be helpful
when planning a slide show. Use one card for
each slide and put the number of the slide and a
simple sketch of the photo on the card with the
statement. You can lay these cards out on a ta-
ble (or on a frame "story board") to help you
organize your presentation. In this way you can
visualize the whole presentation and add, re-
move, or rearrange as needed.

Other speakers find an outline typed on one or
two sheets to be effective. Type in caps and tri-
ple space for ease in reading. Use red underlin-
ing to emphasize important material.

You should include a conclusion in your outline
as it is usually a good idea to summarize your
talk. You may then wish to state what actions
may be taken by the audience or where addi-
tional information may be obtained. Be sure to
have a meaningful closing.

The outline is recommended instead of a verba-
tim script so that you can quickly refer to the
outline and then talk directly to the audience.
Few persons can read a script well and maintain
eye contact.

2. Prepare a brief biographic sketch of yourself for
the host to use in introducing you. This should
be a short, one-page statement of your educa-
tion and experience. (I find that typing in all
caps and triple spacing makes it easier to read.)
A good introduction by the host is important, as
the audience must know who you are and that
you are qualified to justify their attention.

Verbal Presentation

831

3. Prepare appropriate and effective visual aids
(see Visual Aids Section following). Also, prac-
tice using them with the equipment that you
will use at the talk.

"How many slides should I use?" I believe there
is no firm rule as your subject and schedule will
usually control the situation. However, too
many slides rapidly changed can provide an un-
pleasant overkill. Likewise, it is not a good idea
to spend too much time on one slide. In the lat-
ter situation, use two or three similar slides to
provide some variety when a discussion is nec-
essary. Good planning and rehearsing can re-
duce this problem so that you have a clear and
concise presentation.

4. Now that you are prepared, take time to re-
hearse. A small cassette recorder will provide a
good tape for your rehearsal. Review it for tim-
ing, organization, technical terms, jargon, and
verbal distractions. Based on this review, revise
your outline and try again. You will be pleased
at the improvement.

Some offices have portable TV equipment. If you
can borrow equipment, tape yourself and then
review it on a monitor (TV set). This can be a
great learning experience — "to see yourself as
others see you."

If you are scheduled for a major presentation in-
volving a potentially controversial subject, I sug-
gest an additional rehearsal. Have several com-
petent colleagues serve as an audience for you
and act as members of various interest groups at
the discussion period after your talk. Make your
presentation to them and then have a "no holds
barred" question and answer session. Because
they have probably been exposed to these situa-
tions, they will be able to ask many slanted and
embarrassing questions. This type of rehearsal
will give you a chance to answer them, then re-
view how effectively you replied. Such an activ-
ity can help you respond better to the real audi-
ences, as you will have been able to anticipate
some of the comments. The participating staff
can also help you with suggestions for appropri-
ate replies.

Always run through your slides before the meet-
ing to be sure that they are right side up, not
backwards, and that you have the appropriate
slides in the cartridge. I also suggest carrying
the slides with you; do not put them in your
baggage if you are flying or give them to some-
one else to bring.

Fred forgot to rehearse.
The audience forgot Fred.

I mentioned earlier that you should have pre-
pared an outline for your presentation. This will
be most useful in your rehearsal and for your in-
troduction and closing, but the written outline
of your major points will be difficult to follow in
a darkened room. With ample rehearsing, you
should be familiar enough with your subject ma-
terial that the slides serve as your outline and

832

Verbal Presentation

you will not need to refer to any written mate-
rial. However, there is usually enough light re-
flecting from the screen onto a correctly placed
podium so that you will often be able to refer to
some of your notes if necessary. A podium light
is usually available, but the light reflecting from
your note paper and hitting the bottom of your
face may give an unpleasant appearance.

6. On the day of your presentation, be sure to
check out the equipment before you leave your
office. Arrive early and check the facilities, i.e.,
seating, podium, heating, ventilation, and light-
ing. Make adjustments before the audience ar-
rives. Also, set up visual aids and audio equip-
ment and test them before the meeting starts.
Position your podium at the front of the room
so that you do not block the audience's view of
the screen. Do not speak from the projector;
stay front and center near the screen.

If you are using a slide or overhead projector,
make sure there is an extra bulb and practice
changing the bulb. If a bulb burns out during
the program, request the lights on and take a 5-
minute break. This will give you an opportunity
to change the bulb quickly, and get organized so
that the show can continue smoothly. It is quite
awkward to change a bulb when you do not
know how and have to have the audience help
you. Also, never grab the bulb with bare fingers
as the only thing the audience will remember is
the smell of burning flesh and what you said
when you grabbed it.

7. Be sure to tie down the cords from the projec-
tor, both power and control cords, to the stand
or table leg before continuing on to the outlet
or podium. This prevents people from tripping
over the cords or pulling the projector off the
table.

8. When setting up the equipment before the
meeting, be sure to have your slide image fill
the screen and keep it as high as possible. If you
have a number of vertical slides, you must ad-
just your image for the verticals which means
that your horizontals will be lower on the
screen. However, if you have just a few verti-
cals, you can usually position the verticals with
part of the top cut off in order to have the hori-
zontals as high and wide on the screen as
possible.

9. Practice with the microphone before the audi-
ence arrives. Have someone stand in the back of
the room to advise you on the correct volume

level. Adjust the microphone for your height
and practice so that you know the correct dis-
tance from "mouth to mike" for best sound.

Presentation

1. Check your appearance. As mentioned previ-
ously, the impression you give to an audience
depends a great deal on your appearance, partic-
ularly before the talk begins. A professional
should dress as a professional, which means that
you should be neatly groomed and appropriately
dressed. It is not what the audience thinks you
are, it is what you convey to the audience of
what you think of yourself. You should convey
the impression that you are a professional and
that you know you are. The too casual approach
can turn an audience off.

If at all possible, try to take it easy, leave your
office early and plan to get to the meeting site
well ahead of time. If appropriate, freshen up
and rest in a motel before the meeting in order
to arrive more relaxed and less harried.

2. Give the introduction with the lights on before
going to the slides. This is important as the audi-
ence can become acquainted with you, your
gestures, and expressions before the lights go
off. Provide a verbal transition to your slides as

a clue to the persons you have appointed to op-
erate the projector and the lights ("Now with
some slides, I would like to show you . . ." or
"May I have the projector, please . . ."). The cor-
rect sequence is for the projector to go on and
then the lights off. This prevents having the au-
dience in total darkness with the projector oper-
ator scrambling to find the switch.

Likewise, it is important to have a transitional
sentence at the end as a clue to the light opera-
tor to turn the lights on and then the projector
off ("That's my last slide, but I would like to dis-
cuss further . . ." or "May we have the lights . . .").
Also, it is advisable to have a meaningful closing,
as described previously, after the slides, and
with the lights on.

3- Change the slides yourself. Be sure to have an
extension cord so that the control can be at the
podium where you are. It is very awkward to
give verbal or sound clues (stamping the floor,
snapping your fingers, etc.) to the projector op-
erator. However, it is advisable to have the pro-
jector operator do the fine focusing of the slides
on the screen as you will be too close to the
screen to know whether or not they are in
sharp focus.

Verbal Presentation

833

4. Meet your host and give him or her your bio-
graphic sketch. Whenever possible, have your
host introduce you to others and talk with them
before your talk if you are at a new place with a
new audience. This is helpful when you get in
front of the group and find that there are a few
friendly faces in the audience. The audience will
also appreciate the fact that you took time to
talk with some of their group. One of the worst
things that you can do is to arrive at a presenta-
tion, sit in a back corner, and not be involved
with the group until you are introduced. Such
actions convey the impression that you care lit-
tle about them and are there just to do a job.

5. Maintain eye contact with the audience. Refer to
your notes, then return to the audience. Try to
look at everyone during each minute or two.
This is hard for some, but practice will help. Eye
contact is an important factor in personal pres-
entation, as it makes the presentation personal.
The audience wants to understand what you are
telling them and not have something read to
them. Only by talking to an audience can you
reflect your interests, concern, and credibility.

Do not talk to the screen as this ruins projec-
tion of your voice for the people in the back of
the room. Look at the screen, make sure you are
talking about the right thing, and turn back to
the audience to continue your presentation.
One way to break talking to the screen is to
point your feet toward the audience, not the
screen.

6. When using a pointer, stand to the side of the
screen and lightly touch the screen. This pre-
vents confusion for the audience as to what you
are pointing at. If you stand away from the
screen, the audience can see the pointer plus
the pointer shadow on the screen which may be
indicating different things.

7. Avoid distracting mannerisms. You should have
picked up on some of these during your re-
hearsals with the tape recorder. Such things and
repetitious terms ("and uh . . . , and . . . , you
know, this is a slide, this is a typical . . .") can be
most distracting, especially when the audience
starts counting them. Excessive or unnatural
movements, hands in pockets, playing with keys
or pointers, scratching, pacing, can also distract
an audience from your message.

8. Maintain enthusiasm. If you come across with a
"don't give a damn" attitude, the audience will
respond in the same way. Monotones or very
fast delivery can also turn them off. You need
not be a Johnny Carson to be effective, but you

should convey your interest and sincerity by a
lively delivery. The audience will respond like-
wise. An increase in volume or an occasional
pause to emphasize a point are often good
techniques.

9. Watch the audience and be responsive to them.
They will convey messages to you by gestures,
expressions, comments, etc., and you will know
when they do not understand a term or are not
following your train of thought. If heads start
nodding and some fall asleep, you know you
have lost them.

The use of humor can release tension and pro-
vide a relief during a serious or technical pro-
gram. However, the humor should be appropri-
ate, low-key, and not distracting. There is no
place for the unrelated, lengthy, or off-color
joke.

10. Keep on schedule. Your rehearsal should have
given you an idea of the length of your talk and
you should stay within the allotted time sched-
ule. This is especially important if you are one
of a series of speakers with assigned time pe-
riods. Some speakers put their watches beside
their outline on the podium to remind them of
the time. Others have the host provide an inob-
trusive signal, i.e., turn over a sheet of paper
with "2 Minutes" written on it when you have

, that time left.

1 1. Instruct your projector operator to turn the pro-
jector completely off with the last slide. Most
projectors give off a buzzing sound which is dis-
tracting to the audience. Also, there is little dif-
ference in the life of your bulb whether the fan
is left on or off.

12. It is usually a good idea to have agreed before-
hand that the host will direct the discussion ses-
sion. The host should not let one or two per-
sons dominate and should dismiss inappropriate
or unrelated comments. This protects you and
permits you to stay on the subject. Avoid argu-
ing as some persons try to bait the speaker. Do
not drag the discussion session on unnecessarily,
and stop while the group is still attentive, not
when they start drifting away.

Evaluation

Evaluation is necessary for self-improvement and
to determine if you have achieved your purpose.
Most of this can be informal. You should know if you
got your message across to the audience by their at-
tentiveness and overall reactions, comments, and
questions.

834

Verbal Presentation

It is also advisable to write a "memo to the files"
on your presentation as soon as possible. Include the
subject, purpose, date, place, audience analysis, reac-
tion, performance by the host and other staff mem-
bers, the quality of the visual aids, and suggestions
for improvements. These will provide reminders
when preparing for the next presentation and may
be helpful to other staff members who will be con-
tacting the same audience later.

You may also consider placing a small cassette
recorder near the podium during your talk so that
you can review it later and make comments on the
memo, if necessary.

If you know someone in the audience who
could objectively review your presentation, you
might consider asking them to do so.

IN-HOUSE PRESENTATION TO SMALL
GROUPS

A situation often faced by resource professionals
is explaining their ideas to personnel of other disci-
plines or to high-level managers, administrators, and
other decision makers. Unless these people can un-
derstand your views on impacts of various alterna-
tives, value of a management practice, or the unique-
ness of a special habitat, your input will have little
meaningful value. Also, you must be effective as this
may be the only chance you will have to make your
case.

The basic principles discussed previously for
general public audiences apply also for presentations
of this type, but there are some modifications that
should be considered. A good general rule is "the
smaller the audience, the less formal the presenta-
tion." However, this does not mean that you should
not be as prepared or organized. With small groups,
(even a single person may be your "audience") you
need to be more conversational, and you must pro-
vide opportunity for appropriate audience participa-
tion. This can be done with ample pauses which en-
courage comments, and with direct questions. These
enable good two-way communications during a
presentation and are effective for many situations.

Visual aids are as helpful for these situations as
for larger audiences. However, color slides may not
be appropriate, especially if you only have a 20-min-
ute appointment with the district manager or with
personnel who know the area and program well. In
these situations, a neatly prepared series of overhead
transparencies or pages on a portable flip chart may
be suitable to support your presentation.

Whenever possible, try to give your presenta-
tion in a small room other than an office where you
can arrange the seating for the expected group size
and set up your visual aid equipment before the
scheduled time. This is especially relevant if your au-
dience is high-level management. Their offices are
usually busy with telephones and other interruptions
that greatly detract from your message.

VISUAL AIDS

Appropriate and properly-prepared visual aids
can greatly increase the effectiveness of your presen-
tation. As the term implies, these provide visual assis-
tance for the audience in understanding your verbal
message.

There are five types of visual aids:

1. Actual objects are the best and should be used
when appropriate. When discussing big game
foods, bring along some clippings of mountain
mahogany, bitterbrush, aspen, etc.; when talking
about sex and aging of wildlife, take along some
deer jaws or grouse wings; or, show them the
specialized equipment used in gaining informa-
tion about your project, such as telemetry gear,
ear tags, big game collars, immobilizing darts,
and live traps.

Verbal Presentation

835

2. Models can be used when the actual object is
too large to bring or too small for the audience
to see. A miniature deer trap that really works
would be a good example.

3. Static graphics are maps, charts, and diagrams
that may be shown to the group or hung on the
wall. Unless these are specially prepared, they
generally are not effective, as they are too com-
plex or the printing is too small. Static graphics
are generally good to have available for the au-
dience to view after your presentation. They
should not be displayed at the front of the room
during the talk as they can be distracting.

4. Active graphics are graphics that progressively
disclose items you are discussing. Examples are
flip charts, hook and loop (velcro) boards, flan-
nel boards, or a chalkboard. These can be effec-
tive in outlining major points for the audience
or a sequence of events or elements of a pro-
gram.

5. Projected visuals include slides, overheads,
television, and movies.

Although you may use several types of visual aids in
your presentation, the following rules should be fol-
lowed in planning and preparing graphics for pre-
senting facts and figures:

1. Boldness. The elements (numbers of words)
must be seen by the audience to be understood.
How many of you have heard a speaker say, "I
know you can't see this, but . . ." or have
strained your eyes trying to read visuals at a
professional conference or staff meeting?

For elements to be easily read by the audience,
they must be 1-1/2 inches high for small rooms
(up to 30 feet), 2 to 2-1/2 inches for medium
rooms (up to 60 feet), and 3 inches or larger
for distances up to 100 feet. It is important to
be able to read the visual from the podium.
Move to the back of the room and see if you
can see it from there. However, if you see that
the letters are only 1/2 inch high on the screen,
you will know that most of the audience can
never see them.

2. Simplicity. Because you want your audience to
recognize the elements immediately, use simple
block letters or numbers as script; fancy letter-
ing delays recognition.

Presentation of complex data is confusing. You
can greatly facilitate understanding by rounding
off, condensing, and abbreviating material. For
example, on a visual the reader has difficulty
comparing

3.

ANNUAL DEER HARVEST IN MONTANA
1960-1980

but when shown as

the reader can immediately see that one is twice
as large. Because you are explaining the graphic,
you do not have to include everything on the
visual as you would in a publication. The
graphic should illustrate your main point.

Also, because you are generally trying to have
your audience understand concepts (projec-
tions, trends, comparisons, etc.), there may be
no need to present precise data. The use of dec-
imal points is seldom necessary and often con-
veys an impression of precision that seldom ex-
ists in natural resources. I suggest rounding off
to the nearest 10, 100, 1000, etc., as appropriate
for fast comprehension. Also, avoid complex,
cluttered graphics. It is better to have several
graphics with a few items than to jam them all
onto one.

Color. Use of color should be considered for
several reasons. Correct use of color can attract
attention and present a more pleasant visual.
However, be aware of contrast as the elements
must stand out from the background if they are
to be easily read. Use dark colors against bright
ones such as yellows, tans, golds, and light blues
and greens.

836

Verbal Presentation

Preparing Graphics

Overhead Transparencies. Overheads are suitable
for many presentations. The advantages are that they
can be used with the room lights on. The speaker
can easily maintain eye contact with the audience as
a transparency can be read correctly on the projec-
tor. Overheads are especially appropriate for smaller
meetings (in-house management) where you may
wish to encourage discussion during the
presentation.

A disadvantage is that in most rooms the
speaker blocks the view of the screen for some of
the audience while at the projector. This can be
avoided by moving to the side of the screen once
the overhead has been positioned correctly. Items
can be identified on the screen with a pointer.

Material for overheads can be easily made from
black elements on a white sheet of standard sized pa-
per (8 1/2 x 11 inches). The sheet is then run
through the office copy machine using a transparent
sheet instead of paper for the copy.

Standard type in office typewriters does not pro-
vide suitable size elements as the letters and num-
bers are too small to be seen when projected. How-
ever, the IBM "orator" type is large enough for
making transparencies for use in a small room. When
preparing overheads using the "orator" type, place
two sheets of paper in the typewriter for better
strike, use all caps, leave a space between each letter
to prevent bleeding, and use triple spacing between
lines.

The best transparencies are made from letters
that are 3/16 inch high or larger. This produces
the correct size image on the screen when pro-
jected. These can be made from dry transfer (press-

on) letters, from a lettering machine (many offices
now have them), from type prepared by a commer-
cial print shop, or made with a hand-lettering stencil
set. The best size type is 24-point (a printer's term
for size), but 18- or 20-point is also acceptable.

Make a rough layout sketch of your overhead.
Use this to position your printing on a white sheet of
paper. I use a light table with a lined sheet under the
white paper to keep spacing uniform and parallel.
With prepared type, you can cut and paste the items
as the edges will not show on the transparency
when copied. You can add lines with a good inking
or felt-tip pen. Color can be added to these black
and white transparencies (underlining, etc.) with
bright colored felt-tip pens.

Suitable graphs, charts, and diagrammatic maps
can be made by using these colored felt-tip pens. Be
sure the pens are made for transparencies or you
will have a smeary mess. Rough out your artwork in
pencil, then place a clear transparency over it and
draw in with the colored pens.

Another method for diagrammatic maps and
some charts is to do the lettering, numbers, and any
lines (roads, streams, x/y axes of graphs, etc.) in
black on a white sheet of paper and make the trans-
parency on the copy machine. Then, add color to
this transparency with the colored pens. Different
habitat types can be cross-hatched in various colors,
or several colored lines can be used on a line graph.

More sophisticated materials can be made with
the help of an experienced graphic artist. The new
generation computers have programs that can pre-
pare high quality artwork on printers. However, re-
member to keep it simple and be sure the elements
are large enough to be seen when projected.

Equipment and materials used in making an overhead
presentation. Left, a thermofax machine that transfers
printed material onto transparencies. Center, transparen-
cies that can be reproduced in a copying machine. Right,
felt-tip markers for writing on acetate film, and an over-
head projector.

Verbal Presentation

837

35mm Slides. Color slides are appropriate visual
aids for many presentations as much field work can
easily be photographed with a 35 mm camera. Slides
can be used to show contrasts, seasonal differences,
relationships, or impacts; explain management prac-
tices and research techniques; and describe a species
or management area. Because slides are so useful in
public presentations, I consider the 35 mm single
lens reflex camera an important part of every biolo-
gist's personal and professional equipment. High
quality equipment is now reasonably priced, and I
suggest you own your own and know how to use it.

it makes your presentation "first person,"
stract "third person."

not an ab-

Effective slides showing changes from pioneer
days can be made from old black and white photos
usually available from the local historical society.
Photograph them with color slide film.

Use only sharp, correctly exposed slides and
keep them clean. Dust can be removed with a camel-
hair brush or a blower. Smears should be removed
with film cleaner.

As you plan a slide presentation, try to include
variety whenever possible. A presentation on deer or
sage grouse management when using only slides of
deer or sage grouse and preferred habitats can be
visually boring. I suggest including photos of other
animals in the habitat: birds, small mammals, snakes,
etc. A few good wildflowers are almost always well
received when describing plants in the habitat in ad-
dition to the main food or cover items. Also, try for
variety in lens angle. When describing a habitat, start
with a wide angle scenic or panoramic, then go to a
slide with a normal lens, move on to a smaller area,
and maybe even closer to a plant community or a
few key plants.

Include people in your photos when appropriate
and yourself, if possible. You in a photo can improve
your credibility and should be supported by verbal
descriptions of your involvement. This is effective as

In addition to pictorial slides, suitable graphic
slides can help clarify items in your presentation.
You can prepare suitable artwork easily and photo-
graph it to make good graphic slides. One method is
to use an 8-1/2- x 1 1-inch or 9- x 12-inch sheet (this
is about the proportion of the 35 mm slide) of col-
ored paper as background for the artwork. I find
light colored art paper in earth tones (golds, or-
anges, tans, and greens) to be good. Printing can be
added by using 24-point (or 20) dry transfer letters
and 1/8-inch matte press-on tapes in various colors
for lines. These tapes are the right proportion in re-
lation to the letters for lines or bars in graphs. The
tape is also available as contour tape that can be
used for making curved lines. The paper, letters, and
tapes are available in graphic art or drafting supply
stores.

Make a rough pencil layout of your design be-
fore starting on the colored paper. Use the edge of a

838

Verbal Presentation

thin sheet of white paper as a line for placing letters,
not a penciled line. The pencil line, when erased,
will photograph a smudged line. A piece of masking
tape will lift off mistakes with the dry transfer letters
when lightly touched to the letter. Keep all impor-
tant material in at least 1-1/4 inches from each edge
to permit easy camera adjustment when photograph-
ing the artwork.

Effective diagrammatic maps can be made by us-
ing different colors of paper to show various owner-
ships, habitat types, seasonal ranges, etc. These are
usually better than showing a slide of a part of a
complex topographical sheet. Pie charts are also ef-
fective graphics as they show comparisons much bet-
ter than numbers on a chart.

A typewriter can be used for making acceptable
slide graphics if the material is kept within a 3- x 5-
inch space on a 4- x 6-inch area of a light colored
sheet. You must come in close and photograph the
4- x 6-inch space when using typed letters in order
for the printing to be large enough when projected.
This is the reason you cannot photograph a page
with a chart or table directly from a publication to
make a slide. The elements will be too small. Take
the material from the publication, simplify, condense,
round off, and abbreviate to make a new graphic, as
described above.

You can do your own photography of artwork
to make your slides. A copy stand is handy, but you
can use a tripod and shoot material held with mask-
ing tape to a wall or on a drawing board. Natural
light works, but flash or flood lamps can be useful.
Practice shooting at various exposures to get the
best results. Unless you have a lot of graphics to
shoot, it may be cheaper to have a commercial pho-
tographer shoot them for you.

CONCLUSION

This chapter is intended to help you with effec-
tive verbal presentations regardless of the audience,
i.e., peers, administrators, advisory groups, or special
interest publics.

A professional is one who knows his subject
area, is experienced, and can communicate ideas
well to other professionals and lay people. Unfortu-
nately, training in communications receives minor
emphasis in university preparation. This situation is
being improved somewhat by in-service training in
some states and by communications workshops of-
fered annually by the Arizona Chapter and the West-
ern Section of the Wildlife Society.

In addition to these possibilities, I suggest you
consider joining the local Toastmasters Club. This
group meets weekly for the sole purpose of improv-
ing members' communication skills. If there is not
one near you, start a club in your locality. Contact
Toastmasters Club International, P.O. Box 10400,
Santa Ana, CA 92711.

The only way to become a good speaker is
through practice behind the podium. Toastmasters
provides such an opportunity. You might also con-
sider attending speech classes offered at the local
university, junior college, or through adult education
programs. In addition to practicing, you will also re-
ceive objective critiques which can be most useful.

Good communication is essential in modern nat-
ural resource management. As professionals, your
challenge is to obtain an understanding of complex
wildlife/habitat interrelationships and communicate
this understanding to people whose support you
need for your management program. I hope you will
accept this challenge.

Verbal Presentation

839

APPENDIX I.

APPENDIX II.

Checklist for slide show presentations.

Suggested Readings.

A. Planning

1. Prepare statement of purpose.

2. Analyze audience — education, number,
experience, etc.

3. Outline presentation.

4. List visual aids and related equipment needs.

5. Clear calendar of conflicts, confirm date.

B. Preparation

1. Prepare final outline.

2. Write biographic sketch for host.

3. Prepare appropriate graphics.

4. Check slides for relevance, quality, timeliness,
dirt, etc.

5. Check out equipment:

• projector (extra bulb)

• projector stand

• screen

• extension cords — control, power

• pointer

• lens cleaning kit — brush, lens cleaner,
tissue

6. Rehearse — check timing, order of slides. Use
tape recorder.

7. Arrive early on day of presentation:

• check facilities — heat, ventilation, seating

• position projector and screen, tie down
cords

• focus and position first slide

• adjust PA system and microphone

• locate light switches

• minimize distracting light

Of the many publications on public speaking and vis-
ual aids, I suggest several that are excellent:

Making Effective Slides for Meetings; Fisher, H.L., Leaf-
let 21119, Cooperative Extension, U.S. Dep. Agric,
University of California, Berkeley, CA 94720

The following are selections from many fine publica-
tions available from Eastman Kodak Company, 343
State Street, Rochester, NY 14650:

Presenting Yourself (S-60)
Effective Lecture Slides (S-22)
Legibility — Artwork to Screen (S-24)
Planning and Producing Slide Programs (S-30)
Slides with a Purpose (VI- 15)

D.

Presentation

1. Check appearance, grooming, etc.

2. Meet host — give biographic sketch, mix with
group.

3. Appoint projector and light-switch assistants.

4. Give introduction with lights on.

5. Begin slides. (Transition: projector on, then lights
off.)

6. Give meaningful closing with lights on. (Tran-
sition from slides: lights on, then projector off —
no fan.)

7. Keep on schedule.

8. Have host handle discussion period.

Evaluation

Write memo to files — include information and
slides needed, response, personnel perfor-
mance.

840

Verbal Presentation

GLOSSARY

This glossary was prepared by using definitions provided by the authors or from the following dictionaries:

The Dictionary' of the Biological Sciences. Peter Gray. Van Nostrand Reinhold Company New York, NY. 1967 edition

A Dictionary of Ecology, Evolution, and Systematics. R.J. Lincoln, G A. Boxshall. and P.F. Clark. Camhridge University Press. London.

McGraw-Hill Dictionary of the Life Sciences. McGraw-Hill Book Company. New York, NY. 1974 edition.

The Random House College Dictionary, 1st ed. Random House, Inc. New York, NY. 1975 edition.

Resource Conservation Glossary, 3rd ed. Soil Conservation Society of America. Ankeny, IA. 1982 edition.

Webster's New Collegiate Dictionary. G&C Merriam Company. Springfield, MA. 1979 edition

Webster's Third New International Dictionary (unabridged) G&C Merriam Company. Springfield, MA. 1966 edition.

ABIOTIC devoid of life; non-living.

ABLATION the removal of a surface layer, as of ice by melting or evaporation.

ABNEY LEVEL a surveying clinometer consisting of a short telescope, bubble tube, and graduated vertical arc-
used especially for measuring tree heights.

ACCURACY relates to the freedom from bias.

AGL (ABOVE GROUND LEVEL) used to define heights that certain aerial censuses should be flown.

ALBEDO a measure of surface reflectivity, that fraction of incident electromagnetic radiation that is reflected
by a body or surface, usually expressed as a percentage.

ALKALI a substance having marked basic properties in contrast to acid properties; a substance with a pH
above 7.

ALKALINITY the quality or state of being alkaline (having a pH balance above 7).

ALTRICIAL helpless and naked when hatched, as young pigeons are (compare to precocial).

ARBOREAL living in or adapted for living in trees.

AREAL any particular extent of space or surface, as a geographical region.

ASYNCHRONY lack of synchronism.

AUFEIS German term for sheets of ice formed by the freezing of overflow water.

AUM (ANIMAL UNIT MONTH) the quantity of forage required by one mature cow ( 1,000 lbs.) or the equiv-
alent for one month.

AUTOECOLOGICAL the study of the individual, or members of a species collectively, in relation to environ-
mental conditions.

AUTOTROPH an organism capable of synthesizing organic nutrients directly from simple inorganic sub-
stances such as carbon dioxide and inorganic nitrogen.

BENEFITS FOREGONE the benefits that would have come from options not chosen when a choice between
options has to be made.

BENTHIC pertaining to or living on the bottom or at the greatest depths of a large body of water.

BEQUEST VALUE the amount an individual would be willing to spend to ensure that a resource would be
present in the future for future generations.

Glossary g4l

BIOGEOGRAPHY the biological study of the geographical distribution of plants and animals.

BIOMASS the total amount of living material, plants and/or animals, above and below the ground in a particu-
lar habitat or area.

BIOTIC pertaining to life or living organisms, caused or produced by or comprising living organisms.

BOGS a plant community that develops and grows in areas with permanently waterlogged peat substrates;
also known as moor or quagmire.

BOREAL pertaining to cool or cold temperate regions of the northern hemisphere, the northern coniferous
zone and taiga.

CARNIVOROUS flesh eating

CHEMOCLINE the transition in a meromictic lake between the mixolimnion layer (at the top) and the moni-
molimnion layer (at the bottom).

CHIONOPHILES animals that thrive in snow-covered habitats.

CHIONOPHOBES animals intolerant of snow-covered habitats.

CIENEGAS an area where the water table is at or near the surface of the ground; standing water occurs in
depressions in the area, and it is covered with grass or sometimes with heavy vegetation.

CLASSIFICATION the act or method of distributing portions of land and their attendant biotic/abiotic attri-
butes into groups or divisions wherein each portion within a group has similar or identical characteristics;
also, naming systematically.

CLINOMETER any of various instruments for measuring angles of elevation or inclination.

COAXIAL CABLE a transmission line that consists of a tube of electrically conducting material surrounding a
central conductor held in place by insulators and that is used to transmit telegraph, telephone, and television
signals of high frequency.

COLD-WATER FISH fish that can stand temperatures ranging from 33 degrees Fahrenheit to 70 degrees Fahr-
enheit.

COLINEARITY having corresponding parts arranged in the same linear order.

CONDUCTIVITY a measure of the ability of a material to conduct an electrical charge; the reciprocal of re-
sistivity.

CONSPECIFIC of the same species.

COOL-WATER FISH fish that can stand temperatures ranging from 33 degrees Fahrenheit to 75 degrees Fahr-
enheit.

CORM an underground stem, such as that of the gladiolus, similar to a bulb but without scales.

CREPUSCULAR active in twilight, at dawn or dusk.

CRYOPEDOGENIC caused by or associated with permanently frozen ground or intensive frost action.

CURSORIAL adapted for running.

DEMAND CURVE a curve that shows the amount demanded at each price level.

DENDROGRAM a branching diagrammatic representation of the interrelations of a group of items sharing
some common factors (as of natural groups connected by a common ancestral form).

842 Glossary

DEPILATORY a liquid or cream used to remove unwanted hair from the body.

DETRITIVORES organisms that feed on fragmented particulate organic matter (detritus).

DETRITUS fragmented particulate organic matter derived from the decomposition of plant and animal re-
mains, organic debris; accumulated in sand, water, or on mud or soil.

DIATOMS any of numerous microscopic, unicellular (single-celled), marine or freshwater algae having sili-
ceous cell walls.

DIMICTIC used to describe a lake having two seasonal overturn periods of free circulation, with accompany-
ing description of the thermocline.

DIURNAL pertaining to those organisms that are active during daytime.

DO (DISSOLVED OXYGEN) refers to the amount of dissolved oxygen in water.

DOMINANCE TYPE species, that by their activity, behavior, or number have considerable influence or con-
trol upon the conditions of existence of associated species; a species that controls its habitat and food web.

DRUMMING to make a succession of strokes or vibrations that produce sounds like drumbeats by beating
the wings, like a male grouse.

DYSTROPHIC pertaining to an environment that does not supply adequate nutrition.

ECHOLOCATION the perception of objects using high frequency sound waves, used by some animals for
navigation and orientation within the environment.

ECOLOGICAL SITES areas of land described by polygons on a map or photograph that have potential for
producing similar potential vegetation, developed from the ecological site system of the U.S. Soil Conservation
Service.

ECOTONE a transition line or strip of vegetation between two communities, having characteristics of both
kinds of neighboring vegetation as well as characteristics of its own.

ECTOTHERMIC poikilothermic, having a body temperature determined primarily by the temperature of the
environment; cold-blooded.

EDAPHIC of or pertaining to soil, especially as it affects living organisms.

EMERGENT PLANTS ( aquatic ) an aquatic plant having most of its vegetative parts above water.

EPHEMERAL ( 1 ) relating to a stream or a portion of a stream that flows only in direct response to precipita-
tion and receives little or no water from springs or no long-continued supply from snow or other sources,
and its channel is at all times above the water table; (2) lasting for only a day; short-lived or transient.

EPILIMNION a freshwater zone of relatively warm water in which mixing occurs as a result of wind action
and connection currents.

ESTIVATE ( a variant of AESTWATE ) to pass the summer, especially in a state of dormancy.

ESTUARINE of, relating to, or formed in an estuary ( water passage ) where the tide meets a river current.

EULITTORAL a subdivision of the benthic division of the littoral zone of the marine environment.

EURY- a prefix meaning wide.

EUTROPHIC rich in nutrients, characterized by abundant plankton and high turbidity.

EUTROPHIC LAKES lakes that are rich in nutrients and organic materials, therefore, highly productive for
plant growth. These lakes are often shallow and seasonally deficient in oxygen in the hypolimnion.

Glossary 843

EUTROPHICATION a means of aging of lakes whereby aquatic plants are abundant and waters are deficient
in oxygen. The process is usually accelerated by enrichment of waters with surface runoff containing nitrogen
and phosphorus.

EXISTENCE VALUE the amount of money an individual would be willing to spend to know a resource pres-
ently exists and would continue to exist.

FACULTATIVE contingent; assuming a particular role or model of life but not restricted to that condition.
Used of organisms having the facility to live, or living, under atypical conditions.

FECUND capable of producing offspring; fertile, productive.

FELL-FIELD a type of tundra ecosystem having sparse, dwarfed vegetation and flat, very stony soil.

FENS a peaty, moist tract, usually derived from the aging or drainage of a swamp.

FISTULA an abnormal passage leading from an abscess or hollow organ to the body surface or from one
hollow organ to another. Sometimes done surgically to enable scientists to examine stomach contents.

FLORISTICS the plant life of a given region, habitat, or geological stratum.

FOSSORIAL adapted for or used in burrowing or digging.

GALLINACEOUS belonging or pertaining to the Order Galliformes, comprising the grouse, pheasants, tur-
keys, partridges, domestic fowls, etc.

GEOMORPHIC of or pertaining to the figure of the earth or the forms of its surface. Of or like the earth, its
shape or surface configuration.

GEOMORPHOLOGY a science that deals with the land and submarine relief features of the earth's surface
and seeks a genetic interpretation of them.

GLEY a bluish-gray or olive-gray sticky layer of clay formed under the surface of certain waterlogged soils.

GREGARIOUS tending to aggregate actively into groups or clusters.

GUILD a group of species having similar ecological resource requirements and foraging strategies and there-
fore having similar roles in the community.

HARD WATER water containing dissolved salts of calcium and magnesium, especially water containing more
than 85.5 parts per million of calcium carbonate.

HEP (HABITAT EVALUATION PROCEDURES) a planning and evaluation technique which focuses on the
habitat requirements of fish and wildlife species. It is a standardized procedure for conducting habitat evalua-
tions in the field developed and used by the U.S. Fish and Wildlife Service.

HERBrVORE an animal that feeds on plants.

HERBrVOROUS (HERBrVORY) feeding on plants; phytophagous.

HERPETOFAUNA animals that are either amphibians or reptiles.

HETEROGENEITY differing in kind; having unlike qualities, possessed of different characteristics.

HETEROTROPHIC obtaining nourishment from exogenous organic material. Used of organisms unable to
synthesize organic compounds from inorganic substrates.

HETEROTROPHS an organism that obtains nourishment from the ingestion and breakdown of organic
matter.

HETEROZYGOSITY having two different alleles at a given locus of a chromosome pair.
844 Glossary

HIBERNACULA the domicile in which an animal hihernates or overwinters; winter quarters.

HOLARCTIC of or designating the zoogeographic region that includes the northern areas of the earth and is
divided into Nearctic and Palearctic regions.

HOME RANGE the area, usually around the domicile, over which an animal normally travels in search of
food.

HOMEOTHERMIC maintains a relatively constant and warm hody temperature that is independent of envi-
ronmental temperatures; warm-blooded.

HORIZON a layer of soil or soil material approximately parallel to the land surface and differing from adja-
cent, genetically related layers in physical, chemical, and biological properties or characteristics.

HQI (HABITAT QUALITY INDEX) this index relates habitat quality to fish biomass in streams. (Refer to
Binns, N.A. 1979- A habitat quality index for Wyoming trout streams. Fishery Res. Rep. Monogr. Series 2.
Wyoming Game and Fish Dep. Cheyenne. 75pp.)

HSI (HABITAT SUITABILITY INDEX) MODEL a specific type of habitat model used by the U.S. Fish and
Wildlife Service in its Habitat Evaluation Procedures (HEPs). A set of habitat variables are used to predict an
index from 0 to 1 of habitat suitability for a species.

HYDRIC SOILS soils that are characterized by or thriving in an abundance of moisture.

HYDROGEN IONS the positively charged ion of hydrogen, H + , formed by removal of the electron from
atomic hydrogen.

HYDROLOGIC relating to the science dealing with the properties, distribution, and circulation of water on
the surface of the land, in the soil and underlying rocks, and in the atmosphere.

HYDROMETEOROLOGY the study of ensembles of liquid in the atmosphere — rain, drizzle, snow, sleet.

HYDROPERIODS the control of vegetative processes in plants by periodic dryness; seasonal hydroperiodism.

HYDROPHYTE ( 1 ) a plant that grows in a moist habitat; (2) a plant requiring large amounts of water for
growth.

HYPOLIMNION the cold bottom water zone below the thermociine in a lake.

INFRALITTORAL the depth zone of a lake permanently covered with rooted or adnate macroscopic vegeta-
tion, often divided into upper (with emergent vegetation), middle (with floating vegetation), and lower zones
(with submerged vegetation).

INTERFACE the ability of computing devices or programs to communicate information from one system to
another.

KROTOVTNAS (also CROTOVTNAS) a former animal burrow in one soil horizon that has been filled with
organic matter or material from another horizon.

LACUSTRINE of, relating to, or growing in lakes.

LAMINAR FLOW streamline flow in a viscous fluid near a solid boundary (compare to turbulent flow).

LANDSAT IMAGE photographic images from satellites, observing earth resources. A satellite program that
developed the Earth Resources Technology Satellite— 1 (ERTS-1 ), which is now called LANDSAT 1.

LAVA TUBES a tubular-like extrusion of lava.

LEK an assembly area for communal courtship displays.

LENTIC pertaining to static, calm, or slow-moving aquatic habitats.

Glossary 845

LIMNETIC of, pertaining to, or inhabiting the pelagic region of a body of fresh water.

LIMNOLOGICAL relating to lakes, ponds, and other standing waters and their associated biota.

LINCOLN INDEX an estimate of population size obtained after release and recapture of marked animals; the
estimated population size (N) is calculated from the number of marked animals released (M), the number
captured in a sample (n) after release, and the number of marked individuals in sample (m) using the formula:

m
This is also known as Lincoln/Peterson index or Peterson estimator.

LINEAR COMPREHENSIVE MANAGEMENT to measure all components of habitats and populations and
know all limiting factors (have comprehensive, quantitative, and continuous knowledge of the system they
manage ).

LITTORAL of or pertaining to the biogeographic zone between the high- and low-water marks.

LOCI (plural of LOCUS) ( 1 ) the position that a gene occupies on a chromosome; (2) the set or configura-
tion of all points satisfying specified geometric conditions.

LOTIC pertaining to fast-running water habitats, such as rivers and streams.

MACROINVERTEBRATES large or exceptionally prominent animals that lack a spinal column.

MACROPHYTES a large macroscopic plant, used especially of aquatic forms such as kelp.

MARGIN the change that would occur with the addition or subtraction of one unit.

MARINE pertaining to the sea.

MARSH a periodically wet or continually flooded area where the surface is not deeply submerged, covered
dominantly with sedges, cattails, rushes, or other hydrophytic plants.

MAST the fruit of the oak and beech or other forest trees, used as food by hogs, birds, and other animals.

MEROMICTIC used of a lake that is permanently stratified due to the presence of a density gradient resulting
from chemical stratification.

MESIC having a moderate rainfall.

MESOTROPHIC having intermediate levels of primary productivity, pertaining to waters having intermediate
levels of the minerals required by green plants.

METALIMNION the zone of steep temperature gradient (thermocline) between the epilimnion and the hypo-
limnion in a lake

MICROCOMPUTER a compact and inexpensive computer relatively limited in capability and capacity, con-
sisting of a microprocessor and other components of a computer.

MICROHABITAT the smallest unit of a habitat, like in a clump of grass or a space between rocks.

MICROPROCESSOR a miniaturized integrated circuit that performs all of the functions of a central process-
ing unit.

MOBBING a collective attack, by a group of animals, on a predator that is too large or aggressive to be
repelled by individual effort.

MOLLUSCS any invertebrate of the Phylum Mollusca comprising the chitons, snails, bivalves, squids, octo-
puses, etc., typically having a calcareous shell of one, two, or more pieces that wholly or partly enclose the
soft, unsegmented body.

846 Glossary

MONOCULTURE raising crops of a single species year after year on the same land.

MONOMICTIC used of a lake having a single period of free circulation or overturn per year, with consequent
disruption of the thermocline; may be either cold monomictic or warm monomictic.

MORAINIC of, pertaining to, forming, or formed by a moraine (an accumulation of earth and stones carried
and finally deposited by a glacier).

MOSAIC an assemblage of overlapping aerial photographs whose edges have been matched to form a contin-
uous photographic representation of an area.

NALED a short-lived insecticide of relatively low toxicity to warm-blooded animals that is used especially to
control crop pests and mosquitos.

NEONATES newborn, recently hatched.

NEOTROPICS a zoogeographical region comprising South America, West Indies, and Central America, south
of the Mexican border.

NONPERSISTENT an item that is not persistent, does not exist for a longer than usual time or continuously.

OBLIGATE essential, necessary, unable to exist in any other state, mode, or relationship.

OLIGOTROPHIC of a lake, lacking plant nutrients and usually containing plentiful amounts of dissolved oxy-
gen without marked stratification.

OMNIVORE an animal that feeds on a mixed diet of plant and animal material.

OMNrVOROUS feeding on a mixed diet of plant and animal material; pantophagous.

OPTION VALUE the option demand is that demand from individuals who are not now consumers or are not
now consuming as much as they anticipate consuming, and who therefore would be willing to pay to perpetu-
ate the availability of the commodity. The option value is the amount an individual is willing to pay.

OXBOW a bend in a river that resembles the U-shaped frame forming a collar about an ox's neck and sup-
porting the yoke.

PALUSTRINE living or thriving in a marshy environment; being or made up of marsh.

PELAGIC ( 1 ) pertaining to water of the open portion of a lake; ( 2 ) pertaining to water of the open portion
of an ocean, above the abyssal zone and beyond the outer limits of the littoral zone.

PERCOLATION the downward movement of water through the soil.

PERIGLACIAL applies to an area bordering the edge of an ice-sheet, to the climate of that area, to physical
processes involving freeze-thaw activity and to their results.

PERMAFROST permanently frozen subsoil.

PERMUTATIONS the act of altering a given set of objects in a group.

PERSISTENT refusing to give up or let go; lasting past maturity without falling off, like certain leaves or
flowers.

PERTLIRBATIONS the state or condition of being perturbed; agitation.

PETERSON ESTIMATOR see Lincoln Index

pH a numerical measure of acidity or hydrogen ion activity. Neutral is pH 7. All pH values below 7 are acid,
and all above 7 are alkaline.

Glossary 847

PHENOLOGICAL the relationship between climate and periodic biological phenomena such as bird migra-
tion or plant flowering.

PHENOLOGY the study of the temporal aspects of recurrent natural phenomena and their relationships to
weather and climate.

PHREATOPHYTES a plant that absorbs water from the permanent water table.

PHYSIOGNOMY the characteristic features or appearance of a plant community.

PHYTOPLANKTON unattached microscopic plants of plankton, subject to movement by wave or current
action.

PIEDMONT a district lying along or near the foot of a mountain range.

PINGO a low hill or mound forced up by hydrostatic pressure in an area underlain by permafrost.

PIONEERING an animal or plant species that establishes itself in a previously barren environment.

PLXEL derived from picture element; the smallest unit on the ground that can be detected by a multispectral
scanner. For the LANDSAT MMS, the pixels are located on the ground on 57 x 79 meter centers.

PLANKTON suspended, floating, or weakly swimming microscopic plants and animals in the water that pro-
vide a basis for the aquatic food chain.

PLAYA the sandy, salty, or mud-caked flat floor of a desert basin having interior drainage, usually occupied by
a shallow lake during or after prolonged, heavy rains.

POINT-CENTER QUADRAT a method of plotless sampling of vegetation in which lines are erected at right
angles from the sampling point to produce four quarters in each, from which a measure is taken of the dis-
tance from the sampling point to the nearest neighbor.

PRECISION relates to the repeatability of a result.

PRECOCIAL of or characterizing birds that are covered with down and capable of moving about when first
hatched.

PROFILE BOARD a scaled board that when photographed behind vegetation displays the physiognomy of a
plant or group of plants. Sequential photographs can be used to display vegetation changes.

PROFUNDAL pertaining to the deep zone of a lake below the level of effective light penetration, and hence
of vegetation.

QUADRATS a small plot or sample area, frequently 1 square meter or 1 milacre in size.

RAIN SHADOW an area of light rainfall situated on the lee side of a range of mountains or hills.

REDD the spawning ground or nest of various fishes.

REDOX POTENTIAL reduction potential; a measure of the tendency of a given system to act as an oxidizing
(electron acceptor) or reducing (electron donor) agent.

RELEVES a random sample of vegetation.

RELICT a remnant or fragment of a flora that remains from a former period when it was more widely distrib-
uted.

RESPIRE CUTANEOUSLY to breathe in and out (inhale and exhale) through the skin.

REVETMENT a facing of masonry or the like, especially for protecting an embankment.

848 Glossary

RIPARIAN HABITAT relating to or living or located on the bank of a natural watercourse (like a river) or
sometimes of a lake or a tidewater.

RIVERINE pertaining to a river.

ROSETTES any structure or marking resembling a rose.

ROTENONE a white, crystalline, water-insoluble, poisonous heterocyclic compound obtained from derris
root; used in certain insecticides.

ROTIFER any of various minute, multicellular aquatic organisms of the Phylum Rotifera, having at the ante-
rior end a wheel-like ring of cilia.

SALINITY the concentration of dissolved solids or salt in water.

SALMONID any of a Family Salmonidae of elongate soft-finned fishes that have the last vertebrae upturned.

SAPROPEL used of organisms inhabiting mud, rich in decaying organic matter.

SAVANNA the tropical and subtropical grassland biome, transitional in character between grassland or desert
and rain forest, typically having drought-resistant vegetation dominated by grasses with scattered tall trees.

SCANNER any of various electronic or optical devices by which images or recorded information are sensed
for subsequent modification, integration, or transmission.

SERAL pertaining to a succession of plant communities in a given habitat leading to a particular climax asso-
ciation; a stage in a community succession.

SERES the entire sequence of ecological communities successfully occupying an area.

SERPENTINE PARENT MATERIAL a soil parent material derived mainly from serpentine rock and domi-
nated by serpentine minerals such as antigorite, chrysolite, fibrolite, and talc.

SOFT WATER water containing little or no dissolved salts of calcium or magnesium, especially water contain-
ing less than 85.5 parts per million of calcium carbonate.

SOIL CONSISTENCE the feel of the soil and ease with which a lump can be crushed by the fingers. Terms
commonly used to describe soil consistence are loose, friable, firm, plastic, sticky, hard, soft, or cemented.

SOIL ERODIBILITY FACTOR (K) a measure of the susceptibility of soil particles to detachment and trans-
port by rainfall and runoff. The K factor is used in the Universal Soil Loss Equation.

SOIL STRUCTURE the combination or arrangement of primary soil particles into secondary particles, units,
or peds.

SPATIALLY the pattern of distribution of organisms in space.

SPECIES RICHNESS the absolute number of species in an assemblage or community.

SPECIFIC CONDUCTANCE the quality of living matter responsible for the transmission of and progressive
reaction to stimuli within the living system.

STATISTICAL TYPE II ERROR when a true alternative hypothesis is rejected.

STENO- a prefix meaning narrow.

STOCHASTIC pertaining to randomness.

STRATA ( 1 ) the divisions/groups into which homogeneous polygons that describe a mapped land area are
placed so that all portions or polygons within a division appear to have similar or identical attributes; (2)
divisions of a classification system.

Glossary 849

STRATIFICATION to become layered, like layers of water temperature in a body of water.

STRIP TRANSECT a method of sampling that entails walking a predetermined line, counting the animals ob-
served, and recording the distances at which they are seen or flushed. The average of the flushing distance is
calculated and used to determine strip width.

SUBMERGENTS pertaining to a plant or plant structure growing entirely underwater.

SUBNIVEAN situated or occurring under the snow.

SUBSTRATE ( 1 ) (biology) the base of substance upon which an organism is growing; (2 ) (hydrology) the
bottom material of a waterway.

SWAMPS areas saturated with water throughout much of the year but with the surface of the soil usually not
deeply submerged.

SYMPATRIC used of populations, species, or taxa occurring together in the same geographical area. The pop-
ulations may occupy the same habitat (biotic sympatry) or different habitats (neighboring sympatry) within
the same geographical area.

SYNCHRONY a synchronous occurrence, movement, or arrangement ( occurring at the same time ).

SYNECOLOGICAL a subdivision of ecology that deals with the study of groups of organisms associated as a
unit.

TAIGA northern coniferous forest biome. The ecosystem adjacent to the arctic tundra, but used with varying
scope to include only the arctic timberline ecotone through to the entire subarctic north temperate forest.

TALUS fragments of rock and other soil material accumulated by gravity at the foot of cliffs or steep slopes.

TDS (TOTAL DISSOLVED SOLIDS) dissolved solids are anhydrous residues of the dissolved substances in
water.

THERMOCLINE a horizontal temperature discontinuity layer in a lake in which the temperature falls by at
least 1 degree Celsius per meter depth; thermal layer.

THERMOREGULATE a mechanism by which mammals and birds attempt to balance heat gain and heat loss
in order to maintain a constant body temperature when exposed to variations in cooling power of the external
medium.

THERMOREGULATION the maintenance or regulation of temperature; the maintenance of a particular tem-
perature of the living body.

TRAMMEL NET a vertically set fishing net of three layers, consisting of a finely meshed net between two nets
of coarse mesh.

TRANSDUCER a device that is activated by power from one system and supplies power in another form to a
second system.

TURBIDITY ( 1 ) the cloudy condition caused by suspended solids in a liquid; (2) a measurement of the sus-
pended solids in a liquid.

TYPING the act of delineating homogeneous areas of land and their attendant vegetation cover on aerial pho-
tographs or maps.

ULTRASONIC pertaining to acoustic frequencies above the range audible to the human ear, or above approx-
imately 20,000 cycles per second

UNGULATE an animal having hooves.

850 Glossary

USLE (UNIVERSAL SOIL LOSS EQUATION) an equation used to design water erosion control systems: A =
RKLSPC wherein A is average annual soil loss in tons per acre per year; R is the rainfall factor; K is the soil
erodibility factor; L is the length of slope; S is the percent slope; P is the conservation practice factor; and C is
the cropping and management factor.

UTM (UNIVERSAL TRANSVERSE MERCATOR) a land data geographic referencing system that is based on a
series of 60 zones worldwide, each covering 6 degrees of longitude in a north-south strip.

WARM -WATER FISH fish that can stand temperatures ranging from 33 degrees Fahrenheit to 80 degrees
Fahrenheit.

WEG (WIND ERODIBILITY GROUP) a soil erodibility grouping for soils based on the stability of soil aggre-
gates against breakdown by tillage and abrasion from wind erosion. The WEG is used in the wind erosion
equation to estimate soil loss due to wind.

WEIR a small dam in a river or stream or a fence, as of brush, set in a stream, channel, etc. for catching fish.

YAGI ANTENNA a highly directional and selective shortwave antenna consisting of a horizontal conductor of
one or two dipoles connected to the receiver and a set of nearly equal insulated dipoles parallel to and on a
level with the horizontal conductor.

ZOOPLANKTON unattached microscopic animals of plankton having minimal capability for locomotion.

Glossary 851

INDEX

Absolute density, 353, 354, 356, 357, 361-364, 366, 448,

463, 465, 488, 490, 497, 499, 504, 515
Aerial counts, 266, 363, 382, 535-538, 540, 544, 548
Aerial photography, 49, 54, 61, 62, 85, 99, 103, 109, 160,

161, 164, 193, 330, 353, 360, 382, 397, 399, 405, 491,

496, 542, 545, 551, 589, 758
Aerial photos, 54, 62, 65, 67, 222, 332, 729
Aerial tracking, 688, 689, 691
Age determination, 264, 266, 559, 562
Air temperature, 568, 714
Alpine tundra, 149, 150, 151, 153-156, 161, 167, 409, 547,

548
Analysis of variance, 656, 747
Anderson, Bertin, 169, 693
Animal-created structures, 140
Animal data collection, 12
Animal-made features, 588, 598, 600
Annual forage cycle, 520
Antennas, 681-685
Aquatic habitat features, 257
Arctic tundra, 149, 152-155, 157-159, 163-166, 351, 353,

356, 373
Aspect, 23, 65, 80, 131, 133, 14 1, 144, 157, 217, 248, 270,

378, 410, 443, 480, 482, 526, 547, 548, 552, 589,

591, 601, 614, 615, 691, 721, 760, 830
Auditory census, 423-425
BIOSTORET, 634
Bailey, James A., 519, 711
Band-tailed pigeon, 411, 420, 426-428
Barriers to communication, 830
Bay ducks, 375
Beaver, 86, 89, 181, 182, 187, 191, 197, 198, 204, 240,

372, 429, 430, 431 434, 435, 440, 444, 445, 447, 451,

452, 485, 588, 598, 599, 710, 836
Belt transects, 328, 398, 399
Bibliographies, 31, 33, 36, 38, 46, 300, 532
Biological limnology, 243, 246-248, 249
Biotic potential, 715

Bird species diversity, 183, 197, 308, 310, 657, 660
Birthing habitat, 528

Blue grouse, 407, 409, 413, 414, 423, 425, 426-428, 722
Boyd, Raymond J., 519
Breeding bird atlases, 297

Breeding bird survey, 297, 303, 306, 309, 311,714
Breeding population, 306, 347, 363, 380, 381
Broad-scale systems, 84
Brood counts, 342, 381, 382, 385
California quail, 411, 419, 420, 722
Call, Mayo, 429

Calling surveys, 334, 339, 343, 348
Canonical correlation, 79, 753, 756
Canopy closure, 73, 82, 84, 483, 765, 770
Canopy volume, 82

Capture methods, 263, 266, 278, 284, 504, 516
Carrying capacity approach, 523, 524
Catch per unit effort, 488, 490
Cave-like structures, 500, 501, 502
Caves, 50, 80, 90, 99, 101, 104, 115, 124, 131, 136, 393,

399, 486, 500, 501 502, 515, 517, 527, 533, 546, 552,

568, 592, 593, 595, 760, 772
Census accuracy, 305, 399
Chapman, Joseph A., 453
Chi-squared test, 15, 19

Chihuahuan Desert, 125, 128, 130, 135, 144, 146, 281,

290, 418, 724
Christmas bird count, 298, 300, 308, 312
Chukar, 130, 408, 409, 412, 426, 427
Classification systems, 50, 54, 66, 74, 77, 89, 96-99, 152,

153, 203, 239,603
Classifying vegetation, 639, 640, 648, 650
Cliffs, 50, 67, 80, 90, 104, 106, 116, 130, 131, 134, 136,

153, 269, 292, 332, 333, 339, 373, 378, 388, 389, 393,

399, 405, 431, 437, 499, 501, 502, 517, 527, 546-

548, 553, 587, 588, 592-595, 740, 760, 772, 796
Cluster analysis, 21, 79, 540
Coefficients of variation, 315, 327
Cold-temperate deserts, 125
Cold-water fishes, 258
Colony site selection, 392
Colony size stability, 390
Colvin, Bruce A., 679
Common bird census, 300
Community measurements, 20, 23
Community suitability index, 761, 765
Community surveys, 286
Complete census, 423

Computer system, 67, 450, 729, 730, 733-736, 738
Condition of individual animals, 20, 545
Confidence intervals, 235, 315, 329, 340, 398, 604, 745,

790
Conners, Peter G., 351
Contingent value method, 789, 790, 802
Continuity of riparian habitat, 174
Cooperrider, Allen Y„ 519, 587, 699, 757, 777
Correction factors, 395-399, 402, 703, 706
Cottontails, 422, 423, 427, 451, 453, 460, 461, 462, 465,

466, 467, 468, 470 471, 472, 473, 584, 722, 724
Counts at roosts and colonies, 330
Covariant analysis, 71 1, 717, 718, 719
Cover board technique, 644
Cover mapping, 75, 217, 758

Critical habitat features, 79, 99, 101, 130, 183, 207, 240
Cross, Stephen P., 497
Cuplin, Paul, 225, 257, 603, 633
Dabbling ducks, 207, 220, 224, 371, 374-376, 379, 383,

384, 770
Data analysis, 1, 2, 5, 6, 66, 67, 227, 236, 326, 605, 612,

662, 689, 715, 716, 756, 805, 809
Data analyzing, 691
Data collection, 1, 2, 4, 5, 7, 9, 1 1, 12, 20, 21, 23, 25, 27,

32, 76, 183, 243, 268, 271, 286, 315, 340, 523, 603,

637, 696, 721, 746, 755, 768, 769, 777, 779
Data collection priorities, 183
Data processing cycle, 730, 731, 734
Data recording, 689, 737
Data storage, 7, 37, 64
Dbh, 81, 82, 86, 101, 117, 176, 296, 596
Dead and down woody material, 80-82, 85, 86, 88, 90.

595, 596, 601, 760
Decker, Eugene, 829
Density estimates, 24, 182, 274, 327, 354, 357, 361. 363.

366, 439, 450, 463, 465, 488, 640, 644, 646-648
Density-frequency relationship, 304
Depletion sampling, 19
Descriptive statistics, 742, 743
Descriptors, 77, 203, 640, 744

Index

853

Detection of echolocation calls, 514

Developing habitat models, 765

Difference methods, 701, 702

Digestible energy, 521, 525, 707

Direct effects, 712

Direct observation, 19, 330, 364, 379, 465, 497, 514, 526,

680, 681, 694, 703, 704, 710
Direct search methods, 274
Discriminant analysis, 91, 311, 752, 753, 756
Dissolved materials, 212, 231, 258
Dissolved oxygen, 135, 231, 240, 241, 244, 246, 247, 258,

634, 636, 638
Disturbance, 107, 124, 133, 143, 144, 152, 157, 160, 161,
165, 167, 190, 192, 230, 330, 331, 339, 342, 348, 373,
390, 395, 396, 400-405, 443, 485, 500, 504, 514,
516, 528, 534, 536, 558, 583, 642, 687, 762, 775
Diversity index, 22, 28, 90, 304, 648, 649, 666, 674, 764
Down logs, 596

Drinking sites, 499, 500, 502, 504, 508, 512, 515
Dropping counts, 422, 540
Economic values of wildlife, 785, 786
Ecotone(s), 23, 131, 133, 134, 142, 151, 153, 174, 175,

270, 276, 348, 409, 413, 417, 581
Edaphic habitats, 90, 147, 584, 587, 588, 592, 601, 760
Edge(s), 23, 28, 73, 80-82, 84, 85, 88, 90, 91, 134, 137,
142, 147, 174, 175, 20-4, 215, 220, 221, 230, 355, 360,
362, 364, 372, 393, 401, 408, 4 10, 41 1, 413, 414,
416, 421, 437, 445, 450, 452, 507, 522, 527, 561, 581,
605, 609, 618, 619, 648, 660, 682, 713, 715, 716,
728, 837-839
Eiders, 376-378

Elevation, 23, 25, 65, 67, 80, 133, 141, 142, 157, 174, 177,
181, 182, 190, 226, 231, 238, 258, 270, 361, 363,
410, 413, 414, 417, 438, 443, 480, 482, 523, 526, 548,
552, 581, 589, 591, 606, 614, 615, 619, 622, 623,
636, 664, 665, 718, 721, 753, 760
Emergents, 156, 207, 212, 213, 215, 216, 218, 219, 220,

372, 376
Encounter rate method, 354, 356, 357
Endangered fish species, 257, 261
Eng, Robert L., 371, 407
Erodibility, 575, 577, 582, 614
Eutrophic, 220, 239, 241, 243, 372

Evaluation, 2, 3, 5, 6, 7, 25, 28, 37, 77, 85, 86, 89-91, 108,
121, 161, 162, 195, 201, 202, 217, 219-222, 224,
245, 265, 266, 290, 299, 308, 312, 343, 346, 348, 353,
361, 366, 369, 381, 385, 421, 426, 428, 476, 492,
496, 534, 537, 539, 558-562, 564, 575, 581, 601, 606,
612, 628-630, 632, 660, 696, 697, 698, 710, 724,
738, 757, 758, 768, 770, 772, 775, 776, 777, 778, 780,
782, 785, 791, 792, 796-798, 802, 818, 821, 830,
834, 840
Featured species, 3, 77, 78, 85, 86, 307, 772
Fecal analysis, 558, 703-706, 710
Field sampling, 4, 6, 245, 628, 662, 663
Field techniques, 246, 289, 646
Field-tracking, 686
Fish and wildlife reference, 32
Fish collection, 263
Fish preservation, 263
Fishery studies, 249
Flight line counts, 396, 399
Floating platform. 393
Flood tide counts. 365. 366

Foliage density. 173, 1 "6, I 78, 189,640,642,644-651,
653,655,656.659

Foliage height diversity, 24, 82, 176, 178, 179, 640, 647,

648
Food supply, 87, 89, 207, 231, 258, 260, 299, 344, 373,

407, 408, 418, 476, 479, 524, 694, 699-701, 761, 766
Forage utilization, 699-701, 703, 704, 707, 709
Foraging area approach, 523
Forest habitat classification, 74
Forest openings, 80, 82, 420
Frequency distributions, 264, 744
Frequency of measurements, 644
Galton's log-normal model, 304
Gambel's quail, 113, 177, 411, 419, 420, 423, 424, 426,

428, 658, 722
Geese, 160, 165, 166, 206, 207, 371, 373-375, 379-382,

384, 385, 723
Geographic information system, 67, 591, 606, 728, 740,
Geomorphic habitat features, 592, 760
Goodness-of-fit test, 19
Gray partridge, 408, 409, 426-428
Grazed-plant methods, 701, 702
Great Basin Desert, 123-125, 128-130, 134, 138, 142-145,

147
Ground nest counts, 397, 398
Ground surveys, 264, 334, 338-340, 342, 544
Ground-water, 170, 192, 193, 261, 613, 614, 623, 628
Guilds, 3, 10, 20-22, 26, 28, 74, 78, 79, 85-90, 96, 108-

112, 116, 121, 147, 188, 224, 288, 290,601, 761, 765,

770, 772, 774, 776
HEP (see Habitat evaluation procedures)
HSI (see Habitat suitability index)
Habitat capability models, 772, 775
Habitat capability relation, 25
Habitat components, 3, 7, 17, 18, 21, 22, 23, 24, 25, 26,

52, 82, 90, 124, 130, 131, 133, 136, 137, 141, 144,

174-177, 179, 181-183, 268, 270, 271, 288, 428, 472,

482, 491, 552, 587, 598, 600, 662, 757, 758, 759, 760,

761, 762, 763, 764, 765, 767, 797
Habitat delineation, 580

Habitat evaluation procedures, 25, 86, 90, 91, 121, 221,
222, 224, 265, 266, 343, 770, 772, 775, 776, 796, 802,
803, 830

Habitat layer index, 103, 109

Habitat management, 1, 2, 27, 49, 54, 66, 79, 85, 124, 146,
147, 163, 181, 183, 194-196, 199, 225, 289, 346,
349, 408, 421, 424, 425, 524, 528, 529, 531, 557, 562,
564, 567, 575, 578, 580, 581, 583, 662, 679, 699,
754, 759, 762, 792, 822

Habitat mapping, 4, 6, 25, 49, 54, 55, 67, 166

Habitat models, 28, 82, 86, 96, 315, 342, 343, 757, 759,
760, 761, 762, 763, 764, 765, 768, 770, 775, 776

Habitat quality index (HQI), 233, 265

Habitat requirements, 5, 21, 26, 77, 78, 80, 86, 147, 188,
260, 292, 315, 347, 364, 374, 375, 379, 384, 419, 428,
454, 475, 482, 491, 520, 528, 529, 533, 534, 537,

539, 541, 542, 543, 544, 545, 546, 550, 551, 563, 564,
661, 694, 695, 699, 715, 758, 767, 770

Habitat suitability, 7, 18, 25, 28, 86, 89, 90, 221, 223, 265,
266, 343, 346, 356, 392, 426, 492, 520, 525, 534,

540, 557, 570, 573, 580, 581, 583, 596, 757, 764, 765,
770, 776, 796, 797, 802, 803

Habitat suitability index, 7, 9, 18, 25, 28, 86, 89, 90, 96,
116, 221, 222, 223, 265, 266, 343, 346, 426, 596, 757,

762, 764, 765, 770, 771, 772, 775, 776, 796, 797,
802, 803

Habitat type, 6, 12, 14, 16, 21, 23, 25, 36, 54, 66, 80, 274,
286, 326, 335, 338, 408, 421, 427, 434, 443, 444, 449,

854

Index

450, 475, 482, 491. 538, 583, 584, 649, 655, 686,

691, 770, 797
Hand capture, 504
Hand tracking, 686
Handling techniques, 504, 505
Hegdal, Paul L„ 679
Hiders, 527
Horizonation, 568
Horizontal diversity, 647-649

Horizontal vegetation structure, 138, 139, 268, 271
Hydrologic soil group, 575, 581
Hydrology of lakes, 623
Hypothesis formulation, 746, 747
Impacts on desert habitats, 142
Impoundments, 185, 191, 218, 223, 224, 239, 240, 251,

261, 372, 385
Indicator species, 22, 78, 79, 85-88, 288, 307, 311, 439,

665
Indirect effects, 124, 71 1, 712, 713
Indirect observation, 703
Inferential statistics, 66, 742, 745, 746
Instream flow needs, 227, 605
Integrated mapping, 64
International bird census, 300, 310
Interpretation, 1, 2, 5, 6, 21, 62, 65, 85, 99, 103, 107, 165,

222, 232, 274, 304, 313, 343, 344, 355, 494, 514,

554, 575, 585, 589, 601, 616, 623, 628, 662, 696, 707,

709, 727, 751, 753, 758, 777, 778, 782, 805, 811, 822
Jaccard's similarity coefficient, 21
Jackrabbits, 453, 454, 455, 465, 467, 468, 470, 471, 472,

473, 478, 479
Jones, K. Bruce, 1, 11, 123, 267
Jumping mice, 433
Kangaroo mice, 432

Kangaroo rats, 104, 136, 137, 432, 572, 582
Kerr, Pochard M., 49
Key species approach, 523, 524
King strip census, 422
Kochert, Michael, 313
Laboratory analysis, 662, 665
LaBounty, James, 237

Landforms, 491, 588, 589, 590, 591, 592, 600, 601
I.andsat imagery, 158, 160, 166
Landsat images, 62, 5-»0
Late nesters, 380, 390, 391
Lava flows, 80, 592, 593, 594, 760
Lemmings, 432, 433, 451
Length of fish, 262, 264
Lent, Peter C, 149, 519
Lentic habitats, 134, 135
Levels of data needed, 1,4, 11
Levels of inventory, 243
Life-form approach, 79, 772, 775
Life stages, 258, 261, 263, 360, 607
Life zones of a lake, 241
Lincoln estimator (see Lincoln index)
Lincoln index, 19, 306, 398, 436, 437, 444, 446, 448, 449,

450, 467, 490
Line transect, 24, 219, 302, 310, 326-328, 345, 356, 358,

366, 368, 444, 451, 473, 558
Linear correlation, 656, 717
Linear density transects, 362, 363
Linear regression, 657, 749, 750, 767
Literature searching, 29, 31, 33, 38
Litter, 22-24, 65, 80, 101, 115, 116, 130-132, 138, 141,

178, 182, 267-269, 271, 276, 283, 288, 441, 443, 470,

472, 478, 582, 584, 588, 594-596, 760, 761

Litter and mulch, 595, 596

Livestock grazing, 2, 4, 142, 143, 171, 173, 180, 185-190,

195, 197-199, 234, 426, 534, 595, 605, 662, 772
Logging and roads, 192
Logs, 22, 80-82, 118, 124, 131, 132, 135, 138, 182, 183,

187, 205, 231, 267-269, 274, 275, 276, 281, 432, 437,

440, 443, 484, 486, 503, 596, 597, 609, 761
Loomis, John, 785
Lotic water, 1 34
Macroinvertebrates, 46, 197, 202, 216-218, 221, 227, 229-

231, 258, 607, 611, 661-663, 665, 669-671, 761
Mangum, Fred, 661
Man-made structures, 132, 140, 141, 326, 392, 417, 419,

433, 485, 500, 502, 514, 588, 596, 598-600
Mann-Whitney U-Test, 304, 750, 751
Manning equation, 622
Mapping variables, 578
Mark and recapture, 15, 19, 20, 264, 281, 306, 436, 437,

444, 446, 449, 463, 467, 530, 536, 559
Marked plot strip transect, 355, 362
Marking techniques, 285, 289, 507, 694
Marmots, 153, 432, 437, 587

Marsh succession, 215

Masked bobwhite, 99, 146, 411, 418, 426, 428

Matney, Iris, 727

Matrix models, 772

Maximum effort monitoring, 245

Maximum level inventory, 244

Mearn's quail, 410, 417, 418, 426

Measurement of evaporation, 615

Measurement of precipitation, 615

Measurement of snow, 615

Measures of central tendency, 743

Measures of dispersion, 743

Mergansers, 207, 221, 376-379

Minimum effort inventory, 243

Minimum effort monitoring, 245

Mining, 2, 7, 21, 143, 190, 234, 241, 309, 345-347, 580,
599, 611, 614, 633, 637, 638, 662

Mist-netting, 499, 504, 508, 512, 514

Mojave Desert, 28, 125, 127, 130, 275, 289, 471

Moles, 431, 440, 591, 636

Monitoring, 1, 2, 4-9, 11, 12, 14, 15, 17-19, 24-27, 29, 30,
32, 33, 49, 54, 61, 62, 67, 73, 74, 77-80, 84-91, 96-
99, 123, 124, 131, 138, 141, 142, 144, 145, 152, 153,
158-161, 163, 164, 166, 201, 202, 217, 218, 222,
225, 231, 233, 237, 243, 245, 251-253, 257, 265, 266,
268, 275, 286, 288, 289, 291, 292, 295, 296, 304,
305, 306, 307. 31 1, 313, 314, 315, 328, 330, 331, 339,
342-344, 348, 351, 356, 363, 369, 371, 379, 387,
390, 391, 394, 395, 401, 402, 407, 430, 434, 436, 437,
438, 439, 442, 443, 444, 447, 448, 450, 453, 476,
491, 497, 508, 516, 517, 519, 534, 538, 543, 544, 548,
550, 551, 554, 559, 585, 595, 598, 600, 601. 610-
613, 628, 632, 634, 637, 638, 639, 660-663, 6^4, 681,
695-697, 699, 707, 709, 710, 711-718, 721, 722, 724,
741, 742, 754, 756-759, 762, 765, 767-769, 775, 776,
777, 779, 782, 787, 805

Monitoring methods for lakes, 243, 252, 253

Monitoring studies, 1, 2, 4, 7, 9, 12, 19. 27, 29, 33, 54, 62.
73, 123, 124, 144, 145, 245, 265, 268, 387, 430, 442,
443, 450, 497, 699, 711, 754, 782

Mourning dove, 1 1 3, 4 1 1 , 42 1 , 423, 426, 427, 658, 659

Multivariate analysis, 28, 89, 90, 147, 311, 562, 741, 751,
754, 767

Muskrats, 204, 206, 207, 218-221, 223, 429, 431. 435, 436,

445, 446, 451, 452, 484, 576

Index

855

Nest counts, 219, 304, 381, 397, 398, 400, 404
Nest monitoring, 296, 305, 306
Nest searches, 306, 337, 339, 367
Nest site selection, 345, 385, 392, 405
Nest spot mapping, 357

Nesting, 3, 7, 9, 13, 24, 62, 80, 81, 85, 109, 110, 116-120,
130, 131, 135, 136, 138-141, 146, 157, 158, 160,
164-166, 176, 178, 179, 194, 195, 198, 204, 207, 218,
220, 222, 224, 292, 293, 296, 306, 307, 309-312,
314, 315, 326, 329-343, 345-349, 351-358, 364, 366,
367, 369, 371-376, 378, 380, 381, 383-385, 387-405,
408, 409, 411, 412, 419-421, 428, 587, 592, 594, 596,
598, 599, 691, 694, 697, 704, 762, 763, 764
Nesting surveys, 326, 329, 331-334, 340, 342, 345
Non-nesters, 391

Nonparametric statistics, 745, 750
Nutritional approach, 523-525
Ohmart, Robert D., 169, 639
Old World rats, 433
Oligotrophic, 220, 239, 243
Oligotrophic-eutrophic series, 239
Open water counts, 366
Optimal estimation, 754
Ordination analysis program, 153
Overlap indexes, 22
PATREC models (see Pattern recognition), 25, 763, 770,

772, 775
Parent material, 568, 582

Patchiness, 24, 28, 139, 147, 152, 164, 176, 178-180, 189,
311, 358, 361, 471, 473, 640, 641, 642, 646, 647, 648
Pattern recognition, 25, 28, 763, 770, 772, 775, 776
Permafrost, 151-153, 157-160, 164, 166
Permanent lentic, 131, 134, 269
Peterson, Larry, 727

Ph, 90, 196, 197, 199, 207, 208, 221, 231, 240, 246, 247,
258, 308, 311, 345, 346, 368, 369, 385, 428, 451, 525,
559, 562, 568, 573-575, 581, 582, 584, 633-638, 665,
697, 761, 789, 824
Phenological characteristics, 520, 542
Photographic counts, 363

Physical features, 23, 80, 106, 258, 292, 314, 352, 353,
372-375, 378, 409-411, 475, 482, 499, 500, 515, 523,
587-589, 595, 598, 600, 603, 760, 761, 772
Physical-chemical surveys, 246

Pikas, 104, 153, 453, 462, 463, 465, 467, 471, 472, 473
Plant communities, 22, 54, 73, 77, 79-82, 84, 85, 89, 91,
121, 122, 147, 158, 159, 161, 163, 167, 169, 173, 175,
179, 183-185, 191, 192, 195, 201, 207, 222, 223,
242, 271, 310, 417, 453, 455, 456, 468, 475, 479, 480,
482, 484, 486, 581, 583-584, 639, 660, 772
Plant community size, 174
Playas, 80, 115, 116, 588, 592, 593, 594
Plot sampling, 24
Plotless sampling, 24

Pocket gophers, 432, 486, 572, 582, 584, 591
Pocket mice, 136, 432, 569, 582
Point count method, 304. 326
Point data, 23, 593
Point-ccnter-quarter, 18, 23, 24, 145
Pool quality, 234, 603, 607, 608
Pools, 13, 131, 134-136, 145, 156, 215, 224, 227, 229,

231, 234, 269, 275, 277, 279, 360, 362, 503, 513, 607,
609, 724
Population levels, 81, 183. 32S, t24, 453. -+54, 467, 468,

470, 471, 496, 533. 540, 545, ^21. 723, 821
Population measurements. 20, 261. 271, 284, 297, 315,
342, 353, 357, 361, 364, 379, 394, 121, 433, 463, 487,

497, 504, 529, 531-534, 537, 538, 540, 542, 544,
545, 548, 551, 554, 765

Population structure, 16, 20, 24, 165, 195, 275, 348, 473,
531, 549, 555, 560

Porcupines, 433, 438, 698

Prairie chicken, 410, 415, 416, 426, 427, 492

Prairie dogs, 104, 432, 436, 437, 585, 687

Precipitation, 12, 20, 26, 123, 125, 127-129, 131, 133,
141-144, 190, 207, 239, 270, 272, 282, 284, 286, 291,
326, 329, 335, 448, 467, 470, 472, 482, 523, 543,
550-552, 568, 574, 575, 581, 603, 613-616, 623, 637,
713, 715, 716, 720, 722, 723. 753, 824

Preliminary field work, 4, 5

Presence, 12, 14, 16, 17, 21, 23, 25, 67, 73, 80, 86, 99-
101, 104, 106, 116, 134, 139-141, 151, 152, 157, 159,
161, 162, 173, 175, 179, 182, 185, 203, 215-217,
221, 258, 263, 264, 271, 274, 288, 297, 304, 314, 315,
330-332, 341, 353, 356, 357, 360, 361, 363, 364,
366, 379, 393, 395, 421-423, 425, 430, 431, 432, 433,
434, 435, 436, 437, 438, 441, 443, 447, 463, 465,
467, 487, 488, 489, 491, 497, 499, 500, 504, 512, 513,
514, 515, 525, 526, 529, 530, 546, 548, 552, 569,
573, 575, 576, 581, 593, 598, 600, 611, 639, 642, 643,
648, 685, 703, 759-764, 770, 772

Principal component analysis, 296, 656, 753

Prioritizing of objectives, 4

Probability of capture, 19, 444, 446, 448, 449, 450

Problem definition, 1, 2

Productivity, 2, 6, 16, 20, 24, 26, 84, 146, 152, 160, 164-
167, 170, 173, 177, 186, 187, 197, 207, 223, 234, 239,
247-249, 251-253, 260, 309, 311, 313, 332, 333, 340-
342, 344-348, 374, 383, 385, 426, 447, 448, 452,
494, 524, 525, 531, 533, 548, 550, 554, 559, 598, 611,
614, 637, 664, 665, 698, 711, 754, 764, 803

Productivity surveys, 340-342

Properties of water, 258, 360, 613, 633, 634, 761

Ptarmigan, 153, 154, 157, 162, 163, 165, 409, 414, 426,
428

Quality criteria, 633, 635, 638

RAIDS (Riparian/Aquatic Information Data Summary), 265

Receivers, 681, 683, 685

Recreational activities, 173, 190

Regression analysis, 659, 719, 749, 788

Relative abundance, 11, 19, 22, 24, 66, 67, 84, 185, 198,
245, 262-264, 274-276, 280, 281, 295, 304, 315, 326,
331, 337, 357, 380, 421, 422, 425, 463, 465, 486, 487,
488, 489, 492, 494, 529, 540, 759, 761, 763-765

Relative density, 291, 335, 337, 354, 356, 357, 361, 363-
366, 430, 433, 434, 437, 438, 439, 443, 444, 447, 450,
488, 489, 499, 504, 515, 529, 530, 548, 554, 648,
765, 770, 772

Renesting, 381, 391

Repeatability, 233-235, 604, 645

Riffles, 13, 134, 227, 229, 231, 607, 609

Ring-necked pheasant, 408, 409, 412, 426, 428, 582

Riparian habitat classification, 173

Riparian vegetation, 54, 68, 170, 172, 175, 184, 187-189,
192, 194-196, 225, 227, 228, 232, 236, 258, 480, 642,
650, 655, 660

Riparian zones, 80, 81, 91, 180, 198, 234, 257, 336, 772

Road counts, 326, 327, 328, 337

Roadside counts, 306, 422, 423, 427, 465

Rock, 16, 23, 64, 80, 104-106, 112, 113, 115, 116, 119,
124, 131, 134, 136-138, 14 1, 142, 153, 157, 197, 207,
229, 231, 234, 235, 242, 269, 314, 378, 384, 389,
409, 414, 432, 442, 467, 501, 502, 507, 533, 552, 568,
569, 574, 578, 579, 581, 583-584, 589, 590, 593,

856

Index

594, 595, 598, 599, 604, 608, 609, 619, 760, 831
Rock fragments, 569, 583, 609
Roosting flock counts, 363
Roosting sites, 361, 362, 484, 500, 501, 502, 503, 504,

512, 514, 515
Rope-drag nest mapping, 354, 358, 367
Ruddy ducks, 215, 375, 376, 381
Ruffed grouse, 410, 413, 414, 416, 421-423, 426-428, 698,

722, 760
Rumen analysis, 703, 706

Runs, 13, 134, 229, 231, 264, 326, 327, 440, 441, 616, 753
RUNWILD, 32, 37, 90, 120, 121, 772, 776
Ryder, Ronald A., 291
STORET, 46, 634

Salinity, 188, 204, 208, 221, 360, 575, 581, 583, 585
Sample counts of nests, 396
Sample plots, 327, 362, 363, 396, 399, 422, 441
Sample size, 5, 12, 19, 144, 295, 326, 327, 340, 449, 628,

648, 666, 745-747 750
Sampling considerations, 1 1 , 266
Sampling design, 11, 12, 28, 87, 303, 313, 344, 355, 660,

662, 675, 742, 756
Sampling methods, 5, 14, 23, 24, 66, 272, 274, 662
Sampling of desert wildlife, 144
Samson, Fred B., 475
Sand dunes, 80, 157, 486, 592-594, 760
Scaled quail, 146, 411, 418, 419, 423, 426-428
Scoping, 1, 2, 768, 769
Scoters, 376-378
Seasonal precipitation, 715
Seasonal water table, 575
Selecting specific methods, 4, 5
Sex and age ratios, 20, 524, 534, 535, 557, 559
Shannon- Weaver index, 22

Sharp-tailed grouse, 410, 413, 416, 424, 427, 428
Short, Henry L, 93

Shrews, 154, 429, 432, 440, 441, 442, 447, 487
Sight distance, 82, 84
Simpson's similarity coefficient, 21
Siren-elicited vocalization, 489
Slope, 23, 65, 80, 119, 131, 133, 141, 143, 144, 165, 166,

181, 192, 270, 271, 352, 353, 356, 360, 361, 393, 409,

415, 443, 453, 480, 482, 526, 548, 553, 573, 575,

578, 579, 587, 589-591, 594, 599, 606, 609, 614, 615,

622, 748-750, 760, 766, 787, 794
Snags, 18, 23, 24, 80-82, 85, 86, 88, 91, 101, 134, 138,

175, 176, 178, 179, 183, 190, 197, 310, 315, 484, 503,

588, 595, 596, 599, 601, 760, 761

Snow cover, 62, 153, 154, 160-162, 164-167, 258, 330,
409, 416, 487, 526, 542, 543, 550, 551, 562, 563, 724

Snow depth, 416, 470, 526, 614, 615, 616, 715-720, 722

Snow quality, 721

Snowshoe hare, 181, 427, 457, 468, 469, 470, 472, 473,
482, 492, 723

Social interactions, 345, 392, 393, 487

Soil(s), 2, 3, 5, 12, 23, 51, 52, 54, 62, 65, 66, 75, 76, 79,
80,89,93,94,99-101, 105, 106, 115, 119, 124, 125,
131, 132, 133, 135-138, 141, 142, 144, 146, 153, 157-
159, 165, 166, 170, 173, 175, 177, 178, 180, 182,
183, 186, 187, 191, 192, 199, 201, 203, 207, 213, 215,
216, 217, 220, 222, 228, 232, 234, 235, 243, 257,
258, 268-271, 276, 283, 284, 353, 389, 393, 39-4, 401,
432, 434, 439, 443, 453, 455, 491, 567-585, 588,

589, 591, 592, 594, 596, 598, 600, 604, 606, 613, 614,
616, 652, 716, 718, 720, 747, 748, 753, 756, 760,
761, 826

Soil color, 570, 572, 584

Soil consistence, 570, 571

Soil depth, 23, 137, 138, 268, 572, 576, 584

Soil drainage classes, 573

Soil fertility, 567, 575, 576, 584, 585

Soil reaction, 573

Soil structure, 570, 571, 573

Soil taxonomy, 576, 578, 585

Sonoran Desert, 123-125, 127-130, 138-140, 142-144, 147,

182, 198, 271, 274, 275
Special lake types, 239
Species composition, 22, 23, 84, 128, 158, 161, 172, 173,

177, 184, 185, 188, 190, 208, 218, 219, 263, 271, 282,

286, 296, 328, 353, 360, 372, 373, 375, 376, 378,

387, 395, 409-411, 413, 420, 424, 522, 523, 554, 640,

642, 701, 759, 760
Species diversity, 17, 20, 22, 23, 129, 130, 139, 140, 142-

144, 146, 147, 156, 183, 197, 222, 304, 308, 310, 515,

657, 660, 761, 764, 765
Specific conductance, 212, 231, 636, 638
Speich, Steven M., 387
Spot-mapping census, 300
Spowart, Richard A., 475
Spruce grouse, 409, 412, 413, 427, 428
Squirrels, 104, 136, 432, 437, 438, 439, 440, 444, 446,

451, 486, 576, 582, 591, 697, 724
Stand size, 73, 82, 84, 86, 655
Stone, James E., 567
Stream channel stability, 225, 227, 231, 232, 603, 606,

607, 614, 628, 629-630
Stream classification, 225, 226
Stream diversity, 227, 231, 234
Stream gradient, 227, 230, 258, 603, 606, 611, 666
Stream habitat features, 225, 227, 234
Stream habitat inventory, 231, 232, 612
Stream order, 226, 227
Stream water chemistry, 227, 230
Stream width and depth, 225, 230
Streambank stability, 196, 227, 231, 603, 605
Streambed, 174, 225, 227, 229-232, 234, 603, 607, 608,

611, 618, 663, 665, 779
Streambed material, 229, 607, 608
Streamflow, 227, 233, 263, 603-605, 612, 616, 619, 622-

626, 628, 632, 634
Streamflow measurement, 605, 616, 619, 623, 628
Streamflow pattern, 227, 603
Strip census, 422, 423, 427, 444, 544
Strip transect, 309, 355, 358, 362, 398, 399, 722
Structural conditions, 17, 73, 77, 82, 85, 310
Structural types, 173, 174, 177, 184, 185, 651, 652, 653,

655
Structure, 4, 12, 16, 17, 20, 21, 24, 25, 28, 62, 66, 74, 80,

81, 90, 93, 94, 96, 97, 99-101, 103, 104, 106-110, 115,

116, 119, 121, 123, 128, 131-142, 146, 147, 165,

172, 179, 181, 182, 184, 188, 189, 195, 201, 206, 207,

212, 213, 258, 268-271, 275, 286, 289, 292, 296.

303, 304, 308-312, 314, 328-330, 348, 352, 353, 359.

372-376, 378, 392-394, 404, 405, 419, 420, 424, 453.

473, 476, 480, 494, 495, 499, 500, 504, 512. 515. 520.

523, 526, 531, 534, 540, 544, 549, 551, 554, 5SS

557, 560, 568, 570, 571, 573, 574, 588, 596, 599, 614,

619, 640, 642, 652, 660, 665, 674, 699, 705, 732,

736, 739, 759, 760, 764, 772, 797
Study objectives, 1, 4, 5, 145

Submergents, 203, 213, 215, 216, 218-221, 372, 375
Subsurface chamber, 393

Index

857

Succession, 25, 68, 89, 161, 173, 196, 215, 220, 224, 304,

308, 554, 556, 644, 772, 775, 776
Successions stages, 73, 77, 80-82, 85, 91, 203, 221, 420,

772, 775
Suspended sediment, 6 14, 628, 631, 638
Suspended solids and turbidity, 230
Swans, 166, 207, 371-373, 384, 385
T-test, 747, 750

Talus, 80, 85, 90, 101, 104, 115, 116, 131, 136, 153, 157,
269, 393, 432, 438, 453, 462, 463, 465, 483, 486, 501,
502, 517, 568, 587, 592-596, 599, 760
Talus slopes, 104, 136, 157, 269, 393, 432, 453, 462, 465,

483, 501, 594
Tame animals, 547, 703, 704, 706
Tape playback response, 365, 366
TDS, (see Total dissolved solids)
Temporary lentic, 23, 131, 135, 269
Territory spot mapping, 354
Texture, 62, 80, 159, 182, 268, 271, 500, 568-570, 572,

574-576, 578, 579, 581-584, 592, 597
Thomas, Jack Ward, 73
Time series analysis, 754
Toe-pace transect, 24

Total dissolved solids, 172, 192, 231, 233, 525, 636
Total population counts, 363 , •": „*,'-' • ..

Transect methods,' 24, 302, 363,' .426:.- , ' "oV-
Transmitters, 681, 682,-683, 685;, 68'7; 689,'.691, 697, 698
Trapping, 14; 15, 100, 145V163,,181,T85, 2^9, 273, 274,
276, 279-284; 289; -306^ 307; 4jQ8,, 4 17,^430, 434, 435,
436, 437, 438, 439, '440, 441, 442, 443,'444, 446,
448, 450, 449, 452, 488, 489, 490, 504; 5,1 2, 514, 536,
556, 562, 563, 680, 686, 694, 697, 714 '
Travel cost method, 788, 790, 802
Tree height, 82, 753
Tropical-subtropical deserts, 125
Turbidity, 131, 192, 202, 220, 230, 233, 241, 261, 269,

603, 611, 634, 637, 638 711, 761
Understory volume, 82, 84
Unmarked strip transect, 355, 362
Van Haveren, Bruce P., 613
Variable-circular, 295, 305, 306, 309, 311, 366
Variable width line transect, 356, 358, 366
Vegetation dynamics, 215
Vegetation maps, 184, 330

Vegetation measurements, 2, 3, 5, 6, 12, 13, 21, 23-25, 28,
50-54, 61, 62, 65-68, 74, 76, 80-82, 84, 89, 90, 93-101,
103, U0, 115, 119, 121, 123-125, 128, 130-133, 135,
137-144, 146, 147, 149, 151-153, 155, 156, 158-167,
170-172, 174, 175, 177-197, 201-213, 215-225, 227,
228, 232, 233, 236, 241, 242, 250, 258, 261, 266, 268-
271, 276-278, 288, 291, 292, 296, 304, 308-311, 314,
327, 330, 335, 336, 338, 345, 347, 352, 353, 355-

360, 362, 364-366, 372-376, 378-380, 382, 389, 393-
395, 397, 398, 408-411, 414, 416, 419, 420, 424, 427,
429, 430, 431, 432, 433, 434, 435, 438, 439, 447,
455, 457, 460, 461, 466, 469, 470, 471, 473, 475, 477,
479, 480, 482, 483, 486, 487, 491, 503, 510, 523,
526, 527, 533, 537, 538, 540, 543-548, 550-554, 556,
558, 567, 572, 575, 580, 581, 583, 587-589, 591,
592, 594-596, 598-600, 606, 609, 614, 615, 633, 639,
640, 642, 643, 644, 646, 647, 648, 649, 650, 651, 652,
653, 654, 655, 656, 657, 659, 660, 665, 679, 687,
699, 700, 702, 708, 710, 743, 744, 753, 759-761, 764,
770, 772, 776, 779, 810, 811

Vegetative structure, 81, 292, 296, 314, 495

Vehicle tracking, 683, 687, 688

Verner, Jared, 73

Vertical diversity, 647, 649,

Vertical habitat availability, 268

Vertical vegetation structure, 139

Visual aids, 831-832, 833, 835, 836, 838, 840

Voles, 104, 429, 432, 433, 440, 451, 478, 484, 582, 585,
698

Wagoner, Ora, 29

Warm-temperate deserts, 125

Warm-water fish, 242, 258, 636, 661

Watchers, 527

Water chemistry, 155, 172, 207, 227, 230, 231, 233, 246,
633, 634, 662, 664

Water depth, 135, 155, 203, 206, 212, 216-219, 235, 241,
249, 262, 352, 360, 361, 366, 604, 605, 619

Water quality, 135, 172, 181, 187, 190, 192, 194, 196,

208, 225, 226, 236, 244, 252, 253, 258, 485, 525, 575,
612, 616, 628, 632-638, 661, 662, 665, 666, 674,
761, 818

Water temperature, 131, 172, 230, 231, 232, 241, 258,
269, 603, 610, 611, 664

Watershed stability, 61 4

Weight of fish, 262, 264

Weller, Milton W., 201

West, William H., 741

Wild turkey, 410, 4 16, 427, 428, 578

Wildlife guilds, 10, 28, 90, 96, 108-112, 116, 121, 147,
765, 776

Willner, Gale R, 453

Winter bird population study, 299, 309-311

Winter bird survey, 305, 31 1

Winter habitat, 166, 315, 330, 372, 379, 383, 384, 413,
470, 484, 550, 551, 557, 558, 767

Winter surveys, 382
Wintering habitat, 378, 382, 383
Woodrats, 431, 433, 437, 439, 572
Writing strategy, a, 806, 811
Zimmerman, Donald E., 805

£Be5$:i£ cenie*

858

Index

Q

W

U 2

H K

< D

Q H

U

c
c

<o

CD

c*.

1^.
00

o

Q.

O

BLM LIBRARY
PDEONv!°RXCO0fc0047

m -^

•i 7

U.S. Department of the Intern
Bureau of Land Management *

--■-

&'&

'Wi

_* ■■■■*■»<

jr^^pP^

y&t

. ■ ■ -

•*
<*

J

JET 'jv-.

'- 'r «*

;

</ # ■ •■■"■

£3&S

i

F*>

i£ \ fc" :

it.

v.- ''■

T

^9

jp*»?.\i+ '■