Amphibians Ecology and Conservation OK

Amphibian Ecology and Conservation
Techniques in Ecology and Conservation Series

Series Editor: William J. Sutherland
Bird Ecology and Conservation: A Handbook of Techniques

William J. Sutherland, Ian Newton, and Rhys E. Green
Conservation Education and Outreach Techniques

Susan K. Jacobson, Mallory D. McDuff, and Martha C. Monroe
Forest Ecology and Conservation: A Handbook of Techniques

Adrian C. Newton
Habitat Management for Conservation: A Handbook of Techniques

Malcolm Ausden
Conservation and Sustainable Use: A Handbook of Techniques

E.J. Milner-Gulland and J. Marcus Rowcliffe
Invasive Species Management: A Handbook of Techniques

Mick N. Clout and Peter A. Williams
Amphibian Ecology and Conservation: A Handbook of Techniques

C. Kenneth Dodd, Jr
Amphibian Ecology and
Conservation
A Handbook of Techniques
Edited by
C. Kenneth Dodd, Jr
1
3
Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© Oxford University Press, 2009
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2009
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose the same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India
Printed in Great Britain
on acid-free paper by
CPI Antony Rowe, Chippenham, Wiltshire
ISBN 978–0–19–954118–8 (Hbk.)

ISBN 978–0–19–954119–5 (Pbk.)
10 9 8 7 6 5 4 3 2 1
Preface
As this volume is completed, more than 6400 amphibian species have been
recognized, with new taxa being described nearly every day. The last few dec-
ades have seen an explosion in systematic research, particularly in the tropics.
Long-recognized centers of diversity have been explored using increasingly
sophisticated sampling techniques, yielding many new taxa. At the same time,
new centers of speciation, such as Sri Lanka and the Western Ghats of India,
have been discovered, while molecular techniques have yielded previously
unsuspected diversity within some well-known taxa, such as the plethodontid
salamanders of southeastern North America and the green toad (Bufo viridis)
complex of Eurasia. For amphibian systematists, these are exciting times.
Unfortunately, amphibians are now at greater peril than at any time in recent
geologic history, a situation chronicled in two recent data-rich books (Lannoo
2005; Stuart et al. 2008). Habitats are being lost at alarming rates because of
expanding human populations and generally favorable economic conditions
fostering development; emerging infectious diseases, particularly amphibian
chytrid fungus (Batrachochytrium dendrobatidis), threaten worldwide impacts;
non-indigenous species proliferate, affecting amphibians and their habitats; and
amphibians, with their permeable skins, diverse life histories, and often biphasic
life cycles requiring both terrestrial and aquatic habitats, are being saturated by
a host of lethal and sublethal toxic substances. New threats, such as the effects
of global climate change, further imperil amphibians, especially those with
limited distributions and dispersal capabilities. Fully one-third of all amphib-
ians are now considered threatened (Stuart et al. 2004), and 168 species have
become extinct within the last two decades. Clearly, these are treacherous times
for many frogs, salamanders, and caecilians.
Amphibians are, quite frankly, engaging animals. Despite Linnaeus’ early
characterization of amphibians in the context of “Terrible are thy works, O
God”, biologists have come to appreciate that their diverse life histories and
shear numbers offer a wealth of material for research on basic ecological princi-
ples, such as trophic interactions, phenotypic plasticity, predator–prey interac-
tions, community structure, mate choice and recognition, water balance, and
many others. In response to threats, conservation biologists have probed these
and other questions in hopes of understanding amphibian biology in order to
prevent declines and extinctions. The basic and applied themes of biology merge
vi | Preface
in these disciplines: understanding ecology leads to conservation options (see

Gascon et al. 2007), and conservation-based research leads to a better appreci-
ation of ecological principles.
To say that there are a great many techniques available in ecological and
conservation-based research on amphibians is an understatement. The pages
of journals such as Herpetological Review and Applied Herpetology contain tech-
niques papers with every issue. Specialized books, such Heyer et al. (1994),
Henle and Veith (1997), and Gent and Gibson (1998), offer additional summar-
ies that are as applicable today as when they were published. No one volume can
include all techniques. The current volume is meant not to supplant these earlier
works, but to supplement them and add new areas not previously summarized,
such as occupancy modeling, landscape ecology, genetics, telemetry, and dis-
ease biosecurity. Our objectives have been to delineate important new develop-
ments, to give an idea as to what the techniques tell or do not tell a researcher,
to focus attention on biases and data inference, and to get readers to appreciate
sampling as an integral part of their science, rather than just a means of captur-
ing animals. The techniques used will set the boundaries within which results
can or should be interpreted.
As noted earlier, amphibian systematics is a flourishing field, with many
new opportunities made available by combining large datasets using molecular
and morphological data with powerful computer analysis. The phylogeny of
amphibians is undergoing increasingly sophisticated analysis. Some analyses,
such as those of Frost et al. (2006), suggest relationships that differ substantially
from “traditional” concepts. If accepted, extensive nomenclatural changes will
be warranted. Although Frost et al. (2006) have advocated substantive changes
in nomenclature, many of which will likely be accepted with further study, other
amphibian biologists disagree with automatically accepting every change pro-
posed by these authors. In this book, I have decided to retain the older nomen-
clature rather than make a taxonomic decision each time a name is mentioned.
The intended audience of this volume (biologists starting their careers in ecology
and conservation) likely will be more familiar with the older generic names of
Bufo, Rana, and Hyla, and initially may be confused by the unfamiliar replace-
ment names. Readers should be aware, however, that names such as Lithobates
( Rana, in part), Anaxyrus ( Bufo, in part), and others currently unfamiliar
may soon be more commonplace.
I wish to thank the following for taking their valuable time to review manu-
scripts and offer suggestions for improving this volume: James Austin, Larissa
Bailey, Bruce Bury, Dan Cogălniceanu, Sarah Converse, Steve Corn, Rafael
Ernst, Alisa Gallant, Marian Griffey, Kerry Griffis-Kyle, Richard Griffiths,
Preface | vii
Margaret Gunzburger, Tibor Hartel, Robert Jehle, Steve Johnson, Y.-C. Kam,
Sarah Kupferberg, Frank Lemckert, Harvey Lillywhite, John Maerz, Joseph
Mitchell, Clinton Moore, Erin Muths, James Petranka, Benedikt Schmidt,
Ulrich Sinsch, Kevin Smith, Lora Smith, Joseph Travis, Susan Walls, and
Matthew Whiles. I greatly appreciate the support from Ian Sherman and Helen
Eaton at Oxford University Press and editorial help from freelance copy-editor
Nik Prowse, and thank series editor, Bill Sutherland, for inviting me to edit the
amphibian volume. This volume is dedicated to all the biologists who take up
the challenge of amphibian ecology and conservation.
C. Kenneth Dodd, Jr
References
Frost, D. R., Grant, T., Faivovich, J., Bain, R. H., Haas, A., Haddad, C. F. B., de Sá, R. O.,
Channing, A.,Wilkinson, M., Donnellan, S. C. et al. (2006). The amphibian tree of
life. Bulletin of the American Museum of Natural History, 297, 1–370.
Gascon, C., Collins, J. P., Moore, R. D., Church, D. R., McKay, J. E., and Mendelson,
III, J. R. (eds) (2007). Amphibian Conservation Action Plan. IUCN/SSC Amphibian
Specialist Group, Gland and Cambridge.
Gent, T. and Gibson, S. (eds) (1998). Herpetofauna Worker’s Manual. Joint Nature
Conservation Committee, Peterborough.
Henle, K. and Veith, M. (eds) (1997). Naturschutzrelevante Methoden der Feldherpetologie.
Mertensiella 7.
Heyer, W. R., Donnelly, M. A., McDiarmid, R. W., Hayek, L.-A., and Foster, M. S.
(eds). (1994). Measuring and Monitoring Biological Diversity. Standard Methods for
Amphibians. Smithsonian Institution Press, Washington DC.
Lannoo, M. J. (ed) (2005). Amphibian Decline. The Conservation Status of United States
Species. University of California Press, Berkeley, CA.
Stuart, S., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues, A. S. L., Fishman, D. L.,
and Waller, R. W. (2004). Status and trends of amphibian declines and extinctions
worldwide. Science, 306, 1783–6.
Stuart, S., Hoffman, M., Chanson, J. S., Cox, N. A., Berridge, R. J., Ramani, P., and
Young, B. E. (eds) (2008). Threatened Amphibians of the World. Lynx Edicions,
Barcelona; IUCN, Gland; Conservation International, Arlington, VA.
This page intentionally left blank
Contents
List of contributors xxv
Part 1 Introduction 1
1 Amphibian diversity and life history 3
Martha L. Crump
1.1 Introduction 3
1.2 Amphibian species richness and distribution 3
1.3 Amphibian lifestyles and life history diversity 5
1.3.1 Caecilians 6
1.3.1.1 Aquatic 7
1.3.1.2 Combination aquatic and terrestrial 7
1.3.1.3 Terrestrial and/or fossorial 7
1.3.2 Salamanders 7
1.3.2.1 Aquatic 8
1.3.2.2 Combination of aquatic and terrestrial 8
1.3.2.3 Terrestrial 9
1.3.2.4 Fossorial 10
1.3.2.5 Arboreal 10
1.3.3 Anurans 10
1.3.3.1 Aquatic/aquatic 10
1.3.3.2 Terrestrial/aquatic 11
1.3.3.3 Arboreal/aquatic 13
1.3.3.4 Fossorial/aquatic 14
1.3.3.5 Terrestrial/non-aquatic 14
1.3.3.6 Arboreal/non-aquatic 14
1.3.3.7 Fossorial/non-aquatic 15
1.4 Amphibian declines and why they matter 15
1.4.1 Economics 16
1.4.2 Ecosystem function 16
1.4.3 Esthetics 17
1.4.4 Ethics 17
1.5 References 17
2 Setting objectives in field studies 21

Dan Cogălniceanu and Claude Miaud
2.1 Basic concepts for a good start 21
x | Contents
2.2 Steps required for a successful study 24

2.2.1 Temporal and spatial scales 24
2.2.2 Choosing the model species 25
2.2.3 Pilot/desk study 26
2.2.4 Elaborate a conceptual model 26
2.2.5 The SMART approach 27
2.2.6 Applying the SMART approach to plan an
amphibian inventory 28
2.2.7 Experimental versus field studies 29
2.2.8 Methods for sampling, data storage, and analysis 29
2.3 Trade-offs and pitfalls 30
2.4 Ethical issues 31
2.5 Acknowledgments 32
2.6 References 32
Part 2 Larvae 37
3 Morphology of amphibian larvae 39
Roy W. McDiarmid and Ronald Altig
3.1 Background 39
3.2 Larval caecilians 40
3.2.1 Morphology and ontogeny 40
3.2.2 Coloration 42
3.2.3 Diversity 42
3.3 Larval and larviform salamanders 42
3.3.2 Coloration 44
3.3.3 Diversity 44
3.4 Anuran tadpoles 45
3.4.2 Coloration 49
3.4.3 Diversity 49
3.5 Summary 50
3.6 References 51
4 Larval sampling 55
David K. Skelly and Jonathan L. Richardson
4.1 Introduction 55
4.1.1 Why sample larvae? 55
4.1.2 Target responses 56
4.1.3 Timing 57
4.1.4 Sampling effort 57
4.2 Sampling techniques 58
4.2.1 Box/pipe sampler 58
Contents | xi
4.2.1.1 Description 58
4.2.1.2 Application 60
4.2.1.3 Considerations 60
4.2.2 Dip net 60
4.2.3 Seine 61
4.2.4 Leaf litterbags 62
4.2.5 Trapping 64
4.2.6 Mark–recapture 65
4.3 Other techniques 66
4.3.1 Bottom net 67
4.3.2 Electroshocking 67
4.3.3 Visual encounter survey 67
4.4 Conclusions 67
4.6 References 68
5 Dietary assessments of larval amphibians 71

Matt R. Whiles and Ronald Altig
5.1 Background 71
5.2 Larval caecilians and salamanders 72
5.3 Anuran tadpoles 72
5.4 Assessing food sources and diets 73
5.4.1 Category I: preparatory studies 73
5.4.2 Category II: gut contents 74
5.4.3 Procedures: anurans and small predators 74
5.4.4 Processing and analysis of gut contents 75
5.5 Category III: assimilatory diet 78
5.5.1 Stable-isotope analysis 78
5.5.2 Procedures 80
5.5.3 Fatty acid analyses 81
xii | Contents
5.6 Summary 82
5.7 References 83
6 Aquatic mesocosms 87
Raymond D. Semlitsch and Michelle D. Boone
6.1 Introduction 87
6.2 Historical background 88
6.3 Why use mesocosms? 90
6.4 Types of mesocosm 93
6.5 Setting up mesocosms 95
6.6 Common experimental designs 96
6.7 Case studies 98
6.7.1 Community ecology 98
6.7.2 Evolutionary ecology 99
6.7.3 Ecotoxicology 100
6.7.4 Land use and management 100
6.8 Conclusion 101
6.9 References 102
7 Water-quality criteria for amphibians 105

Donald W. Sparling
7.1 Introduction 105
7.2 Dissolved oxygen 105
7.3 Temperature 108
7.4 pH 110
7.5 Conductivity, hardness, and salinity 112
7.6 Total and dissolved organic carbon 113
7.7 Pollutants 114
7.7.1 Fertilizers and nitrogenous compounds 114
7.7.2 Pesticides 114
7.7.3 Metals 115
7.7.4 Organic pollutants and halogenated hydrocarbons 115
7.7.5 Pharmaceuticals 116
7.8 Summary and conclusions 116
7.9 References 117
Part 3 Juveniles and adults 121

8 Measuring and marking post-metamorphic
amphibians 123
John W. Ferner
8.2 Toe-clipping 125
Contents | xiii
8.2.1 Anurans 125

8.2.2 Salamanders 129
8.2.3 Ethical issues related to toe-clipping of amphibians 129
8.3 Branding 130
8.3.1 Anurans 130
8.4 Tagging and banding 131
8.4.1 Anurans 131
8.4.3 Caecilians 133
8.5 Trailing devices 133
8.6 Pattern mapping 134
8.6.1 Anurans 134
8.7 Passive integrated transponder (PIT) tags 136
8.7.1 Anurans 137
8.7.3 Caecilians 137
8.8 Taking measurements 138
8.9 Recommendations 138
8.10 References 139
9 Egg mass and nest counts 143

Peter W. C. Paton and Reid N. Harris
9.1 Background: using egg mass and nest counts
to monitor populations 143
9.2 Oviposition strategies 144
9.3 Egg-mass counts 145
9.4 Amphibian nests and nest counts 147
9.5 Clutch characteristics 150
9.6 Spatial distribution of eggs 152
9.7 Breeding phenology 152
9.8 Number of surveys needed 153
9.9 Estimating egg-mass detection probabilities 154
9.10 Variation in counts among observers 154
9.11 Marking eggs 155
9.12 Situations in which nest counts are not practical 155
9.12.1 Nest destruction 155
9.12.2 Desertion of attendant 155
9.13 How to count eggs in a nest 156
9.14 Estimating hatching success 156
9.15 Analysis of egg-mass count data 157
9.16 Summary 157
9.17 References 158
xiv | Contents
10 Dietary assessments of adult amphibians 167

Mirco Solé and Dennis Rödder
10.1.1 Adult anurans 167
10.1.2 Adult caudata 168
10.1.3 Adult gymnophiona 169
10.2 Methods to obtain prey items: historical overview 169
10.2.1 Stomach flushing: materials 170
10.2.2 When should sampling be conducted? 171
10.2.3 How to perform the flush 171
10.3 Data analysis 172
10.3.1 Measuring prey availability and electivity 174
10.3.2 Comparing trophic niche structure 176
10.3.3 Fatty acid and stable-isotope analyses 177
10.3.4 Further biochemical analyses 178
10.4 General considerations 178
10.4.1 Ontogenetic changes 179
10.4.2 Seasonal changes 179
10.5 Conclusions 180
10.7 References 180
11 Movement patterns and radiotelemetry 185

Dale M. Madison, Valorie R. Titus, and Victor S. Lamoureux
11.2 Equipment 186
11.2.1 Receivers and antennas 186
11.2.2 Transmitters 186
11.2.2.1 External transmitters 187
11.2.2.2 Internal transmitters 188
11.3 Surgical techniques 189
11.3.1 Surgery 189
11.3.2 Recovery and healing 191
11.4 Tracking procedures 191
11.4.1 Animal release 191
11.4.2 Locating signal source 192
11.4.3 Data collection 193
11.5 Analysis of movement data 193
11.6 Validation of telemetry procedures 197
11.6.1 Internal condition and mass 197
11.6.2 Movements 198
11.6.3 Reproduction 198
11.6.4 Injury and survivorship 199
Contents | xv

11.8 References 200
12 Field enclosures and terrestrial cages 203

Elizabeth B. Harper, Joseph H.K. Pechmann, and
James W. Petranka
12.1 Introduction: amphibians in the terrestrial environment 203
12.2 What are the purposes of terrestrial enclosures? 204
12.3 Defining the research question 205
12.3.1 Questions related to habitat types 205
12.3.2 Questions with treatments that can be assigned
within enclosures 207
12.4 Constructing enclosures 208
12.4.1 Location of enclosures 208
12.4.2 Size and number of enclosures 209
12.4.3 Building enclosures that minimize escapes and
trespasses 216
12.4.4 “Standardizing” conditions among enclosures 218
12.5 Study species 218
12.5.1 Choice of species 218
12.5.2 Source and age of animals 219
12.6 Census techniques 219
12.6.1 Individual marks 219
12.6.2 Methods of capture 220
12.6.3 Frequency of censuses 220
12.7 Response metrics 221
12.7.1 Vital rates: survival, growth, age at reproductive
maturity, fecundity 221
12.7.2 Physiological responses 222
12.7.3 Behavioral responses 222
12.8 Thinking outside the box 223
12.9 References 223
Part 4 Amphibian populations 227

13 Drift fences, coverboards, and other traps 229
John D. Willson and J. Whitfield Gibbons
13.2 Drift fences, funnel traps, and other passive
capture methods 229
13.2.1 What are passive traps? 229
13.2.2 How are passive traps constructed, aligned,
and monitored? 231
xvi | Contents
13.2.3 What can passive traps tell you?

What can they not tell you? 235
13.3 Coverboards and other traps that require active capture 236
13.3.1 What are coverboards and other active traps? 236
13.3.2 How are active traps constructed, aligned, and
monitored? 237
13.3.3 What can active traps tell you? What can they
not tell you? 240
13.4 References 241
14 Area-based surveys 247

David M. Marsh and Lillian M.B. Haywood
14.1 Introduction: what are area-based surveys? 247
14.1.2 Why use area-based surveys? 247
14.2 Kinds of area-based survey 249
14.3 Specific examples of area-based surveys 249
14.3.1 Leaf-litter plots 249
14.3.2 Natural-cover surveys for terrestrial salamanders 250
14.3.3 Nocturnal transects 251
14.3.4 Quadrats for stream amphibians 252
14.3.5 Soil quadrats for caecilians 252
14.4 Modifications 253
14.4.1 Distance sampling 253
14.4.2 Adaptive cluster sampling 253
14.5 Design issues: choice of sampling unit 254
14.5.1 Design issues: how many replicates? 255
14.5.2 Reducing variation among replicates 255
14.5.3 How many times to survey each replicate? 256
14.6 An example of study design 257
14.7 Assumptions of area-based surveys 259
14.8 Summary and recommendations 260
14.9 References 260
15 Rapid assessments of amphibian diversity 263

James R. Vonesh, Joseph C. Mitchell, Kim Howell, and
Andrew J. Crawford
15.1 Background: rapid assessment of amphibian diversity 263
15.1.1 When is an RA needed? 264
15.2 Planning an RA 265
15.2.1 Developing objectives 265
15.2.2 Costs and funding 266
15.2.3 Team selection and training 266
15.2.4 Permits 267
Contents | xvii
15.2.5 Data management 267

15.2.6 Developing the sampling plan 268
15.2.6.1 Selecting sampling sites 268
15.2.6.2 Selecting sampling time 270
15.2.6.3 Selecting sampling techniques 270
15.2.6.3.1 Visual encounter surveys (VESs) 270
15.2.6.3.2 Assumptions and limitations of VESs 272
15.2.6.4 Determining what data to collect 273
15.2.6.5 Field logistics 274
15.3 In the field 274
15.4 Compiling data and interpreting results 274
15.5 Recommendations and reporting 275
15.6 Summary 275
15.7 References 276
16 Auditory monitoring of anuran populations 281

Michael E. Dorcas, Steven J. Price, Susan C. Walls, and
William J. Barichivich
16.2 MCS 281
16.2.1 History and current status of MCS 282
16.2.2 Study objectives: what can MCS tell us? 283
16.2.3 Survey design 283
16.2.4 Other survey-design issues to consider:
the efficiency of MCS 284
16.2.5 Limitations of MCS data 287
16.3 ARS 288
16.3.1 Sources for ARS 289
16.3.2 Construction, deployment, and retrieval of data 289
16.3.3 Monitoring environmental data 292
16.3.4 Answering questions using ARS 292
16.3.5 Limitation of ARS 294
16.6 References 295
17 Measuring habitat 299

Kimberly J. Babbitt, Jessica S. Veysey, and George W. Tanner
17.2 Habitat selection 299
17.3 Spatial and temporal scale 300
17.4 Approaches for examining habitat selection 300
17.5 Determining availability 301
xviii | Contents
17.6 What to measure 302

17.7 Weather variables 302
17.8 Aquatic habitat 304
17.9 Physical habitat variables 304
17.10 Chemical variables 307
17.11 Measuring vegetation 309
17.11.1 Tree (overstory) measurements 310
17.11.2 Shrub (midstory) measurements 312
17.11.3 Ground (understory) measurements 313
17.12 Edaphic features 314
17.13 Conclusion 314
17.14 References 315
Part 5 Amphibian communities 319

18 Diversity and similarity 321
C. Kenneth Dodd, Jr
18.2 Data transformation 322
18.3 Species diversity 323
18.3.1 Sampling considerations 323
18.3.2 Species richness 324
18.3.3 Species accumulation curves 325
18.3.4 Heterogeneity 327
18.3.5 Evenness and dominance 329
18.4 Similarity 329
18.5 Software 331
18.6 Summary 334
18.7 References 334
19 Landscape ecology and GIS methods 339

Viorel D. Popescu and James P. Gibbs
19.1.1 Relevance of landscape ecology to amphibian
biology and conservation 339
19.1.2 Defining a “landscape” from an amphibian’s
perspective 341
19.1.3 GIS 342
19.2 Applications of spatial data for amphibian conservation 346
19.2.1 Multiscale predictors of species occurrence and
abundance 346
19.2.2 Landscape thresholds 347
19.2.3 Connectivity/isolation in amphibian populations 349
Contents | xix
19.2.4 Landscape permeability 351

19.2.5 Landscape genetics 352
19.3 Spatial statistics 353
19.4 Limitations and future directions 354
19.5 References 356
Part 6 Physiological ecology and genetics 361

20 Physiological ecology: field methods and
perspective 363
Harvey B. Lillywhite
20.1.1 Important features of amphibians 364
20.2 Heat exchange, body temperature, and thermoregulation 365
20.2.1 Thermal acclimation of physiological function 365
20.2.2 Early methods in field studies of amphibian
thermal relations 366
20.2.3 Precautions for temperature measurements 367
20.2.4 Telemetry 371
20.3 Water relations 372
20.3.1 Cutaneous water exchange: terrestrial environments 372
20.3.2 Cutaneous water exchange: aquatic environments 374
20.4 Measuring water exchange 374
20.5 Energetics 375
20.6 Modeling amphibian–environment interactions 376
20.6.1 Mathematical models 376
20.6.2 Physical models 377
20.7 Other issues and future research directions 379
20.8 References 382
21 Models in field studies of temperature and

moisture 387
Jodi J. L. Rowley and Ross A. Alford
21.1.1 The importance of understanding the temperature
and moisture environments of amphibians 387
21.1.2 Physical models of amphibians 389
21.2 Field models for investigating temperature and moisture 390
21.2.1 Models with zero resistance to EWL 390
21.2.2 A system of models that allows for variable EWL 390
21.2.2.1 Rationale 390
21.2.2.2 Model design and construction 391
xx | Contents
21.2.2.3 Example of model construction and

validation in the laboratory 392
21.2.2.4 Field validation of use of models to
characterize available temperatures 395
21.2.2.5 The importance of matching the spectral
properties of models to living animals 399
21.2.2.6 Models as indicators of relative moisture
availability of microenvironments 399
21.3 Summary and future developments 402
21.4 References 403
22 Genetics in field ecology and conservation 407

Trevor J. C. Beebee
22.1 Background: the importance of genetics in ecology
and conservation 407
22.2 Molecular methods for investigating amphibian
populations 408
22.2.1 Laboratory facilities 408
22.2.2 Sampling 409
22.2.3 DNA extraction 409
22.2.3.1 Proteinase K/phenol/chloroform method 410
22.2.3.2 Kit-based DNA extraction 410
22.2.3.3 Chelex method 410
22.2.3.4 FTA cards 410
22.2.3.5 Summary of DNA-extraction methods 410
22.2.4 Quantifying DNA recoveries 411
22.2.5 The basis of PCR analysis 411
22.2.6 Choice of analytical methods 412
22.2.6.1 Mitochondrial DNA (mtDNA) 412
22.2.6.2 Random amplification of polymorphic
DNA (RAPD) and amplified fragment
length polymorphism (AFLP) analyses 413
22.2.6.2.1 RAPD analysis 413
22.2.6.2.2 AFLP analysis 414
22.2.6.2.3 Microsatellite analysis 414
22.3 Analysis of genetic data 417
22.3.1 Cryptic species or life-stage identification 417
22.3.2 Genetic diversity, inbreeding, and bottlenecks 417
22.3.3 Identification of barriers to movement 419
22.3.4 Defining populations and population size 420
22.3.5 Historical issues 421
22.3.6 Behavior and sexual selection 423
22.4 Future developments 423
22.5 References 425
Contents | xxi
Part 7 Monitoring, status, and trends 429

23 Selection of species and sampling areas:
the importance to inference 431
Paul Stephen Corn
23.2 Sampling 433
23.3 Study sites and consequences of convenience sampling 435
23.4 Abundance and inference 439
23.7 References 443
24 Capture–mark–recapture, removal sampling,

and occupancy models 447
Larissa L. Bailey and James D. Nichols
24.2 Estimating amphibian population size and vital rates 448
24.2.1 Marking: tag type and subsequent encounter 448
24.2.2 Capture–mark–recapture 449
24.2.3 Removal sampling 455
24.3 Estimating amphibian occupancy and vital rates 455
24.3.1 Occupancy estimation 456
24.3.2 Estimation of occupancy vital rates:
extinction and colonization 457
24.4 Summary and general recommendations 458
24.5 Disclaimer 459
24.6 References 459
25 Quantifying abundance: counts, detection

probabilities, and estimates 465
Benedikt R. Schmidt and Jérôme Pellet
25.1 Background: imperfect detection in amphibian
ecology and conservation 465
25.2 Imperfect detection 466
25.2.1 Counts underestimate abundance 467
25.2.2 Per-visit and cumulative detection probabilities 468
25.2.3 Temporal and spatial variation in
detection probabilities 469
25.3 Components of imperfect detection 470
25.4 How to deal with imperfect detection 471
xxii | Contents
25.4.1 Estimation of abundance 471

25.4.2 Other approaches to dealing with
imperfect detection 473
25.5 Designing a sampling protocol 476
25.6 Software 476
25.7 Outlook 477
25.8 References 477
26 Disease monitoring and biosecurity 481

D. Earl Green, Matthew J. Gray, and Debra L. Miller
26.2 Amphibian diseases of concern 482
26.2.1 Infectious diseases 482
26.2.2 Parasitic diseases 487
26.2.3 Toxins 487
26.3 Disease monitoring: detection and diagnosis 487
26.3.1 Disease surveillance 487
26.3.2 Sample size 490
26.3.3 Sample collection and shipment 490
26.3.4 Diagnostics 494
26.4 Biosecurity: preventing disease transmission 497
26.4.1 Human and animal safety 497
26.4.2 Washing and disinfecting equipment 498
26.4.3 Movement of animals and disease management 499
26.6 References 501
27 Conservation and management 507

C. Kenneth Dodd, Jr
27.1.1 Statutory protection 508
27.1.2 Protecting habitats 508
27.2 Managing amphibian populations 509
27.3 Wetland breeding sites 510
27.3.1 Wetland integrity and hydroperiod 510
27.3.2 Wetland creation and restoration 512
27.3.3 Core habitat and buffer zones 514
27.3.4 Vegetation structure, composition, and
canopy cover 516
27.3.5 Water quality 516
27.4 Terrestrial habitats 517
27.4.1 Contiguous habitats and edge effects 517
27.4.2 Silviculture 517
Contents | xxiii
27.4.3 Restoring degraded lands 518

27.5 Migratory and dispersal routes 519
27.5.1 Corridors between habitat fragments 519
27.5.2 Crossing transportation corridors 519
27.6 Intensive manipulation of individuals 521
27.6.1 Captive breeding 521
27.6.2 Relocation, repatriation, translocation (RRT) 521
27.6.3 Disease and biosecurity 522
27.7 Conclusion 523
27.8 References 523
Index 529
Contributors
Ross A. Alford School of Marine and Tropical Biology, James Cook University,
Townsville, Queensland 4811, Australia. E-mail: [email protected]
Ronald Altig Department of Biological Sciences, Mississippi State University,
Mississippi State, Mississippi 39762-5759, USA. E-mail: [email protected]
Kimberly J. Babbitt Department of Natural Resources and the Environment, University
of New Hampshire, Durham, NH 03824, USA. E-mail: [email protected]
Larissa L. Bailey Department of Fish, Wildlife and Conservation Biology, Colorado
State University, Fort Collins, CO 80523, USA. E-mail: [email protected]
William J. Barichivich Florida Integrated Science Center, US Geological Survey, 7920
NW 71st Street, Gainesville, FL 32653, USA. E-mail: [email protected]
Trevor J. C. Beebee School of Life Sciences, University of Sussex, Falmer, Brighton
BN1 9QG, UK. E-mail: [email protected]
Michelle D. Boone Department of Zoology, 212 Pearson Hall, Miami University,
Oxford, Ohio 45056, USA. E-mail: [email protected]
Andrew J. Crawford Smithsonian Tropical Research Institute, Apartado Postal
0843-03092, Balboa, Ancón, Republic of Panama. E-mail: [email protected]
Dan Cogălniceanu University Ovidius Constant‚a, Faculty of Natural Sciences, Bvd.
Mamaia 124, Constant‚a, Romania. E-mail: [email protected]
Paul Stephen Corn U.S. Geological Survey, Aldo Leopold Wilderness Research
Institute, 790 E. Beckwith Ave., Missoula, MT 59801, USA. E-mail: [email protected]
Martha L. Crump Department of Biological Sciences, Northern Arizona University,
Flagstaff, Arizona 86011, USA. E-mail: marty.crump@ nau.edu
C. Kenneth Dodd, Jr Department of Wildlife Ecology and Conservation, University
of Florida, Gainesville, FL 32611, USA. E-mail: [email protected]
Michael E. Dorcas Department of Biology, Davidson College, Davidson, NC 28035,
USA. E-mail: [email protected]
John W. Ferner Department of Biology, Thomas More College, Crestview Hills,
Kentucky 41017, USA. E-mail: [email protected]
J. Whitfield Gibbons Savannah River Ecology Laboratory, P.O. Drawer E, Aiken,
South Carolina 29802, USA. E-mail: [email protected]
James P. Gibbs Department of Environmental and Forest Biology, State University
of New York, Syracuse, NY 13210, USA. E-mail: [email protected]
Matthew J. Gray Center for Wildlife Health, Department of Forestry, Wildlife and
Fisheries, University of Tennessee, 274 Ellington Plant Sciences Building, Knoxville,
TN 37996, USA. E-mail: [email protected]
xxvi | Contributors
D. Earl Green U.S. Geological Survey, National Wildlife Health Center, 6006
Schroeder Drive, Madison, WI 53711, USA. E-mail: [email protected]
Elizabeth B. Harper State University of New York, College of Environmental
Sciences and Forestry, 1 Forestry Drive, Syracuse, New York 13210, USA. E-mail:
[email protected]
Reid N. Harris Department of Biology, James Madison University, Harrisonburg,
Virginia 22807, USA. E-mail: [email protected]
Lillian M. B. Haywood Department of Biology, Washington and Lee University,
Lexington, VA 24450, USA. E-mail: [email protected]
Kim Howell Department of Zoology and Wildlife Conservation, PO Box 35064,
University of Dar es Salaam, Dar es Salaam, Tanzania. E-mail: [email protected]
Victor S. Lamoureux Binghamton University, PO Box 6000, Binghamton, New
York 13902, USA. E-mail: [email protected]
Harvey B. Lillywhite Department of Zoology, University of Florida, Gainesville, FL
32611, USA. E-mail: [email protected]fl.edu
Dale M. Madison Binghamton University, PO Box 6000, Binghamton, New York
13902, USA. E-mail: [email protected]
David M. Marsh Department of Biology, Washington and Lee University, Lexington,
VA 24450, USA. E-mail: [email protected]
Roy W. McDiarmid Patuxent Wildlife Research Center, US Geological Survey, National
Museum of Natural History, Washington DC 20036, USA. E-mail: [email protected]
Claude Miaud Laboratoire d’ Alpine UMR CNRS 5553, Université de Savoie, 73376
Le Bourget-du-Lac, France. E-mail: [email protected]
Debra L. Miller Veterinary Diagnostic and Investigational Laboratory, The University
of Georgia, College of Veterinary Medicine, 43 Brighton Road, Tifton, GA 31793,
USA. E-mail: [email protected]
Joseph C. Mitchell Mitchell Ecological Research Service, LLC, PO Box 5638,
Gainesville, FL 32627–5638, USA. E-mail: [email protected]
James D. Nichols U.S. Geological Survey, Patuxent Wildlife Research Center, 12100
Beech Forest Road, Laurel, MD 20708, USA. E-mail: [email protected]
Peter W. C. Paton Department of Natural Resources, University of Rhode Island,
Kingston, Rhode Island 02881, USA. E-mail: [email protected]
Joseph H. K. Pechmann Department of Biology, 132 Natural Sciences Building,
Western Carolina University, Cullowhee, North Carolina 28723, USA. E-mail:
[email protected]
Jérôme Pellet A. Mailbach Sàrl, CP 99, Ch. de la Poya 10, CH-1610 Oron-la-Ville,
Switzerland. E-mail : [email protected]
James W. Petranka Department of Biology, University of North Carolina at Asheville,
Asheville, North Carolina 28804, USA. E-mail: [email protected]
Contributors | xxvii
Viorel D. Popescu Department of Wildlife Ecology, University of Maine, Orono,

ME 04469, USA. E-mail: [email protected]
Steven J. Price Department of Biology, Wake Forest University, Winston-Salem, NC
Jonathan L. Richardson School of Forestry and Environmental Studies, Yale
University, 370 Prospect Street, New Haven, Connecticut 06511, USA. E-mail:
[email protected]
Dennis Rödder Zoological Research Museum Alexander Koenig, Adenauerallee 160,
D-53113, Bonn, Germany and Faculty of Geography/Geosciences, Trier University,
Wissenschaftspark Trier-Petrisberg, Am Wissenschaftspark 25–27, D-54286 Trier,
Germany. E-mail: [email protected]
Jodi J. L. Rowley School of Marine and Tropical Biology, James Cook University,
Townsville, Queensland 4811, Australia. E-mail: [email protected]
Benedikt R. Schmidt Zoologisches Institut, Universität Zürich, Winterthurerstrasse
190, CH-8057 Zürich, Switzerland. E-mail: [email protected]
Raymond D. Semlitsch Division of Biological Sciences, 105 Tucker Hall, University
of Missouri, Columbia, Missouri 65211, USA. E-mail: [email protected]
David K. Skelly School of Forestry and Environmental Studies, Yale University, 370
Prospect Street, New Haven, Connecticut 06511, USA. E-mail: [email protected]
Mirco Solé Department of Biology, Universidade Estadual de Santa Cruz, Rodovia
Ilhéus – Itabuna, km 16, Ilhéus, Bahia, Brazil. E-mail: [email protected]
Donald W. Sparling Cooperative Wildlife Research Laboratory, Life Science II,
MS6504, Southern Illinois University, Carbondale, Illinois 62901, USA. E-mail:
[email protected]
George W. Tanner Department of Wildlife Ecology and Conservation, University of
Florida, Gainesville, FL 32611, USA. E-mail: tannerg@ufl.edu
Valorie R. Titus Binghamton University, PO Box 6000, Binghamton, New York
Jessica S. Veysey Department of Natural Resources and the Environment, University
of New Hampshire, Durham, NH 03824, USA. E-mail: [email protected]
James R. Vonesh Department of Biology, Virginia Commonwealth University, 1000
West Cary Street, Richmond, VA 23284, USA. E-mail: [email protected]
Susan C. Walls Florida Integrated Science Center, US Geological Survey, 7920 NW
71st Street, Gainesville, FL 32653, USA. E-mail: [email protected]
Matt R. Whiles Department of Zoology and Center for Ecology, Southern Illinois
University, Carbondale, Illinois 62901-6501, USA. E-mail: [email protected]
John D. Willson Savannah River Ecology Laboratory, P.O. Drawer E, Aiken, South
Carolina 29802, USA. E-mail: [email protected]
Part 1
Introduction
1
Amphibian diversity and life history
Martha L. Crump
When the first crossopterygian crawled out of the rich Devonian waters and cast the
first envious vertebrate gaze at the terrestrial world, a boundless empire awaited colon-
ization. Although the change from an ungainly lobe-finned locomotion to a terrestrial
walking gait . . . was agonizingly slow, generations succeeded generations, archotypes [sic]
gave way to new evolutionary experiments, and the land became the home for the first
quadrupeds—the amphibians.
William E. Duellman (1970)
1.1 Introduction
Over the past 350 million years, amphibian descendents of lobe-finned fishes
have radiated into most habitats on Earth. In doing so, they have acquired spec-
tacular and sometimes bizarre physiological, morphological, behavioral, and
ecological attributes that mold their innovative life histories. Amphibians have
highly permeable skin, which makes them both vulnerable to losing water and
able to absorb water. Their eggs, covered with jelly capsules rather than hard
shells, lose water rapidly. For these reasons, amphibians require relatively moist
environments.
Many sampling techniques have been developed in North America or Europe where
most amphibians exhibit the complex life cycle of aquatic eggs, aquatic larvae, and
metamorphosis into terrestrial adults that return to water to breed. Not all amphib-
ians fit this stereotype. The spectacular diversity of amphibian life histories provides a
focus for studying their natural history, as well as presents a challenge since researchers
must ensure that field methods are appropriate for the target species.
1.2 Amphibian species richness and distribution

Scientists have named approximately 1.75 million species of living organisms
(Groom et al. 2006). About 72% of all named animals are insects; about 5%
are vertebrates. Approximately 0.5% of all animal species—6347 species (see
AmphibiaWeb; http://amphibiaweb.org)—belong to the class Amphibia. The
4 | Amphibian ecology and conservation
class gets it name from the Greek words amphi meaning “two” and bios mean-
ing “mode of life” because many species have a diphasic life history: they spend
part of their lives in water and part on land. Biologists divide the class into three
orders: Gymnophiona (caecilians), Urodela (salamanders), and Anura (frogs)
(Figure 1.1).
Gymnophiona, from the Greek words gymnos and ophis meaning “naked ser-
pent,” encompasses 174 species. Caecilians, which resemble large earthworms,
are long, skinny animals with no legs and reduced eyes. Annuli (grooves) encir-
cle their bodies. Their tails are either greatly reduced or absent, and a sensory
tentacle sits between each eye and nostril. Some caecilians have small dermal
scales beneath the surface of their mucus-covered skin. These scales, composed
mainly of collagen fibers and minerals, are not found in salamanders or anurans.
Adult caecilians range in length from a little more than 10 cm to about 1.5 m.
Most are highly specialized for burrowing. Others live on the ground but are
cryptic and secretive. Some are aquatic. Caecilians occur in tropical habitats
around the world except for Madagascar and the Papuan–Australian region
(Pough et al. 2004).
(a) (b)
(c) (d)
Fig. 1.1 Representatives of the three orders of amphibians. (a) Anura: Rana palmipes,
from Ecuador, (b) Anura: Bufo arenarum, from Argentina, (c) Urodela: Phaeognathus
hubrichti, from Alabama, USA, (d) Gymnophiona: Hypogeophis rostratus, Seychelles.
Photographs (a) and (b) by Martha L. Crump; photographs (c) and (d) by C. Kenneth
Dodd, Jr.
1 Amphibian diversity and life history | 5
Five hundred and seventy-one species belong to Urodela, from the Greek
words uro and delos meaning “tail evident.” All salamanders have tails, and adults
have elongate bodies. Most have front and back legs of about the same length;
the limbs of a few aquatic species are greatly reduced or absent. Salamanders
are completely aquatic, terrestrial, combined aquatic and terrestrial, fossor-
ial, or arboreal. Adults range in size from about 30 mm to nearly 2 m. Most
salamander species occur in eastern and western North America and temper-
ate Eurasia, although plethodontids have radiated extensively in Central and
South America. There are no salamanders in sub-Saharan Africa, Australasia,
Australia, or much of tropical Asia, and they are missing from most islands
(Pough et al. 2004).
Frogs, which include toads, make up the order Anura from the Greek words
an and oura, meaning “without tail.” Although tadpoles have tails, adults do
not. Anurans live in the water, on the ground, underground, and in the trees.
Most have long, strong back legs well suited for jumping. Males of most species
call to attract females for mating. Adults of the 5602 recognized species range
in size from about 13 mm to 30 cm. Anurans live almost everywhere except
where restricted by cold temperatures or extremely dry conditions, and except
for many oceanic islands (Pough et al. 2004).
Duellman identified 43 areas worldwide with exceptionally high numbers
of amphibian species, endemic species (those found nowhere else), or both
(Duellman 1999). Nineteen of these high diversity areas are in the western hemi-
sphere; the others are in Eurasia, Africa, and the Papuan–Australian region.
The neotropical region houses 54% of the world’s amphibian species.
Amphibians live in nearly every habitat except for open oceans, most oceanic
islands, polar regions, and some extremely dry deserts (Wells 2007). These
restrictions are imposed on them because of their highly permeable skin that
loses water, and because they are ectothermic: the energy needed to raise their
body temperatures comes from the sun. Thus, amphibians become inactive at
low temperatures. These characteristics, however, work to their advantage as
well. In dry areas and those with seasonal rainfall, amphibians absorb water
through their skin by contacting moist soil. Their low metabolic rates translate
into low energy requirements and allow them to estivate, often underground,
during unfavorable conditions.
1.3 Amphibian lifestyles and life history diversity

Textbooks, management guides, and monitoring manuals, especially those ori-
ginating in Europe and North America, often give an oversimplified impression
of amphibian life histories. In fact, amphibian life histories are often complex,
and many are still poorly understood. Not all species have an aquatic larval stage
and a terrestrial adult stage.
Reproductive mode, a central aspect of amphibian life history, refers to the
site of egg deposition, egg and clutch characteristics, type and duration of
embryonic and larval development, and type of parental care if any (Duellman
and Trueb 1986). Many amphibians are not tied to aquatic habitats for repro-
duction. Instead, they reproduce on land, underground, or in trees, even in
the temperate zones. The following brief discussion of selected lifestyles and
life histories reveals that similar behaviors have evolved in diverse taxonomic
groups and geographical areas: amphibian “experiments” toward greater inde-
pendence from standing or flowing water and perhaps from lower predation
pressure as well.
Readers wishing more information concerning amphibian life histories
should consult Duellman (2007) and Wells (2007). For reviews of reproduct-
ive modes, see Salthe (1969), Salthe and Duellman (1973), Wake (1982, 1992),
Haddad and Prado (2005), and Duellman (2007). For reviews of parental care
see Crump (1995, 1996).
1.3.1 Caecilians
Caecilian lifestyles include aquatic, combined aquatic and terrestrial, terrestrial,
and fossorial. Fully aquatic caecilians generally have compressed bodies with
well-developed dorsal fins on the posterior portion. Fossorial species generally
have blunt heads, used for pushing and compacting the soil while burrowing.
Although details of reproductive biology are unknown for many species, cae-
cilians display two basic modes: oviparous (egg-laying) and viviparous (bear-
ing live young) (Wake 1977, 1992). Oviparous caecilians lay eggs on land. In
some species the eggs hatch into larvae that wriggle to water. Caecilian larvae
exhibit less dramatic metamorphosis than do salamanders or frogs. They hatch
almost fully developed, and the larval period is short. The eggs of some ovip-
arous species undergo direct development. That is, development occurs within
the egg capsule; there is no free-living larval stage. Some oviparous females stay
with their eggs, which probably protects them from predators and from drying
out. Viviparity evolved independently several times in caecilians. Females retain
eggs in their oviducts until development is complete (Wake 1993; Wilkinson
and Nussbaum 1998). After the developing young exhaust their yolk reserves,
they scrape lipid-rich secretions from their mothers’ oviductal epithelium with
their fetal teeth. Gestation lasts for many months, and the newborn are large
relative to their mothers.
1.3.1.1 Aquatic
All species in the South American family Typhlonectidae are either fully aquatic
or semi-aquatic. Some in the latter group spend the day in burrows they con-
struct next to water, then emerge at night to feed in shallow water. All typhlonec-
tids are assumed to be viviparous. Soon after the young are born, they shed their
gills and quickly acquire adult morphology.
1.3.1.2 Combination aquatic and terrestrial

Caecilians of the two most primitive families—Ichthyophiidae and
Rhinotrematidae—all appear to have complex life cycles encompassing both
water and land. As far as is known, ichthyophiids from southeastern Asia lay
their eggs in burrows or under vegetation at the edge of water. The female coils
around her eggs. The hatchling larvae wriggle to water. Rhinotrematids from
South America likewise appear to be oviparous, and free-living aquatic lar-
vae have been found for some species. Although egg-laying sites are unknown,
females presumably oviposit on land near water and the larvae make their way to
water. The developmental mode is not known for uraeotyphlids, from southern
India, but presumably they also lay eggs on land and have aquatic larvae. Some
species of the large, widespread family Caeciliidae also oviposit on land and have
free-living aquatic larvae.
1.3.1.3 Terrestrial and/or fossorial

Many caecilians burrow underground, although some forage at night on the
ground surface. Some terrestrial and fossorial caeciliids from South America,
Africa, India, and the Seychelles lay direct-developing eggs. In some of these,
females have been found coiled around their eggs. Some terrestrial and fossor-
ial caeciliids are viviparous. All members of the family Scolecomorphidae from
Africa are terrestrial. One is viviparous; in the other five species the mode is
unknown.
1.3.2 Salamanders
Salamander lifestyles include fully aquatic, combined aquatic and terrestrial,
terrestrial, arboreal, and fossorial. Body shapes of aquatic salamanders range
from the slender and eel-like sirenids and amphiumids to the flattened and
robust cryptobranchids. Many permanently aquatic species retain external gills
as adults. Most arboreal salamanders are small and have extensively webbed feet;
some have prehensile tails. Burrowing salamanders generally have long slender
bodies and tails, reduced limbs and feet, and small body size.
As a group salamanders exhibit diverse reproductive modes. Some salaman-

ders go through a complex life cycle with aquatic larvae and metamorphosis.
Others are paedomorphic: they become sexually mature and reproduce while
retaining juvenile characteristics. Still others have direct-developing eggs or give
birth to live young.
1.3.2.1 Aquatic
Some aquatic salamanders lay eggs in still or slowly flowing water. Siren and
Pseudobranchus from southeastern USA and northeastern Mexico live in swamps,
lakes, marshes, and sluggish streams, where they attach their eggs to vegetation.
The larvae develop into eel-like paedomorphic adults: they lack eyelids, have
external gills, and their skin resembles larval skin. When their habitat dries out,
sirenids secrete a mucous cocoon and burrow into the mud where they estivate
until conditions improve. The Mexican axolotl (Ambystoma mexicanum) is per-
manently aquatic and paedomorphic. Its larvae fail to metamorphose fully, and
the gonads mature in the larval body form.
Proteus anguinus, from southeastern Europe, lives in subterranean lakes and
streams of limestone caves where it lays aquatic eggs that hatch into aquatic lar-
vae. Females attend their eggs. In North America, Typhlomolge and Haideotriton
also live in cave waters and lay aquatic eggs that hatch into aquatic larvae. All
these species are paedomorphic.
Some aquatic salamanders oviposit in flowing water of cold streams. Male
hellbenders (Cryptobranchus alleganiensis) from eastern North America con-
struct nests under rocks. More than one female might lay her strings of eggs in
the nest. The male guards the eggs through their early stages. Likewise, male
Andrias japonicus from Japan and male Andrias davidianus from central China
guard their nests. In all three cryptobranchid species, aquatic larvae undergo
incomplete metamorphosis. The adults retain certain larval features such as lid-
less eyes and the absence of a tongue pad.
1.3.2.2 Combination of aquatic and terrestrial

In Eurasia, many salamandrids (e.g. Triturus) live on land but lay their eggs in
ponds, lakes, or streams. Likewise, most Asian hynobiids are terrestrial as adults
but migrate to aquatic sites to lay aquatic eggs that hatch into aquatic larvae. The
same is true for many North American terrestrial salamanders. Tiger salaman-
ders (Ambystoma tigrinum) and spotted salamanders (Ambystoma maculatum)
migrate to breeding ponds in the spring. Once the aquatic larvae metamorphose
and leave the water, they return to their aqueous beginnings only during breed-
ing season.
In contrast, female marbled salamanders (Ambystoma opacum) lay their eggs

under leaf litter or logs in depressions on dry ground. They stay until their eggs
hatch when the nests flood during winter rains. Following aquatic larval devel-
opment and metamorphosis, marbled salamanders are completely terrestrial.
Female Amphiuma from the southeastern USA lay large yolky eggs in dried-
out swamps or in cavities under logs near ponds: places that will flood during
spring and summer rains. Females stay with their eggs until then. Once they
hatch, the well-developed larvae metamorphose within a few weeks. Adults live
in swamps or slow-moving streams, but move overland during rains. They sur-
vive droughts by burrowing underground for 2 years or more.
Some plethodontid salamanders lay eggs on land, and after hatching the lar-
vae make their way to water. Female Hemidactylium scutatum oviposit just above
or at the water line in bogs and swamps. After hatching, the larvae wriggle or flip
into the water. Some Desmognathus lay their eggs under rocks, logs, or leaf litter
at water’s edge. The newly hatched larvae stay with their mother in the nest site
for several days, then wriggle to water.
The complex life cycle of red-spotted newts (Notophthalmus viridescens) from
eastern North America involves several stages. Aquatic females attach their eggs
to underwater vegetation. The eggs hatch into aquatic larvae that eventually
metamorphose into an immature terrestrial stage called an eft. The orange-red
efts stay on land for 1–14 years. Eventually they return to ponds where they turn
dull green with a row of small red dots along each side of their body and trans-
form into the aquatic adult body form. In some populations, however, paedo-
morphic adults reproduce in the larval body form.
1.3.2.3 Terrestrial
Many species of salamander live under leaf litter or logs, retreating into crevices
or holes during dry conditions. Terrestrial salamanders, including many in the
temperate zone, have various ways of reproducing independent of water bodies.
The seepage salamander (Desmognathus aeneus) oviposits in a nest near water. The
larvae hatch at an advanced stage and metamorphose within a few days in the
nest without feeding. Other plethodontid salamanders (e.g. Desmognathus wrighti,
Bolitoglossa, and Plethodon) lay large, direct-developing eggs in cavities or inside
hollow logs. In many species, the female remains with the eggs; in a few species
the male remains instead. Montane populations of European fire salamanders
(Salamandra salamandra) often retain eggs in their oviducts. The young absorb
nutrients from their yolk and are born live, although lowland populations usually
have aquatic eggs and larvae. Salamandra atra are viviparous; after the developing
young exhaust their yolk reserves, they obtain nutrients from the female.
1.3.2.4 Fossorial
Most Oedipina, fossorial or semi-fossorial plethodontid salamanders ranging
from Mexico to northern South America, lay direct-developing eggs under-
ground. Some fossorial plethodontids (e.g. Lineatriton lineola) attend their
direct-developing eggs.
1.3.2.5 Arboreal
Neotropical arboreal plethodontids (e.g. Bolitoglossa and Nototriton) lay their
eggs under mats of mosses and liverworts on tree branches and in bromeliads.
The eggs undergo direct development, and some female Bolitoglossa attend their
eggs. Aneides lugubris from western North America lays direct-developing eggs
as high as 10 m above the ground.
1.3.3 Anurans
Anuran lifestyles include purely aquatic, aquatic and terrestrial, terrestrial, arbor-
eal, and fossorial. A frog with relatively short hind legs is most likely a terrestrial
hopper or fossorial species. One with long hind legs is likely to be aquatic, arbor-
eal, or a jumping terrestrial species. Aquatic anurans tend to have their eyes on the
top of their heads rather than at the sides, and they often have fully webbed feet.
Some have flattened bodies. Arboreal frogs generally have expanded pads on the
ends of their toes. Many fossorial species have small heads with pointed snouts
and depressed bodies. Some have spadelike tubercles on their hind feet used for
burrowing.
Anurans have evolved remarkably diverse life histories, from aquatic eggs to
viviparity. In between those extremes, frogs from numerous families on many
continents lay their eggs out of water yet have aquatic larvae. In the section head-
ings below, the first designation refers to post-metamorphic stages, the second
to egg and/or larval stages.
1.3.3.1 Aquatic/aquatic
Many aquatic frogs lay eggs that hatch into tadpoles. African clawed frogs
(Xenopus) attach their eggs to submerged vegetation in standing water. The
South American paradox frog (Pseudis paradoxa) oviposits among vegetation
in shallow water of ponds and lakes. Aquatic larvae grow to 25 cm—the largest
of any frog—then they metamorphose into relatively small juveniles. Tailed
frogs (Ascaphus truei) from northwestern USA and adjacent Canada live in cold,
torrential streams where they lay eggs under rocks. In some areas larvae require
several years to metamorphose.
Some fully aquatic anurans brood their young. Female South American Pipa
carry eggs embedded in their backs. In some species of Pipa the eggs hatch into
tadpoles. In others they undergo direct development. Female Australian gastric-
brooding frogs (Rheobatrachus) swallowed their late-stage eggs or early-stage
larvae, and the tadpoles absorbed yolk reserves while developing in their mothers’
stomachs. No Rheobatrachus have been seen since the 1980s; both species are
assumed extinct.
1.3.3.2 Terrestrial/aquatic
Terrestrial anurans from many families lay aquatic eggs, and the aquatic larvae
metamorphose into terrestrial juveniles. Species that oviposit in standing water
produce clutches in compact masses (some ranids), floating rafts on the water
surface (some ranids), strings (most Bufo and some pelobatids), scattered indi-
vidually or in small packets on the bottom substrate (Bombina and Discoglossus),
attached individually or in small groups onto submerged plants (some species
of Pseudacris, Acris, Hyla, and Spea), or as a film on the water surface (some
microhylids).
Some anurans breed in moderately fast streams or mountain torrents. Female
Atelopus from Central and South America lay their eggs in strings attached to
rocks. The larvae have ventral sucker-like discs that allow them to adhere to
rocks while feeding. Tadpoles of bufonid stream-breeding Ansonia from Asia
and Werneria from Africa have similar suckers. Other stream-adapted frogs,
such as many ranids, lay large eggs in compact masses attached to rocks in areas
where the current is slow and the eggs are less likely to be swept away.
Some terrestrial anurans produce foam nests in which their eggs are sus-
pended. Male Leptodactylus, Physalaemus, and Pleurodema from Central and
South America kick their hind legs during amplexus, whipping the females’
eggs and mucus and their sperm and mucus into foamy masses (Figure 1.2a).
The outermost layer of foam dries quickly and provides some protection against
desiccation and predation. Some foam-nesters produce their nests on the water
surface, others in cavities or holes next to ponds. Leptodactylus bufonius con-
structs mud nests at the margin of temporary ponds and deposits its foam nests
inside (Figure 1.2b). The eggs hatch into tadpoles that remain in the nests until
rains dissolve the nests and flood the area. Some Australian myobatrachids also
construct foam nests on the water surface.
Some terrestrial frogs oviposit in very small bodies of water on land.
Brazilian Bufo castaneoticus lay their eggs in water-fi lled fruit capsules of the
Brazil nut tree, and the tadpoles feed on detritus. Eupsophus from Chile and
(a) (b)
(c) (d)
Fig. 1.2 Representatives of four anuran modes of reproduction. (a) Pleurodema borelli
pair constructing a foam nest on the surface of water, from Argentina, (b) Leptodactylus
bufonius mud nest by the edge of a depression, from Argentina, (c) Hyla bokermanni eggs
on a leaf above water, from Ecuador, (d) male Rhinoderma darwinii brooding tadpoles in
its vocal sac, from Chile. Photographs by Martha L. Crump.
Crinia georgiana from Australia oviposit in small water-fi lled depressions,

seeps, or crevices on the ground where non-feeding larvae absorb their large
yolks before metamorphosing.
Temperate and tropical frogs with terrestrial eggs and aquatic larvae accom-
plish this feat in diverse ways. Dendrobatids lay their eggs on land, but then a
parent transports the larvae to water. In a few species of Dendrobates, the female
also provides her young with unfertilized eggs as food. Rhinoderma rufum, from
Chile, lays its eggs on land. The male takes late-stage eggs into his vocal sac
where they hatch, then transports the larvae to water. In Europe, male midwife
toads (Alytes) carry their eggs wrapped around their hind legs. Eventually they
hop to ponds where the eggs hatch into aquatic larvae.
1.3.3.3 Arboreal/aquatic
Taxonomically diverse arboreal frogs lay their eggs on vegetation overhanging
water. After the eggs hatch, the larvae fall into the water below where they continue
to develop. Agalychnis, Phyllomedusa, and some Hyla from the New World trop-
ics lay their eggs over standing water (Figure 1.2c), and neotropical centrolenids
oviposit over flowing water. In some centrolenids, a parent protects the eggs from
predators and keeps them moist by resting on them. Female Afrixalus from sub-
Saharan Africa oviposit on leaves above water, then fold the leaf edges together and
glue them in place with oviductal secretions. Some arboreal Old World rhacoph-
orids and hyperoliids construct foam nests on vegetation overhanging temporary
pools or slow-moving streams.
Water-filled basins offer oviposition sites that presumably lessen the risk of
eggs getting swept away and reduce predation. Males of several neotropical
gladiator frogs (e.g. Hyla boans and Hyla rosenbergi) construct basins beside
streams or rivers. Water seeps in and fills the nests, and the frogs lay eggs as a
surface film. After developing in the basin, the tadpoles metamorphose into
froglets that take to the trees.
Some arboreal anurans oviposit in water-filled tree holes and axils of aer-
ial plants. The eggs of many of these frogs (e.g. Anodonthyla, Platypelis, and
Plethodonthyla from Madagascar) have large amounts of yolk. The tadpoles typ-
ically lack mouthparts, and they are non-feeding. In contrast, female Osteopilus
brunneus from Jamaica lay eggs in water-filled leaf axils of bromeliads and con-
tinue to deposit about 250 more eggs in the bromeliad every few days throughout
the tadpoles’ development. The tadpoles—up to about 170 in a clutch—feed on
the later-arriving eggs until they metamorphose into arboreal froglets.
Some arboreal frogs attach their eggs to the walls of water-filled cavities in
trees. After hatching, the tadpoles drop into the water. In Chirixalus eiffingeri,
an Asian rhacophorid, the female returns periodically and deposits fresh eggs
for her tadpoles to eat. In others of this reproductive mode, the aquatic larvae
feed on algae and debris.
In the New World tropics, female Flectonotus carry their eggs in dorsal
pouches. After the eggs hatch as advanced tadpoles, the females transport
them to water-filled bromeliads or bamboo where they complete development.
Females of some neotropical Gastrotheca also brood their eggs in dorsal pouches
and transport the tadpoles to aquatic sites.
1.3.3.4 Fossorial/aquatic
Scaphiopus and Spea, North American spadefoot toads, spend much of their
lives underground but emerge following heavy rains and lay their eggs in newly
formed ponds. The tadpoles develop quickly, which increases the probability
of metamorphosing before the ponds dry. Rhinophrynus dorsalis, from south-
ern Texas to Costa Rica, likewise lives underground and emerges after the first
heavy rains to breed in temporary ponds. Female Hemisis marmoratum from
Africa lay their eggs in subterranean chambers and stay with their eggs until
after they hatch. At that point, the female digs a tunnel into adjacent water for
the tadpoles.
1.3.3.5 Terrestrial/non-aquatic
Terrestrial anurans exhibit diverse life histories that free them from aquatic
breeding sites. Adenomera from South America deposits foam nests under logs
or in terrestrial cavities, and non-feeding tadpoles develop in the nests until they
metamorphose. Neotropical Eleutherodactylus, Oreophrynella from Guyana and
southern Venezuela, and New Guinean microhylids lay direct-developing eggs
under logs or leaf litter. Many attend their eggs.
Some completely terrestrial anurans brood their young. Female Assa dar-
lingtoni from Australia attend their terrestrial eggs. When the eggs are about
12 days old, the father climbs into the egg mass, rupturing the capsules. The
newly hatched tadpoles wriggle into brood pouches, one along each side of
the male’s body. He broods his non-feeding larvae until they metamorph-
ose into froglets. Female Darwin’s frogs (Rhinoderma darwinii) from Chile
and Argentina lay their eggs on moist ground. Just before the eggs hatch the
males gobble them into their mouths and into the vocal sacs (Figure 1.2d).
The young ingest secretions from the vocal-sac lining and emerge from their
fathers’ mouths as froglets.
Several Nectophrynoides, African bufonids, retain eggs in their oviducts and
give birth to live young. In Nectophrynoides occidentalis, after depleting their
yolk reserves the developing embryos feed on “uterine milk” secretion produced
by glands in the mother’s oviduct walls. These frogs live at high elevations,
exposed to long periods of cold and drought. The females have a 9-month gesta-
tion period during which they estivate underground.
1.3.3.6 Arboreal/non-aquatic
Some neotropical Eleutherodactylus lay their direct-developing eggs in tree
holes, bromeliads, moss, or on leaves. Some attend their eggs, others do not.
Eleutherodactylus jasperi from Puerto Rico lived in arboreal bromeliads and gave
birth to live young. The direct-developing eggs were retained in the oviducts,
and nutrition came entirely from the embryo’s yolk reserves. This species has not
been seen since 1981 and is assumed extinct.
Female Cryptobatrachus, Stefania, and Hemiphractus, neotropical hylids,
carry their direct-developing eggs exposed on their backs, secured by mucous
gland secretions. Females of some Gastrotheca brood direct-developing eggs in
dorsal pouches that protect the developing embryos from predators and des-
iccation and also function in gaseous exchange between the females and their
embryos.
1.3.3.7 Fossorial/non-aquatic
The burrowing microhylid Synapturanus salseri from Colombia lays its eggs
in burrows just below the root mat on the forest floor. Non-feeding tadpoles
hatch at an advanced stage and absorb their yolk reserves. The Brazilian bur-
rowing leptodactylid Cycloramphus stejnegeri likewise oviposits in underground
nests and has non-feeding tadpoles. Other fossorial anurans, such as Geocrinia
and Arenophryne (Australian myobatrachids), Callulops (New Guinean micro-
hylids), and Breviceps (African microhylids), lay direct-developing eggs in
underground burrows. Female Breviceps stay with their eggs and presumably
keep them moist.
1.4 Amphibian declines and why they matter

The world is experiencing a “biodiversity crisis”: rapid and accelerating loss of
species and habitat (Ehrlich and Ehrlich 1981; Myers 1990; Raven 1990; Wilson
1992). Amphibians are part of this overall loss. Populations of amphibians are
declining and disappearing worldwide at an increasing rate as compared to pre-
1980 decades, even from protected areas (Blaustein and Wake 1990; Phillips
1994; Stuart et al. 2004). During the late 1980s and early 1990s, many declines
seemed mysterious because there was no obvious cause. Skeptics argued that
declines might be simply natural population fluctuations. Since the late 1980s,
scientists worldwide have focused on determining the extent of declines and
identifying the causes. We now know these declines are real.
The International Union for the Conservation of Nature (IUCN) assesses
the status of species on a global scale and maintains and updates a catalog of
taxa that face a high risk of global extinction: the IUCN Red List of Threatened
Species. The 2008 update lists 30% of described amphibians as threatened with
extinction (IUCN 2008). Since 1500, at least 39 species of amphibians have
become extinct.
Scientists have hypothesized six major threats to amphibians: habitat modi-

fication and destruction, commercial over-exploitation, introduced species,
environmental contaminants, global climate change, and emerging infectious
diseases, especially the chytrid fungus Batrachochytrium dendrobatidis (Collins
and Storfer 2003). Most agree the primary threat is habitat modification and
destruction. For the past 100 years, human population growth has been expo-
nential and has occurred largely in areas with the highest amphibian species
richness: the tropics and subtropics (Gallant et al. 2007). As a result, these land-
scapes are being heavily modified to support agriculture and other human activ-
ities. The chytrid fungus also is exerting a major impact in many areas and on
many species (Smith et al. 2006). Thus far the chytrid has caused the decline
or extinction of about 200 species of frog (Skerratt et al. 2007). Many factors
make the chytrid a significant concern, including its wide distribution in both
the New and Old Worlds, its rapid spread and high virulence, and the fact that
it infects a broad diversity of host species (Daszak et al. 1999).
Why should we care if we lose amphibians? It is for the same basic reasons we
should care if other animals and plants disappear: economics, ecosystem func-
tion, esthetics, and ethics (Noss and Cooperrider 1994; Groom et al. 2006).
1.4.1 Economics
Selfishly, we should care if we lose amphibians because we use them for our own
benefit, including for food and as pets. We use literally tonnes of frogs each year
in medical research and teaching. We have isolated novel chemical compounds
from granular glands of anuran skin and have used these compounds to develop
new drugs.
1.4.2 Ecosystem function

Amphibians play a key role in energy flow and nutrient cycling because they
serve as both predator and prey. By eating huge quantities of algae, tadpoles
reduce the rate of natural eutrophication, the over-enrichment of water with
nutrients, which leads to excessive algal growth and oxygen depletion. Most
adult amphibians eat insects and other arthropods. As ectotherms, amphib-
ians are efficient at converting food into growth and reproduction. Unlike
endothermic birds and mammals that generate heat metabolically, amphibians
expend relatively little energy to maintain themselves. Birds and mammals
use up to 98% of their ingested energy to maintain their body temperatures,
leaving as little as 2% to be converted to new animal tissue: food for preda-
tors. In contrast, amphibians convert about 50% of their energy gained from
food into new tissue, which is transferred to the next level in the food chain
(Pough et al. 2004). If amphibians disappeared, would the world be overrun

with houseflies, mosquitoes, and crop-eating insect pests? Would their preda-
tors go extinct?
1.4.3 Esthetics
Imagine the silence of rainy spring evenings without the lively croaking of male
frogs. The monotonous roads without spring migrations of salamanders. People
worldwide consider frogs to be good luck because of their association with rain.
Amphibians provide inspiration for our artistic endeavors, from literature to
music and the visual arts.
1.4.4 Ethics
Every species is a unique product of evolution. In 1982 the United Nations
General Assembly adopted the World Charter for Nature, which states: “Every
form of life is unique, warranting respect regardless of its worth to man, and,
to accord other organisms such recognition, man must be guided by a moral
code of action” (Noss and Cooperrider 1994). More than 100 nations signed
the charter. Like all other living species, amphibians have intrinsic value and a
right to exist.
Amphibians, amazing descendants of terrestrial pioneers, are fighting for
their lives in a world greatly modified by humans.
1.5 References
Blaustein, A. R. and Wake, D. B. (1990). Declining amphibian populations: a global phe-
nomenon? Trends in Ecology and Evolution, 5, 203–4.
Collins, J. P. and Storfer, A. (2003). Global amphibian declines: sorting the hypotheses.
Diversity and Distributions, 9, 89–98.
Crump, M. L. (1995). Parental care. In H. Heatwole and B. K. Sullivan (eds), Amphibian
Biology, vol. 2: Social Behaviour, pp. 518–67. Surrey Beatty and Sons, Chipping
Norton, NSW.
Crump, M. L. (1996). Parental care among the Amphibia. In J. S. Rosenblatt and
C. T. Snowdon (eds), Advances in the Study of Behavior, vol. 25: Parental Care: Evolution,
Mechanisms, and Adaptive Significance, pp. 109–44. Academic Press, New York.
Daszak, P., Berger, L., Cunningham, A. A., Hyatt, A. D., Green, D. E., and Speare, R.
(1999). Emerging infectious diseases and amphibian population declines. Emerging
Infectious Diseases, 5, 735–48.
Duellman, W. E. (1970). The Hylid Frogs of Middle America, vol. 1. Monograph of the
Museum of Natural History, Number 1. University of Kansas, Lawrence, KA.
Duellman, W. E. (1999). Global distribution of amphibians: patterns, conservation, and
future challenges. In W. E. Duellman (ed.), Patterns of Distribution of Amphibians: a
Global Perspective, pp. 1–30. John Hopkins University Press, Baltimore, MD.
Duellman, W. E. (2007). Amphibian life histories: their utilization in phylogeny and clas-
sification. In H. Heatwole and M. J. Tyler (eds), Amphibian Biology, vol. 7. Systematics,
pp. 2843–92. Surrey Beatty and Sons, Chipping Norton, NSW.
Duellman, W. E. and Trueb, L. (1986). Biology of Amphibians. McGraw-Hill, NewYork.
Ehrlich, P. R. and Ehrlich, A. H. (1981). Extinction: The Causes and Consequences of the
Disappearance of Species. Random House, New York.
Gallant, A. L., Klaver, R. W., Casper, G. S., and Lannoo, M. J. (2007). Global rates of
habitat loss and implications for amphibian conservation. Copeia, 2007, 967–79.
Groom, M., Meffe, G. K., and Carroll, C. R. (2006). Principles of Conservation Biology,
3rd edn. Sinauer Associates, Sunderland, MA.
Haddad, C. F. B. and Prado, C. P. A. (2005). Reproductive modes in frogs and their unex-
pected diversity in the Atlantic Forest of Brazil. BioScience, 55, 207–17.
IUCN (International Union for the Conservation of Nature) (2008). Red List of
Threatened Species. IUCN, Gland. http://www.iucn.org/themes/ssc/redlist.htm.
Myers, N. (1990). Mass extinctions: what can the past tell us about the present and the
future? Global and Planetary Change, 82, 175–85.
Noss, R. F. and Cooperrider, A. Y. (1994). Saving Nature’s Legacy: Protecting and Restoring
Biodiversity. Island Press, Washington DC.
Phillips, K. (1994). Tracking the Vanishing Frogs: an Ecological Mystery. St. Martin’s Press,
New York.
Pough, F. H., Andrews, R. M., Cadle, J. E., Crump, M. L., Savitzky, A. H., and Wells, K. D.
(2004). Herpetology, 3rd edn. Prentice Hall, Upper Saddle River, NJ.
Raven, P. H. (1990). The politics of preserving biodiversity. BioScience, 40, 769–74.
Salthe, S. N. (1969). Reproductive modes and the number and sizes of ova in the urodeles.
American Midland Naturalist, 81. 467–90.
Salthe, S. N. and Duellman, W. E. (1973). Quantitative constraints associated with
reproductive mode in anurans. In J. L. Vial (ed.), Evolutionary Biology of the Anurans,
pp. 229–49. University of Missouri Press, Columbia, MO.
Skerratt, L. F., Berger, L., Speare, R., Cashins, S., McDonald, K. R., Phillott, A. D.,
Hines, H. B., and Kenyon, N. (2007). Spread of chytridiomycosis has caused the rapid
global decline and extinction of frogs. EcoHealth, 4, 125–34.
Smith, K. F., Sax, D. F., and Lafferty, K. D. (2006). Evidence for the role of infectious
disease in species extinction and endangerment. Conservation Biology, 20, 1349–57.
Stuart, S. N., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues, A. S.L., Fischman, D. L.,
and Waller, W. (2004). Status and trends of amphibian declines and extinctions world-
wide. Science, 302, 1783–6.
Wake, M. H. (1977). The reproductive biology of caecilians: an evolutionary perspec-
tive. In D. H. Taylor and S. I. Guttman (eds), The Reproductive Biology of Amphibians,
pp. 73–101. Plenum, New York.
Wake, M. H. (1982). Diversity within a framework of constraints. Amphibian repro-
ductive modes. In D. Mossakowski and G. Roth (eds), Environmental Adaptation and
Evolution, pp. 87–106. Gustav Fischer, New York.
Wake, M. H. (1992). Reproduction in caecilians. In W. C. Hamlett (ed.), Reproductive
Biology of South American Vertebrates, pp. 112–20. Springer-Verlag, New York.
Wake, M. H. (1993). Evolution of oviductal gestation in amphibians. Journal of

Experimental Zoology, 266, 394–413.
Wells, K. D. (2007). The Ecology and Behavior of Amphibians. University of Chicago
Press, Chicago, IL.
Wilkinson, M. and Nussbaum, R. A. (1998). Caecilian viviparity and amniote origins.
Journal of Natural History, 32, 1403–9.
Wilson, E. O. (1992). The Diversity of Life. The Belknap Press of Harvard University
Press, Cambridge, MA.
2
Setting objectives in field studies
Dan Cogălniceanu and Claude Miaud
2.1 Basic concepts for a good start

Considerable financial and human resources may be wasted due to poor research
design and implementation. In addition, poor planning can result in taking
non-repeatable or unreliable data that are of no or limited use, and leads to the
paradox of “data-rich, information-poor.” Rigorous study planning and the def-
inition of clear and realistic goals can help avoid wasted effort. Before undertak-
ing field studies, biologists first must define their objectives and then assess the
availability of resources necessary to accomplish them. This is a key stage, and
any uncertainty or vagueness at this point could prove detrimental to the success
and usefulness of the study.
There are two different approaches in scientific investigation: inductive and
deductive methods. The inductive or exploratory method relies on gathering data
without first identifying hypotheses to be tested; results may then be explained
based on the data gathered. However, science involves more than accumulating
data and then searching for patterns. A useful study must start with an obser-
vation that identifies a problem or question to be resolved. The observation
should have biological significance and be based on current evolutionary theory
(Wolff and Krebs 2008). For example, amphibian populations are declining:
what are the causes? Two related species hybridize on just a narrow contact zone:
why? Habitats are fragmented by human activity: what effects does this have on
population persistence, and why (e.g. dispersal abilities, food resources, popula-
tion demography, genetics, susceptibility to predators, and human activities)?
As science is an activity focused on understanding natural processes and, espe-
cially in modern times, in devising solutions to problems, the hypothesis-based
deductive method was developed (James and McCulloch 1985). An hypothesis
is a statement related to an observation which may be true, but for which proof
has not yet been found. The function of the hypothesis is to direct the search
for order among facts. Fieldwork then serves to test one’s hypothesis, thereby
scientifically demonstrating correlation, causation, or the absence thereof. The
hypothesis must of necessity regard some facts as more significant than others,
based on the researcher’s previous experience, familiarity with the literature,
and individual interpretation.
As scientific knowledge is rarely complete, any list of potential alternative
hypotheses is also unlikely to be complete. Therefore, alternative hypotheses for
the particular problem or question being considered must also be formulated.
Thus, it is necessary to conduct research with both the hypothesis and alter-
native hypothesis in mind, as more than one cause may be contributing to any
single effect (Wolff and Krebs 2008). For example, Jaeger (1972) used enclos-
ures to test the mechanism of interspecific competition between two species of
plethodontid salamanders, hypothesizing that it resulted from either differential
exploitation of food resources or through interference. Both primary and alter-
native hypotheses are formulated and tested so that researchers can determine
which explanations best fit the results obtained. Unsupported hypotheses can be
rejected, but even the most parsimonious hypothesis may not be fully accepted
because there still may be underlying explanations as yet untested (Senar 2004).
Repeated experimentation involving alternative hypotheses eventually allows
researchers to gain a measure of confidence in the validity of empirically sup-
ported hypotheses (Jaeger and Halliday 1998). Both hypotheses and data are
essential for credible science since, as Krebs (1999) concluded, hypotheses with-
out data are not very useful, and data without hypotheses are useful only for an
inductive approach.
Ecology is an empirical science that requires data from the “real” world,
and a field study is a tool which can ultimately lead to the acceptance or rejec-
tion of an hypothesis. After repeated experimentation and validation, data
and information from many field studies are synthesized and organized into
concepts of how nature functions. These concepts are the result of the integra-
tion between what scientists think they know (based on previous research and
observation) and newly acquired data (Ford 2000). Asking the right question
is important, because the type of question asked and the particular techniques
and methods used in hypothesis-driven research strongly influence what is
discovered and the direction of future research. During the course of this
feedback process, scientists come to a better understanding of the phenomena
they study (Figure 2.1).
Scientists present what they want to accomplish during a field study in terms
of goals and objectives. A goal is a statement that explains what the study is
designed to accomplish. It is usually a broad and general statement, inclusive of
2 Setting objectives in field studies | 23
Analysis and hypothesis testing
New theories and Data and

hypotheses information
Synthesis and creativity
New theories and Data and

Theories and Data and

Fig. 2.1 The cycle of scientific investigation and the shift towards the spiral of knowledge.
a long-term direction. The goal is then split into specific, measurable objectives
that indicate how the goal can be achieved within a specified time frame, and
the expected results.
Defining a study’s objectives clearly is the first and probably the most
important single step in research planning, and is a key element in a success-
ful project. The goals of research can be non-applied (i.e. aimed at increasing
or changing existing knowledge) or applied (i.e. focused on solving practical
questions), or both. Rather than addressing an effect, a study should focus
on the cause of the phenomenon in question. Separating effects from causes
is a major challenge when setting objectives. For example, acid rain has well-
known causes, but much research today focuses on the effects or seeks solutions
to limit its impact. In another example, many recent studies have reported on
amphibian declines (Stuart et al. 2004), but many fewer have identified spe-
cific causes (e.g. Becker et al. 2007). It is important that amphibian biologists
do not limit their focus to specific taxa; otherwise, they lose sight of the fact
that similar effects may be occurring in other taxa. Amphibian biologists need
to establish closer links with researchers studying similar problems in other
taxa (Halliday 2005).
Ideally, the cycle of scientific investigation should proceed from data and infor-
mation gathering and analysis, to quantifying knowledge, and to conceptual
understanding. A logical, stepwise approach of scientific inquiry (Lehner 1996)
should be followed: (1) perceive that a problem/question exists; (2) formulate a
possible explanation (i.e. devise an hypothesis); (3) formulate alternative hypoth-
eses; (4) identify the best approach to test the hypothesis (i.e. theoretical models,
experiments, or field observations); (5) collect and analyze data; (6) support or
reject the hypothesis; and (7) understand the meaning and implications of the
results. The original hypothesis can then be modified, experiments repeated and,
with time, conceptual understanding attained.
2.2 Steps required for a successful study

Ecological systems are large and complex and, at times, unpredictable. The more
complex the system, the more uncertainties arise. Data are most often obtained
within the parameters of a limited number of samples, restricted time, or specific
area, and then applied to larger scales. By carefully framing the hypothesis and
deciding the most practical, meaningful, and objective method of study, degree
of error can be minimized (Hayek 1994). The following section presents some of
the most important issues that must be considered when planning a field study.
2.2.1 Temporal and spatial scales

The issue of scale has three components (Schneider 2001): (1) pressing problems
in ecology often exist within timescales of decades or centuries, and cover large
areas; (2) most data can be gathered directly for only short periods of time and
over small areas; (3) patterns measured at small scales do not necessarily hold
true at larger scales; nor do processes observed at smaller scales necessarily exist
at larger ones. Thus, an inappropriate scale limits the inference of the results.
Spatial and temporal scales can be classified as gradient, interval, discrete,
and continuous (Bernstein and Goldfarb 1995). Setting temporal and spatial
scales involves defining their boundaries and helps to answer three practical
questions: where, for how long, and how often? Selecting the appropriate scale
is difficult because of environmental heterogeneity in both time and space.
Amphibian populations and communities also vary widely throughout time
and space, since they are influenced by a multitude of environmental variables
(Meyer et al. 1998; Pechmann et al. 1991; Pellet et al. 2006).
Temporal scales are rarely discussed explicitly, but they are often assumed
to span years rather than generations. Many time-dependent phenomena such
as extinction, predator–prey interactions, competition, and succession reveal
important insights if considered on a generational scale (Frankham and Brook

2004). The current distribution and abundance of animal species inhabiting an
area is the result, in part, of the impact of geological processes (e.g. plate tecton-
ics, orogenesis, sea-level changes, major catastrophes), ecological processes (e.g.
competition, predation, climate), life-history traits (e.g. dispersal), and recent
human impacts (e.g. changes in vegetation and habitat structure, overexploi-
tation, introduction of nonindigenous species). For example, in the Northern
Hemisphere the Pleistocene–Holocene post-glacial recolonizations of species
and human activities (e.g. habitat destruction and alteration) are the two major
factors shaping the current distribution of amphibians. Objectives of a time-
scale component of a study plan should include such factors as short-term (less
than one generation), intermediate (more than one to a few generations), and
long-term (many generations, and inclusive of a large spatial scale and monitor-
ing activities).
Spatial scales and objectives are strongly interconnected. Selection of a study
site includes such factors as the complexity of local macro- and microhabitats
and, at large spatial scales, biogeographical provinces and landscapes (Morrison
2002). Of great importance to the overall study plan are the criteria for describ-
ing vegetation, whether gross (e.g. foliage height, diversity), physiognomy (phys-
ical structure), or floristic (plant taxonomic description). The relative usefulness
of structural or floristic measures depends on the spatial scale of the analysis.
For example, whereas most studies do not require a detailed description of plant
taxa, species lists might be essential for characterizing microhabitat and trophic
(resource) availability.
The importance of choosing the correct scale can be exemplified by a study
assessing population fluctuations within a particular species: the first step would
be to understand the distribution of the species in order to thoroughly sample all
areas, rather than by biasing results by sampling in only one small portion of the
range. In another example, a study of the movement patterns of a species may
focus on individual distributional patterns, daily or weekly movements, sea-
sonal migrations related to reproduction or dispersal, or on region-wide range
shifts through years or decades.
2.2.2 Choosing the model species

Many amphibian species have restricted distributions and complex habitat
requirements. Some amphibian populations, especially among temperate pond-
breeding species, are maintained by episodic reproduction that occurs in a spor-
adic and unpredictable manner (Alford and Richards 1999). If a number of
different species is available for study, selecting an experimental species should
be based on an understanding of its life history (i.e. feeding, habitat use, disper-
sal abilities, reproduction, behavior, predators, phylogeography, population gen-
etics) and its suitability in resolving the particular question being asked (Wolff
and Krebs 2008). For example, a species with low detectability is not a good
choice for a mark–recapture study since it will require a considerable sampling
effort and entail potential capture bias (Weir et al. 2005). The role of keystone
species (e.g. Eleutherodactylus bransfordii in Costa Rican forests), umbrella spe-
cies (e.g. Rana sylvatica in the Milwaukee river basin), or flagship species (e.g.
Salamandra lanzai in the Western Alps) can also be considered. Rare or threat-
ened species are often selected because of conservation applications, but research
on them could incur potential risks due to ethical considerations. Researchers
should avoid choosing a rare or threatened species if common or less vulnerable
surrogate species can be selected.
2.2.3 Pilot/desk study

A preliminary pilot study usually helps to develop and test realistic and achiev-
able objectives, and avoids later shortcomings and failures. Usually a pilot study
is carried out on short temporal and small spatial scales, and allows the testing of
conceptual models and methods. What might seem simple during office plan-
ning might prove completely different or unworkable in the field. Even a simple
review of the literature can prove helpful in avoiding mistakes, however.
2.2.4 Elaborate a conceptual model

A conceptual model (e.g. a simple box diagram showing components and link-
ages) is a simplified model of the system to be studied. There are no ideal methods
to employ; instead, a multitude of models are available to choose from with
various degrees of complexity. Conceptual models are helpful in that they can
be used to select the variables to be measured that might be considered import-
ant to the study. Since professionals with diverse backgrounds have different
philosophies or approaches to using models, it is recommended that each team
member contribute knowledge of and/or expertise with various models, then
develop an integrated study model from which to work (Maher et al. 1994).
After model development, many of the design questions become more obvi-
ous. A model need not embrace all components of the system; it needs only to
be adequate for the scope of the investigation. There is always a possible risk that
the conceptual model is inappropriate or over-simplified; thus, a slightly more
complex model may need to be developed. Adopting a more complex model
focuses researchers on collecting additional data that might prove important,
despite the extra costs involved with sampling and measurements. In other
words, it might be better to collect more data than appears necessary at first,
than to discover later that some important parameter was omitted. A pilot study
helps to avoid under- or over-collecting data. Although conceptual models help
researchers identify what to measure, the timing of studies is determined by the
natural history of the species of interest.
2.2.5 The SMART approach

Specific objectives allow for greater chances of conducting a successful research
project. SMART stands for specific, measurable, attainable, relevant, and time
(Piotrow et al. 1997), an approach used in project writing that also helps in for-
mulating specific and measurable objectives within study plans. SMART helps
devise objectives that are clear and concise, indicates what is to be achieved,
addresses only attainable results, indicates when each stage will be completed,
and is not encumbered by idealistic aspirations.
• S: The specific part of an objective defines what will be done and where it will
occur. When setting objectives, ask simple questions that can be answered,
and avoid ambiguities, the use of buzzwords or jargon (e.g. “cutting-edge”),
and pompous phrasing.
• M: Measurable is an attribute of an activity or its results. The source of
and mechanism for collecting measurable data are identified, and collec-
tion of these data is determined feasible. If the objective of the study is to
document trends (i.e. an increase or decrease of one or more measurable
variables), then a baseline is required to act as a reference point (e.g. habi-
tat availability, characteristics and use; amphibian community structure;
population size). If a baseline is not yet available then it will be useful to
first have it established.
• A: Attainable refers to the probability of conducting the proposed activ-
ities within the established time frame with the available resources and sup-
port. It also includes the external factors critical to success. Doing research
today, for example, can become difficult due, in part, to increasing admin-
istrative restrictions (Prathapan et al. 2008). Other external factors include
unforeseen costs, shifts in exchange rates, obtaining collecting and access
permits, changes in legislation, and political, security, and health (both
human and animal) issues. In coping with external factors, a risk analysis
is useful because it allows for planning alternative strategies.
• R: Some useful measures for the relevance of the study’s objectives are the
utility and value of the results for practical purpose (e.g. management of
protected areas, better conservation measures, and ecological restoration). It
may be difficult to determine how relevant the objective may be, especially
since the true relevance may not be apparent until the study is completed
(e.g. if a drought were to affect a long-term study of amphibian breeding).
Perhaps an easy way to evaluate the relevance is to answer the “so what”
question, thus avoiding undertaking studies of limited or no interest.
• T: Finally, the time required to complete the project will depend on the
parameters discussed above.
2.2.6 Applying the SMART approach to

plan an amphibian inventory
In the following example, the SMART approach is used as the basis for setting
up an inventory of amphibian species in a national park or reserve.
• Specific: determine whether to inventory all amphibian species, or only
those of special interest, such as rare, threatened, or endemic species. Set
the spatial limits of the study, such as the administrative boundary of the
park or reserve, or specific habitats of interest (e.g. temporary ponds; roads
in areas which bisect amphibian dispersal corridors; high elevations con-
taining unique habitats or species richness, such as cloud forests).
• Measurable: select parameters to be measured (e.g. presence/not detected,
relative abundance, density, sex or life stage, percentage of area occupied),
select methods appropriate for each species (especially taking detection
probabilities into account), and decide habitat parameters (e.g. tempera-
ture, humidity, vegetative cover, characteristics of aquatic habitats).
• Attainable: identify resources and available support (e.g. funding, work
force) and administrative considerations (e.g. collecting and access per-
mits, training staff for field data collection, animal care and use require-
ments, security issues).
• Relevance: provide concise statements as to the importance of the research,
and whether it has practical applications. The relevance of the proposed
research is particularly important in field studies, especially if it will aid
in the management of protected areas and species conservation. Utility
also assists in determining research approaches. For example, incorpor-
ating measures of abundance with detection probabilities, in addition to
occupancy models, is much more informative in determining status than
occupancy models alone.
• Time: the research schedule (planning, fieldwork, data analysis, report
and publication preparation) should be clearly stated based on the previ-
ously outlined parameters and limiting factors.
2.2.7 Experimental versus field studies

Biological communities, apart from their high internal complexity, are subject
to random, naturally occurring fluctuations involving both physical and biotic
parameters (e.g. heat, cold, drought, effects of disease, and spread of invasive
species). Stochastic fluctuations such as these are a major source of statistical
variance in nature, resulting in a shift by some researchers towards less complex
and more controlled field and laboratory experimental designs (see Chapter 6).
Such a reductionist approach can be framed within a simpler conceptual model
where the number of variables of interest is reduced in return for minimizing
uncontrolled environmental fluctuation.
Still, there is an experimental continuum between laboratory and field studies
(Figure 2.2). Perhaps some of the most powerful experimental tools which can
provide the statistical rigor required for hypothesis testing are outdoor experi-
ments (Fauth 1998; Rowe and Dunson 1994; Chapter 6). Controlled experi-
ments maximize a researcher’s ability to detect a response to variables of interest
(e.g. food, density of individuals, hydroperiod). The best approach is to use
observations in nature coupled with experimental analysis. For example, Wilbur
(1997) used both natural and artificial ponds to investigate complex food webs in
temporary ponds, and their effects on the larval amphibian community.
2.2.8 Methods for sampling, data storage, and analysis

Selecting an appropriate research technique is not an easy task, especially since
“trendy” methods may not be the best ones to use. Methods selected must be
adequate for the proposed objectives, and it is best to test them first to determine
Least Most Least
Mathematical
model
Control and precision
Laboratory
Unpredictability
experiment Simple
Realism
Outdoor
Complexity
experiment
Unrestricted field
experiment
Complex Most Least Most
Fig. 2.2 Trade-offs between complex and simple experiments.

efficiency and effectiveness. Researchers should adopt complimentary models,

applicable and relevant methods, and an altogether SMART use of study plans,
resources, and results. The most important elements in any study plan are accurate
and precise methods that meet the objectives of the study. A cost-benefit analysis
can prove helpful in the final selection of the methods (Arntzen et al. 1995, 2004).
The sampling design should be both efficient (a better use of sampling effort
in obtaining results) and simple (easily comprehended and easily implemented)
(Scott and Köhl 1993). Statistical analyses often have biases or limitations in
ecological studies (Krebs 1999; Pollock et al. 2002). Statistical significance may
not be the equivalent of biological significance. Focusing too much on statistical
issues may prove deceptive or lead to inaccurate conclusions. One way to min-
imize the potential of error is to incorporate power analysis into study designs
(Yoccoz 1991). Statistical power refers to the ability of a test to correctly reject
a null hypothesis, and is frequently evaluated in the context of the sample size
required to detect an effect of a given magnitude (Michener 2000).
When using probability statistics, it is possible to make two kinds of error:
a researcher may claim there is a difference when one does not exist, or can fail
to detect a difference when one does exist (Underwood 1997). The first type of
error is estimated by the traditional probability value (0.05) associated with
statistical tests. If a researcher rejects the null hypothesis at this level, then there
is still a 5% chance that he/she is in error.
The second type of error is more difficult to estimate because it depends on
the sample size, the magnitude of effect, and sample variability. Researchers
are likely to correctly reject a null hypothesis with larger sample sizes, larger
effect sizes, or less variable samples. Alternatively, a higher critical value may be
selected (e.g. 0.1) as an indication of statistical significance.
Statistical power is a measure of the proportion of time that a researcher
would correctly reject the null hypothesis if an experiment could be repeated
an infinite number of times; statistical power is usually estimated by computer
simulation. The goal of a power analysis is to define a level of confidence in the
research results; it can also be used in determining trends and in setting optimal
sample size. A discussion of power analysis is available through StatSoft (www.
statsoft.com/textbook/stpowan.html).
2.3 Trade-offs and pitfalls

There are several pitfalls that can and should be avoided when establishing object-
ives in field studies (Bardwell 1991). The most common ones are: (1) addressing
the wrong problem; (2) stating the problem in a way that no solution is possible;
(3) prematurely accepting a solution as the only possible answer; and (4) using
data and information that are either incorrect (e.g. inaccurate information in the
scientific literature) or irrelevant. There are several additional pitfalls that may be
avoided by careful planning and a thorough understanding of the questions to be
asked (Tucker et al. 2005). Four of the most frequent are listed below.
1) The statistical framework might be inadequate, since many techniques
developed in the context of controlled experimentation are sometimes
incorrectly applied to field data, resulting in an inappropriate use of the
null hypothesis (Johnson 1999).
2) Researchers and technicians might differ in their skills, use non-comparable
methods, or have different personal goals. Training prior to the start of field
collection of data and a comparison of each person’s abilities helps to min-
imize these problems.
3) Methods may be changed during a study. This could lead to an incompati-
bility of data sets and limit the interpretation of results.
4) The locations of permanent sample sites are not properly recorded so that
different areas are subsequently revisited or sampled.
When designing fieldwork, researchers need to be aware of potential options
and trade-offs, and try to balance them (Hairston 1989). Examples include:
(1) complexity versus simplicity (e.g. choices ranging from theoretical models to
field experiments), (2) confidence in results versus general application (e.g. high
confidence can be achieved at short temporal and small spatial scales and with
relatively simple goals and conceptual models, but the results will be of limited
value), and (3) replication versus sophistication of experimental design, recog-
nizing that it is impossible to simultaneously maximize precision, realism, and
generality (Levins 1968).
2.4 Ethical issues

While ecological field studies have generated a wealth of useful and important
data, the environmental impacts of these studies have rarely been quantified. A
basic assumption is that the relative benefits of research outweigh the potential
short-term costs to the study animals or habitats (Farnsworth and Rosovsky
1993). Even the simple act of observation, however, may affect the behavior
of the study organisms. Repeatedly visiting different sites during fieldwork
can have negative impacts, either by spreading or introducing nonindigenous
species (Whinam et al. 2005) or diseases, such as amphibian chytrid fungus
(Garner et al. 2005), or by microhabitat alteration (e.g. turning logs). Preventive
measures, such as equipment disinfection (Chapter 26) and routine checks for
unwanted “passengers” (e.g. seeds) are now widely used (ARG-UK 2008).
Controversies continue regarding the negative effects of some common tech-
niques. For example, toe-clipping, historically one of the most widely used mark-
ing techniques, has recently received much criticism (Chapter 8; May 2004;
McCarthy and Parris 2004). Critics, however, have not provided much evalu-
ation of the impacts of alternative procedures. Apart from dorsal or ventral pat-
tern mapping by photography or computer imaging, which are non-invasive, all
other marking techniques have disadvantages (Phillott et al. 2007). The effects
of toe-clipping vary among species, and therefore must be assessed accordingly.
Other marking methods for amphibians are available, although they are not as
economic or as easy to use, but which have fewer risks (Chapter 8; Ferner 2007).
Toe-clipping also may be prohibited by regulatory constraints; information on
regulations is best obtained and evaluated during the planning stage. Thus, there
may be a trade-off between the risks associated with methodology and the know-
ledge to be gained, even when the species may be benefited (Funk et al. 2005).
Field studies often involve years of hard work. In the end, the results may
provide few insights compared to the amount of effort to acquire them, unless
careful planning precedes the initiation of research activities. Careful planning
optimizes researcher effort and helps ensure that the data recorded will be stat-
istically accurate, with beneficial results in advancing knowledge of amphibian
ecology and conservation biology.
2.5 Acknowledgments
Cristina Vâlcu, Tibor Hartel, Marian Griffey, and Ken Dodd provided helpful
comments on previous versions of this chapter.
2.6 References
Alford, R.A. and Richards, S.J. (1999). Global amphibian declines: a problem in applied
ecology. Annual Review of Ecology and Systematics, 30, 133–65.
ARG-UK (Amphibian and Reptile Groups of the UK) (2008). Amphibian Disease
Precautions: a Guide for UK Fieldworkers. ARG-UK Advice Note 4, pp. 1–5. www.
arg-uk.org.uk/Downloads/ARGUKAdviceNote4.pdf.
Arntzen, J. W., Oldham, R. S., and Latham, D. M. (1995). Cost effective drift fences for
toads and newts. Amphibia-Reptilia, 16, 137–45.
Arntzen, J. W., Goudie, I. B., Halley, J.J., and Jehle, R. (2004). Cost comparison of
marking techniques in long-term population studies: PIT-tags versus pattern maps.
Amphibia-Reptilia, 25, 305–15.
Bardwell, L. V. (1991). Problem-framing: a perspective of environmental problem solv-

ing. Environmental Management, 15, 603–12.
Becker, C. G., Fonseca, C. R., Haddad, C. F. B., Batista, R. F., and Prado, P. I. (2007).
Habitat split and the global decline of amphibians. Science, 318, 1775–7.
Bernstein, B. B., and Goldfarb, L. (1995). A conceptual tool for generating and evaluating
ecological hypotheses. BioScience, 45, 32–9.
Farnsworth, E. J., and Rosovsky, J. (1993). The ethics of ecological field experimentation.
Conservation Biology, 7, 463–72.
Fauth, J. E. (1998). Investigating geographic variation in interspecific interactions using
common garden experiments. In W. J. Resetaridis Jr and J. Bernardo (eds), Experimental
Ecology, Issues and Perspectives, pp. 394–415. Oxford University Press, New York.
Ferner, J. W. (2007). A Review of Marking and Individual Recognition Techniques for
Amphibians and Reptiles. Herpetological circular no. 35. Society for the Study of
Amphibians and Reptiles, Salt Lake City, UT.
Ford, E. D. (2000). Scientific Method for Ecological Research. Cambridge University Press,
Cambridge.
Frankham, R. and Brook, B. W. (2004). The importance of time scale in conservation
biology and ecology. Annales Zoologici Fennici, 41, 459–63.
Funk, W. C., Donnelly, M. A., and Lips, K. R. (2005). Alternative views of amphibian
toe-clipping. Nature, 433, 193.
Garner T. W. J., Walker, S., Bosch, J., Hyatt, A.D., Cunningham, A.A., and Fisher, M.C.
(2005). Chytrid fungus in Europe. Emerging Infectious Diseases, 11, 1639–41.
Hairston, N. G. (1989). Hard choices in ecological experimentation. Herpetologica, 45,
119–22.
Halliday, T. (2005). Diverse phenomena influencing amphibian population declines. In
M. Lannoo (ed.), Amphibian Declines: the Conservation Status of United States Species,
pp. 3–6. University of California Press, Berkeley, CA.
Hayek, L. A. (1994). Research design for quantitative amphibian studies. In W. R. Heyer,
M. A. Donnelly, R. W. McDiarmid, L. A. Hayek, and M. Foster (eds), Measuring
and Monitoring Biological Diversity. Standard Methods for Amphibians, pp. 21–39.
Smithsonian Institution Press, Washington.
Jaeger, R. G. (1972). Food as a limited resource in competition between two species of
terrestrial salamanders. Ecology, 53, 535–46.
Jaeger, R. G. and Halliday, T. R. (1998). On confirmatory versus exploratory research.
Herpetologica, 54 (supplement), 64–6.
James, F. C. and McCulloch, C. E. (1985). Data analysis and the design of experiments
in ornithology. In R. F. Johnston (ed.), Current Ornithology, vol. 2, pp. 1–63. Plenum
Press, New York.
Johnson, D. H. (1999). The insignificance of statistical significance testing. Journal of
Wildlife Management, 63, 763–72.
Krebs, C. J. (1999). Ecological Methodology, 2nd edn. Benjamin/Cummings, Menlo Park, CA.
Lehner, P. N. (1996). Handbook of Ethological Methods, 2nd edn. Cambridge University
Press, Cambridge.
Levins, R. (1968). Evolution in Changing Environments. Princeton University Press,
Princeton, NJ.
Maher, W. A., Cullen, P. W., and Norris, R. H. (1994). Framework for designing sam-
pling programs. Environmental Monitoring and Assessment, 30, 139–62.
May, R. M. (2004). Ethics and amphibians. Nature, 431, 403.
McCarthy, M. A. and Parris, K. M. (2004). Clarifying the effect of toe clipping on frogs
with Bayesian statistics. Journal of Applied Ecology, 41, 780–6.
Meyer A. H., Schmidt, B.R., and Grossenbacher, K. (1998). Analysis of three amphibian
populations with quarter-century long time-series. Proceedings of the Royal Society of
London Series B Biological Sciences 265, 523–8.
Michener, W. K. (2000). Research design: translating ideas to data. In W. K. Michener
and J. W. Brunt (eds), Ecological Data: Design, Management and Processing, pp. 1–24.
Blackwell Science, Oxford.
Morrison, M. L. (2002). Wildlife Restoration. Techniques for Habitat Analysis and Animal
Monitoring. Society for Ecological Restoration. Island Press, Washington DC.
Pechmann, J.H.K., Scott, D.E., Semlitsch, R.D., Caldwell, J.P., Vitt, L.J., and Gibbons,
J.W. (1991). Declining amphibian populations: the problem of separating human
impacts from natural fluctuations. Science, 253, 892–5.
Pellet J., Schmidt, B.R., Fivaz, F., Perrin, N., and Grossenbacher, K. (2006). Density, cli-
mate and varying return points: an analysis of long-term population fluctuations in the
threatened European tree frog. Oecologia, 149, 65–71.
Phillott, AD, Skerratt, L. F., McDonald, K. R., Lemckert, F. L., Hines, H. B., Clarke, J. M.,
Alford, R. A., and Speare, R. (2007). Toe-clipping as an acceptable method of identifying
individual anurans in mark recapture studies. Herpetological Review, 38, 305–8.
Piotrow, P. T., Kincaid, D. L., Rimon, J. G., and Rinehart W. (1997). Health
Communication: Lessons from Family Planning and Reproductive Health. Johns Hopkins
School of Public Health, Center for Communication Programs. Praeger Publishers,
Westport, CT.
Pollock, K. H., Nichols, J. D., Simons, T. R., Farnsworth, G. L., Bailey, L. L., and Sauer
J. R. (2002). Large scale wildlife monitoring studies: statistical methods for design and
analysis. Environmetrics, 13, 105–19.
Prathapan, K. D., Rajan, P. D., Narendran, T. C., Viraktamath, A. C., Aravind, N. A., and
Poorani, J. (2008). Death sentence on taxonomy in India. Current Science, 94, 170–1.
Rowe, C. L. and Dunson, W. A. (1994). The value of simulated pond communities in
mesocosms for studies of amphibian ecology and ecotoxicology. Journal of Herpetology,
28, 346–56.
Schneider, D. C. (2001). The rise of the concept of scale in ecology. BioScience, 51,
545–53.
Scott, C. T. and Köhl, M. (1993). A method of comparing sampling design alternatives
for extensive inventories. Mitteilungen der Eidgengessischen Anstalt fuer Wald, Schneeund
Landschaft, Birmensdorf, 69, 1–62.
Senar, J. C. (2004). Mucho mas que plumas. Monografies del Museu de Ciències Naturals
2. Museu de Ciències Naturals i l’Institut Botànic de Barcelona, Barcelona.
Stuart, S. N., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues, A. S. L., Fischman, D. L.,
and Waller, W. (2004). Status and trends of amphibian declines and extinctions world-
wide. Science, 306, 1783–5.
Tucker, G., Bubb, P., de Heer, M., Miles, L., Lawrence, A., Bajracharya, S.B., Nepal, R.C.,
Sherchan, R., and Chapagain, N. R. (2005). Guidelines for Biodiversity Assessment and
Monitoring for Protected Areas. KMTNC, Kathmandu, Nepal.
Underwood, A. J. (1997). Experiments in Ecology. Their Logical Design and Interpretation
Using Analysis of Variance. Cambridge University Press, Cambridge.
Weir, L. A., Royle, A., Nanjappa, P., and Jung, R. E. (2005). Modeling anuran detection
and site occupancy on North American Amphibian Monitoring Program (NAAMP)
routes in Maryland. Journal of Herpetology, 39, 627–39.
Whinam, J., Chilcott, N., and Bergstrom, D. M. (2005). Subantarctic hitchhikers: expe-
ditioners as vectors for the introduction of alien organisms. Biological Conservation,
121, 207–19.
Wilbur, H. M. (1997). Experimental ecology of food webs: complex systems in temporary
ponds. Ecology, 78, 2279–2302.
Wolff, J. O. and Krebs, C. J. (2008). Hypothesis testing and the scientific method revis-
ited. Acta Zoologica Sinica, 54, 383–6.
Yoccoz, N. G. (1991). Use, overuse and misuse of significance tests in evolutionary biol-
ogy and ecology. Bulletin of the Ecological Society of America, 72, 106–11.
Part 2
Larvae
3
Morphology of amphibian larvae
Roy W. McDiarmid and Ronald Altig
3.1 Background
The larvae of amphibians are non-reproductive and usually aquatic. Most
undergo metamorphosis prior to attaining an adult morphology and sexual
maturity. Species within each amphibian order that develop by other modes (e.g.
direct development; Altig and Johnston 1989; Thibaudeau and Altig 1999) have
non-feeding larvae or embryos, and we do not discuss them in this chapter.
Amphibian larvae have some generalized morphological features that are use-
ful for identification. In contrast to fish, they lack bony supports in the tail fins,
and the vent is a longitudinal slit (not obvious in tadpoles). Amphibian larvae
also lack eyelids, and most have external gills that are visible at some stage in
their ontogeny.
Most caecilians (Gymnophiona) whether terrestrial or aquatic as adults,
have aquatic larvae that look grossly like the legless, elongate adults. The larvae
of salamanders (Caudata) look much like the adults. Unlike the condition in
caecilians and frogs, salamanders may occur in larval (i.e. larval morphology,
non-reproductive, and metamorphic) or larviform (i.e. larval morphology,
reproductive, may or may not metamorphose) states. Larviform salamanders
may exist as pedotypes (i.e. larval relative to the normal developmental tra-
jectory of the taxon, reproductive, will metamorphose if the environmental
conditions change to the detriment of being in the larval environment; some
ambystomatids and salamandrids; terminology of Reilly et al. 1997) or pedo-
morphs (i.e. larval relative to the developmental pattern of their ancestors, do
not metamorphose; all amphiumids, cryptobranchids, proteids, sirenids, and
some plethodontids). Larval and larviform salamanders grossly resemble meta-
morphosed individuals in general body form but retain a number of larval fea-
tures. Pond-adapted forms have a more bulky body and larger gills and tail fins
than the more streamlined, stream-adapted forms.
The larvae of frogs and toads (Order Anura), called tadpoles, are grossly dif-
ferent from adults and have many developmental (Altig and Johnston 1989) and
morphological (e.g. Altig and McDiarmid 1999a; also various morphologies doc-
umented in staging tables, see Duellman and Trueb 1986, pp. 128–9) features not
seen in other amphibian larvae. Tadpoles live in many kinds of habitats; the most
common types of tadpoles are found in lentic or lotic water, spend most of their
time on the bottom, and feed by rasping material from submerged surfaces.
Because amphibians are ectotherms, their inherent developmental rates are
modified by temperature and other environmental variables; size is thus an
inaccurate estimator of chronological age. Consequently, biologists describe
tadpole ontogeny using a staging table that divides their development into rec-
ognizable stages based on the attainment of specific morphological landmarks.
With such a table the degree of development of morphological features of tad-
poles can be compared among populations and across taxa occurring in the
same or different habitats regardless of chronological age or attainable size.
Larval amphibians are exceptionally variable within and among species, although
the degree and patterns of that variation are poorly documented and their sources
rarely investigated. The many papers on induced morphological changes published
in recent years (e.g. Relyea and Auld 2005, among many others) have made it abun-
dantly clear that every tadpole of a given taxon collected at any site is a variant
within the broad phenotypic range of its taxon. Although results from mesocosm
experiments with controlled combinations and densities of species provide some
insight into understanding phenotypic variation, predicting morphological vari-
ation from random samples of ponds is highly unlikely. The presence of different
sets of predators in natural situations complicates the picture even more, and one
has to keep these factors in mind when evaluating the morphology of tadpoles.
We mention in passing that less is known about amphibian eggs (e.g. Altig
and McDiarmid 2007), hatchlings (Gosner stages 21–24; Altig 1972), and
metamorphs than is known about tadpoles (stages 25–41) and other amphibian
larvae. We urge workers to preserve and describe positively identified samples of
these stages. Here we summarize data on the morphology, ontogeny, and diver-
sity of larvae in each amphibian order.
3.2 Larval caecilians

3.2.1 Morphology and ontogeny
The vermiform body (Figure 3.1) has primary annuli homologous to the costal
folds of salamanders that form during early development; secondary and tertiary
annuli may form later. External gills with branched rami (Figures 3.1a and c) are
3 Morphology of amphibian larvae | 41
(a) (b)
(c) (d)
Fig. 3.1 Morphology of a larval caecilian. (a) Embryo of Ichthyophis glutinosus,

modified from figure in Sarasin and Sarasin (1887–1890); (b) heads of embryos of
Ichthyophis kohtaoensis, stages 27 (left) and 33 (right), modified from drawings in
Dünker et al. (2000); (c) hatchling of I. kohtaoensis, modified from photograph by
A. Summers on front cover, Journal of Morphology 2000, 243(1); and (d) larva of
Ichthyophis banannicus showing the gill slit (upper arrow) and tail fin (lower arrow),
photograph by R. Nussbaum.
present during embryological development and lost at hatching (Figure 3.1b).

The single gill slit (Figure 3.1d) closes at metamorphosis, and a small fin is present
on a short tail (Figure 3.1d). The sensory tentacle, situated adjacent or anterior
to the eye of caecilians, develops at metamorphosis. Some caecilians have small
bony scales (i.e. osteoderms) embedded in the skin in the grooves between the
annuli. These scales that are quite small when they first appear, get progressively
larger as they develop, beginning in the posterior annuli and moving anteriorly.
The definitive pattern of occurrence varies among species. Neuromasts, or lat-
eral line organs, are present throughout larval life; labial folds, which modify the
shape and size of the mouth opening during suction feeding, are present as they
are in larval salamanders. The most complete information on early ontogeny
can be found in Dünker et al. (2000) (also see Sarasin and Sarasin 1887–1890;
Brauer 1897, 1899; Sammouri et al. 1990).
3.2.2 Coloration
Most caecilians are somewhat drab shades of gray, brown, black, or blue. A few
species are more brightly colored and slightly banded or striped (terminology of
Altig and Channing 1993).
3.2.3 Diversity
Larvae of species of caecilians in the families Rhinatrematidae, Ichthyophiidae,
and some Caeciilidae that are known are similar to and grossly resemble adults
in general morphology. The paucity of ontogenetic data, however, makes fur-
ther comparisons impossible.
3.3 Larval and larviform salamanders

With the exceptions of pedomorphs in the families Amphiumidae (cylin-
drical, elongate body with four tiny limbs, each with one to three tiny digits)
and Sirenidae (cylindrical, elongate bodies with front limbs only, each with
three or four fingers), salamander larvae have a typical quadruped morphology
(Figure 3.2a) similar to that of the adults once all four limbs develop. Costal
grooves divide the myotomic muscle bundles of the trunk into costal folds. Gills
of various shapes are often prominent (Figures 3.2a, b, and d), one to four gill
slits open during larval life eventually close in metamorphic taxa, gill rakers are
prominent to absent, and a gular fold that typically is free from adjacent throat
tissue is usually present. Fleshy labial folds modify mouth size and shape during
suction feeding, and neuromasts are present, although sometimes difficult to
see without special techniques (Lannoo 1985). Major metamorphic modifica-
tions include the loss of tail fins, gills, gill slits, gular folds, and neuromasts,
the development of eyelids, and many other integumentary, osteological, and
physiological changes.
Pond-inhabiting larvae (e.g. most species of Ambystomatidae and Hynobiidae)
have robust bodies and heads, tall dorsal fins that can originate as far anterior
as the back of the head, and long gill rami with plumose fimbriae. Stream-
inhabiting larvae (e.g. certain species of Hynobiidae, Plethodontidae, and some
Salamandridae) are more streamlined; they have low fins that usually originate
near the tail/body junction, and shorter, less plumose gills. In species within the
Rhyacotritonidae and desmognathine Plethodontidae, the gill rami are excep-
tionally short with few fimbriae. Gill rami are branched in species in the fam-
ilies Amphiumidae (gills lost soon after hatching) and Sirenidae (gills persist
(a) Gm Df Vf
TL
Gf Cf Cg
V
SVL (d)
Gm
(b) (c)
Lb N
Lf
Gf
Gm
B Gr
Ll
B Is
Fig. 3.2 Measurements and body parts of a larval salamander. (a) Lateral (upper)
and ventral (lower) views of a typical Ambystoma salamander larvae, drawing by
D. Karges; (b) dorsal view of the head and anterior body region of a hatchling
Ambystoma maculatum, photograph by A.M. Richmond; (c) ventral view of head
of A. maculatum; and (d) stylized drawing of gill structure of a larval salamander,
modified from drawing in Pfingsten and Downs (1989). B, balancer; Cf, costal fold;
Cg, costal groove; Df, dorsal fin; Gf, gular fold; Gm, gill ramus with fimbriae attached
to posterior surface; Gr, gill rakers; Is, interbranchial septum; Lb, limb bud; Lf, labial
fold; Ll, lateral line organs (neuromasts); N, naris; SVL, snout–vent length; TL, total
length; V, vent; Vf, ventral fin.
throughout life although they atrophy during aestivation); rami in larvae of

all other families are non-branched. Larvae of sirenid salamanders have low,
fleshy, pigmented fins restricted to the tail, as in adults, and hatchlings have a
transparent dorsal fin that originates well forward on the body. Amphiumid
larvae have a very short, low caudal fin that is lost soon after hatching. All other
salamander larvae have tail fins throughout their ontogeny. Regression of the
dorsal fin usually starts long before other metamorphic changes are noticeable.
Fleshy flaps occur on the trailing edges of the hind legs of Onychodactylus lar-
vae. Keratinized toe tips are found in a number of taxa but are most common in
stream inhabitants. Sirenids have keratinized jaw sheaths (upper and lower) as
do some ambystomatid larvae (lower).
Hatching occurs before the limbs are fully developed, and the front limbs
usually develop faster than the hind ones. Some hatchlings (e.g. in the families
Ambystomatidae and Salamandridae) have a balancer, a fleshy projection on
each side of the lower part of the head (Figures 3.2b and c), that is lost soon after
hatching. Staging tables (compilation in Duellman and Trueb 1986, p. 128)
are available for several species (e.g. Ambystoma maculatum, Harrison 1969;
Ambystoma mexicanum, Cano-Martinez et al. 1994; and Hynobius nigrescens,
Iwasawa and Yamashita 1991).
3.3.2 Coloration
In contrast to Stereochilus marginatus (Plethodontidae), Rhyacotriton spp.
(Rhyacotritonidae), and most pedomorphs, most of which retain something
similar to the larval coloration as adults, larval salamanders often have a color-
ation distinct from that of the metamorph or the adult. Larval sirenids are jet
black with contrasting stripes and bands of red or yellow, while adults have either
a gray or black ground color usually overlain by speckles and small blotches of
gold to greenish iridophores. Color and pattern (i.e. coloration) in larvae of most
species can vary considerably during ontogeny, throughout a day, and among
sites in response to substrate color, temperature, and water clarity. Colors are
typically muted grays, browns, and blacks, and patterns range from none (uni-
colored), blotched, and mottled through striped (longitudinal or diagonal con-
trasting markings) and banded (transverse contrasting markings). The dorsum
of the tail muscle of small Ambystoma is often banded, and the pattern may
be retained throughout ontogeny (e.g. Ambystoma talpoideum) or change to a
totally different pattern sometime after hatching and then again after metamor-
phosis. In some species and populations, larval Desmognathus (Plethodontidae)
have a distinct pattern that is kept throughout life. Although colors are usually
more muted, the adult patterns in other salamanders (e.g. Ambystoma tigrinum
group, many plethodontids) may appear at metamorphosis or a different pattern
may appear (e.g. most Ambystoma) after metamorphosis and slowly develop into
the adult pattern, which is achieved long before sexual maturity.
3.3.3 Diversity
Larval salamanders show much less ecomorphological diversity than tadpoles.
By definition, pedotypes and pedomorphs retain a larval morphology even
though they become reproductive, and species that metamorphose and are
adapted for either pond or flowing water are the most easily recognized groups.
Cannibal morphotypes with enlarged heads and altered dentition occur in some
species (Ambystomatidae, Hynobiidae). Pond inhabitants often do not over-
winter, whereas some stream inhabitants may grow as larvae for several years
before undergoing metamorphosis. In some parts of their range Notophthalmus
viridescens (Salamandridae) larvae metamorphose into a brilliantly colored eft
that lives on the forest floor for several years before returning to the ponds for an
aquatic existence as a reproductive adult.
3.4 Anuran tadpoles

The transition between the body and tail across all stages and taxa of tadpoles is
difficult to define. The demarcation between the two is most consistently and
accurately described as the juncture of the axis of the tail myotomes with the
posterior body surface (Figure 3.3a, bottom). The tails of most tadpoles lack
vertebrae and are composed of dorsal and ventral fins and a long series of pro-
gressively smaller myotomic muscle bundles surrounding the notochord. The
tadpoles of some species of Megophryidae do have tail vertebrae (e.g. Haas et al.
2006; Handrigan et al. 2007). The shapes and extents of the fins vary among
taxa and habitats (i.e. tallest in pond dwellers, lowest in fast-water and semi-
terrestrial forms). The eyes are either dorsal (i.e. lie totally within the dorsal sil-
houette; Figure 3.3b, left) or lateral (i.e. included as part of the dorsal silhouette;
Figure 3.3b, right). All free-living tadpoles have a spiracle(s) through which water
that has been pumped in through the mouth by the buccopharyngeal muscula-
ture and passed over the gills and food filtering system passes out of the body. In
the vast majority of species, the spiracle is single and situated somewhere on the
left side of the body (Figure 3.3a). Tadpoles of Ascaphus (Leiopelmatidae) have
a single, ventral spiracle on the chest, while those of Bombina (Bombinatoridae)
and Alytes and Discoglossus (Alytidae) have a single spiracle located almost
midventrally on the abdomen. In microhylid larvae, the midventral spiracle is
located at the posterior part of the abdomen or near the vent. The tadpoles of
pipids, rhinophrynids and the leptodactylid genus Lepidobatrachus have dual
lateral spiracles. The spiracles of Lepidobatrachus develop differently from the
other two taxa (Ruibal and Thomas 1988).
Scent-laden water passes through the nares, over the olfactory epithelium of
the nasal sacs, and into the buccal cavity via the internal nares. The shape of the
external apertures varies from round to elliptical and may have a variety of papillae
(a)
IND
TMW
IOD
BL TaL
TMH MTH
TL
OD SP LB VT TMA
(b) (c)
Fig. 3.3 Measurements and body parts of a tadpole. (a) Dorsal (upper) and lateral
(lower) views of a typical tadpole, drawing of Rana sp. by D. Karges; (b) dorsal (left)
and lateral (right) eye positions of tadpoles, stylized drawings by D. Karges; and
(c) examples of medial (left, Bufo boreas) and dextral (right, Rana catesbeiana) vent
tubes; white arrow, plane of ventral fin; black arrow, outflow of vent tube. BL, body
length; IND, internarial distance; IOD, interorbital distance; LB, hind limb bud; MTH,
maximum tail height; OD, oral disc; SP, spiracle; TMA, tail muscle axis; TMH, tail muscle
height; TMW, tail muscle width; TL, total length; TaL, tail length; VT, vent tube.
associated with the margins. A vent tube extends posteriorly from the midventral
abdomen. Two major types are recognized: dextral, where the aperture lies to the
right of the sagittal plane of the tail fin (e.g. hylids and ranids), and medial, where
the aperture lines parallel with the plane of the tail fin (e.g. bufonids and scaphi-
opodids; Figure 3.3c). As with the spiracular tube configurations, there are many
subtle variations in the shape and position of the vent tube.
The lateral line system (i.e. neuromasts; Hall et al. 2002; Lannoo 1985) is
composed of many depressions in the skin with sensory cells in the center that
signal the patterns of water flow over various parts of the body and tail. The
distribution and arrangement of neuromasts may be useful in distinguish-
ing between closely related species. In darkly pigmented tadpoles these sites
are often pale and obvious, but evaluation of stitch patterns in most tadpoles
requires separating the epidermis from the underlying dermis, clearing in gly-
cerin, and viewing the skin with dark-field illumination (see Lannoo 1985).
The oral apparatus, the composite of upper and lower labia and all kerati-
nized mouthparts, is highly variable across taxa and ecological types. The most
common oral apparatus (Figure 3.4a and c) occurs in many taxa in lentic and
lotic sites. An assembly of the two infralabial and two Meckel’s cartilages with
three joints forms the lower jaw that is surmounted by a serrated, keratinized
jaw sheath. The supralabial cartilage of the upper jaw is surmounted by a simi-
lar keratinized jaw sheath, and during a bite, the lower jaw passes totally behind
the upper (Figure 3.4d). The interactions of the serrated margins of the sheaths
serve as cutting/gouging surfaces when a tadpole feeds. The highly variable
shapes of the jaw sheaths suggest different feeding abilities.
The face of the oral disc has fleshy transverse tooth ridges (Figure 3.4a,
c, and d) surmounted by a row(s) of keratinized labial teeth. In most cases,
several replacement teeth are interdigitated below a presently erupted tooth
(Figure 3.4b, lower right), and they successively move into position as the
erupted tooth wears out. The tooth rows are numbered from the anterior edge
of the disc toward the mouth on the upper labium and from the mouth to
the posterior edge of the disc on the lower labium. A fractional designation
indicates the number of tooth rows on each labium; some rows have naturally
occurring medial gaps denoted parenthetically. For example, a Labial Tooth
Row Formula (LTRF) of 2(2)/3(1) indicates two upper rows with a gap in the
second one and three lower rows with a gap in the first one (Figures 3.4a and
c). Some tadpoles lack tooth rows (i.e. 0/0), and the maximum LTRF known
is 17/21 in a tadpole of an undescribed hylid frog from the Guayana Highlands
of southern Venezuela.
The papillate margins of the oral disc may be complete and encircle the disc
(e.g. tadpoles of Scaphiopodidae, Pelobatidae, and many stream-inhabiting tad-
poles of several families), have a medial dorsal gap (most common; Figure 3.4a),
or have both dorsal and ventral gaps (e.g. Bufonidae and those of some Hylidae,
Mantellidae, Ranidae, and Rhacophoridae; Figure 3.4c). Although the number
of rows of papillae on different parts of the disc margin may vary, the lengths of
the papillae are typically somewhat uniform; several species of Phrynobatrachus
(Ranidae) have exceptionally elongate papillae along the posterior margin of the
disc. Submarginal papillae occur on the face of the disc away from the margin
(a) (b)
UJS LJS
A-1 A-2
G
SRC
MZJ
MP IRC
A-1
UJS
LJS P-1
P-2
MZT P-3
TS
C
H
TR S-1
B
P-1 P-2 P-3 TR SP EM S-2
(c) (d)
1
2
Fig. 3.4 Components of the oral disc of a tadpole. (a) Oral disc of a typical tadpole,
schematic drawing by D. Karges; (b) sagittal sections of the oral apparatus of a common
benthic tadpole (upper) and a tooth ridge (lower left), schematic drawings modified
from ones in Heron-Royer and Van Bambeke (1889); two labial teeth of Hyla chrysoscelis
(lower right) removed from a tooth series and in natural position; (c) scanning electron
photomicrograph of the oral apparatus of a tadpole of Bufo fowleri, actual oral disc
width, about 2.3 mm, photograph by M. Penuel-Matthews; and (d) lateral view of
the mouthparts of H. chrysoscelis (1, upper jaw sheath; 2, lower jaw sheath; 3, lower
tooth rows; 4, upper tooth row). A-1, A-2, anterior tooth rows 1 and 2; P-1, P-2,
P-3, posterior tooth rows 1 to 3; S-1, S-2, sheaths of presently erupted (1) and first
replacement (2) teeth; B, body of first replacement tooth; C, cusps on first replacement
tooth; EM, lateral emargination in oral disc; G, dorsal gap in marginal papillae; H, head of
presently erupted tooth; IRC, infrarostral cartilage; LJS, lower jaw sheath; M, mouth; MP,
marginal papillae; MZJ, mitotic zone for production of jaw sheath; MZT, mitotic zone for
production of labial teeth; SP, submarginal papillae; SRC, suprarostral cartilage; TR, tooth
ridge; TS, tooth series; UJS, upper jaw sheath.
and form various patterns (Figure 3.4a). The margin of the disc may be emar-
ginate (i.e. indented; Figure 3.4a) or not.
Hatchlings (Gosner 1960; stages 21–24) usually have external gills but lack
eyes and limb buds. The forelimb buds develop beneath the operculum after it
closes, and the hind-limb buds grow from the posteroventral intersection of the
body and tail muscle (Figure 3.3a, bottom). Staging tables have been made for
a number of taxa (Duellman and Trueb 1986, pp. 128–129), but using a com-
mon table allows for meaningful comparisons among taxa. Gosner (1960; gen-
eral) and Nieuwkoop and Faber (1956; Xenopus) are the two most commonly
cited. Recent summaries of tadpole morphology and terminology are included
in Altig (2007b) and Altig and McDiarmid (1999a, 1999b).
3.4.2 Coloration
Except for notations in descriptions, surprisingly little has been written
about tadpole coloration. As in other larval amphibians, three basic popula-
tions of pigment-containing cells interact to produce both color and pattern.
Melanophores contain melanins that produce browns and blacks, iridophores
contain reflective guanine crystals that produce whites and silvers, and xan-
thophores contain carotenoids that produce yellows and reds. The pigments
are retained inside the cells and can be dispersed in various patterns under the
influence of temperature and light. Altig and Channing (1993) summarized the
diversity of colorations in tadpoles, and Caldwell (1982) tested the functions of
coloration in tadpoles experimentally.
3.4.3 Diversity
Most morphological characters of tadpoles reflect their ecology. Suctorial tad-
poles in a number of families have streamlined bodies and mouthparts modi-
fied to maintain position in fast-flowing water as they feed. A typical increase
in the number of tooth rows, to a maximum of 17/21, is usually accompanied
by a larger oral disc with complete marginal papillae. Other unusual morpho-
logical structures found in stream-inhabiting tadpoles include a belly modified
as a sucker (a few bufonid and ranid species), and lateral sacs (or lymphatic
sacs) on the ventrolateral parts of the body of other stream-dwelling tadpoles
(Arthroleptidae). Tadpoles of Mertensophryne (Bufonidae) have a hollow crown
on the head that encircles the eyes and nares, and tadpoles of Schismaderma
carens (Bufonidae) have a semicircular, transverse flap of skin behind the eyes.
Suspension-feeding tadpoles in the families Microhylidae, Pipidae, and
Rhinophrynidae have reduced, soft mouthparts that lack keratinized struc-
tures. They usually hang in midwater and capture suspended particles as water
is pumped in through the mouth and out the spiracles. Even so, not all tadpoles
that lack keratinized mouthparts are suspension feeders, and tadpoles with
keratinized mouthparts that are infected with the amphibian chytrid fungus
(Batrachochytrium dendrobatidis) often lose most or all such structures.
Carnivores and other macrophagous feeders have a diversity of mouth-
parts related to how they feed. For example, tadpoles of the leptodactylid frog
Lepidobatrachus have huge mouths but almost no soft or keratinized mouth-
parts; they engulf entire organisms, including other tadpoles. Tadpoles of
other leptodactylid frogs, Ceratophrys spp., have huge jaws and many tooth
rows and efficiently tear their victims to pieces. Carnivorous tadpoles of Spea
(Scaphiopodidae) feed similarly. A number of other tadpoles (e.g. Hylidae: Hyla
leucophyllata group; Ranidae: Occidozyga) lack all or most tooth rows but have
huge jaw sheaths. Tadpoles that occupy tree holes and bromeliad tanks (e.g.
some species of Dendrobatidae, Hylidae, and Rhacophoridae) are of several
morphological types (Lannoo et al. 1987; Lehtinen et al. 2004) and have differ-
ent diets; some are non-feeding, several are detritovores or macrophagous car-
nivores, many eat frog eggs (fertilized or not) of their own (cannibals) or other
species, and some eat only trophic eggs supplied by their mother. Surface-film
feeders have large oral discs, but keratinized structures are reduced or absent;
the disc is turned upward (i.e. umbelliform) and captures material carried in
the surface film. The oral discs of these surface-feeding tadpoles may be formed
primarily from the lower labium (e.g. Microhylidae) or from parts of both labia
(e.g. Megophryidae). This convergent morphology occurs in six families, and
most tadpoles of this sort occur only in the slow reaches of streams.
Attempts have been made to define ecomorphological guilds or groups of taxa
with suites of common morphological characters that are presumed to reflect
a common ecology (e.g. Altig and Johnston 1989). Because of the lack of eco-
logical data for many species and an incomplete understanding of how some of
their morphologies actually function, we advise caution in assigning species to
specific guilds without some knowledge of their natural history. For example,
the morphologies of Mantidactylus lugubris (Altig and McDiarmid 2006) and
of some taxa that occur in phytotelms suggest that one might find them in fast-
flowing water. In fact, tadpoles of M. lugubris live in leaf packs in slow-flowing
water.
3.5 Summary
Amphibian larvae show considerable morphological diversity from the relatively
conserved forms of caecilians and salamanders to the unusual and often novel struc-
tures found in tadpoles of frogs and toads. The extreme variability of tadpoles is
almost certainly a product of ontogenetic challenges, the recently discovered influ-

ences of predators and competitors (e.g. Relyea and Auld 2005), and the selective
effects of different habitats. The morphological variability manifest under these
different conditions makes identifications and ecological evaluations especially dif-
ficult. This situation, combined with our lack of understanding of geographic vari-
ation in larval amphibians, especially tadpoles, emphasizes the relative poor state
of our knowledge of larval biology. Much remains to be learned. The discovery of
various anomalies (e.g. Drake et al. 2007), often with a weak understanding of their
causes (Altig 2007a; see also Lannoo 2008), adds yet another impediment to our
total understanding of larval morphology.
3.6 References
Altig, R. (1972). Notes on the larvae and premetamorphic tadpoles of four Hyla and
three Rana with notes on tadpole color patterns. Journal of the Elisha Mitchell Scientific
Society, 88, 113–19.
Altig, R. (2007a). Comments on the descriptions and evaluations of tadpole mouthpart
anomalies. Herpetological Conservation and Biology, 2, 1–4.
Altig, R. (2007b). A primer for the morphology of anuran tadpoles. Herpetological
Conservation and Biology, 2, 73–6.
Altig, R. and Channing, A. (1993). Hypothesis: functional significance of colour and
pattern of anuran tadpoles. Herpetological Journal, 3, 73–5.
Altig, R. and Johnston, G. F. (1989). Guilds of anuran larvae: relationships among devel-
opmental modes, morphologies, and habitats. Herpetological Monographs, 3, 81–109.
Altig, R. and McDiarmid, R. W. (1999a). Body plan: development and morphology. In
R. W. McDiarmid and R. Altig (eds), Tadpoles, the Biology of Anuran Larvae, pp. 24–51.
University of Chicago Press, Chicago, IL.
Altig, R. and McDiarmid, R. W. (1999b). Diversity: familial and generic characteriza-
tions. In R. W. McDiarmid and R. Altig (eds), Tadpoles, the Biology of Anuran Larvae,
pp. 295–337. University of Chicago Press, Chicago, IL.
Altig, R. and McDiarmid, R. W. (2006). Descriptions and biological notes on three
unusual mantellid tadpoles (Amphibia: Anura: Mantellidae) from southeastern
Madagascar. Proceedings of the Biological Society of Washington, 119, 418–25.
Altig, R. and McDiarmid, R. W. (2007). Morphological diversity and evolution of egg
and clutch structure in amphibians. Herpetological Monographs, 21, 1–33.
Brauer, A. (1897). Beiträge zur Kenntnis der Entwicklungsgeschichte und Anatomie der
Gymnophionen. Zoologische Jahrbucher Anatomie, 10, 277–472.
Brauer, A. (1899). Beiträge zur Kenntnis der Entwicklung und Anatomie Gymnophionen.
Zoologische Jahrbucher Anatomie, 12, 477–508.
Caldwell, J. P. (1982). Disruptive selection: a tail color polymorphism in Acris tadpoles in
response to differential predation. Canadian Journal of Zoology, 60, 2818–27.
Cano-Martínez, A., Vargas-González, A., and Asai, M. (1994). Metamorphic stages in
Ambystoma mexicanum. Axolotl Newsletter, 23, 64–71.
Drake, D. L., Altig, R., Grace, J. R., and Walls, S. C. (2007). Occurrence of oral defects
in larval anurans from protected sites. Copeia, 2007, 449–58.
Duellman, W. E. and Trueb, L. (1986). Biology of Amphibians. McGraw-Hill, New York.
Dünker, N., Wake, M. H., and Olson, W. M. (2000). Embryonic and larval development
in the caecilian Ichthyophis kohtaoensis (Amphibia, Gymnophiona): a staging table.
Journal of Morphology, 243, 3–34.
Gosner, K. L. (1960). A simplified table for staging anuran embryos and larvae with notes
on identification. Herpetologica, 16, 183–90.
Haas, A., Hertwig, S., and Das, I. (2006). Extreme tadpoles: the morphology of the fos-
sorial megophryid larva, Leptobrachella mjobergi. Zoology, 109, 26–42.
Hall, J. A., Larsen, Jr, J. H., and Fitzner, R. E. (2002). Morphology of the prometamor-
phic larva of the spadefoot toad, Scaphiopus intermontanus (Anura: Pelobatidae), with
an emphasis on the lateral line system and mouthparts. Journal of Morphology, 252,
114–30.
Handrigan, G. R., Haas, A., and Wassersug, R. J. (2007). Bony-tailed tadpoles: the devel-
opment of supernumerary caudal vertebrae in larval megophryids (Anura). Evolution
and Development, 9, 190–202.
Harrison, R. G. (1969). Harrison stages and description of the normal development
of the spotted salamander, Ambystoma punctatum (Linn.). In R. G. Harrison (ed.),
Organization and Development of the Embryo, pp. 44–6. Yale University Press, New
Haven, CT.
Héron-Royer, L. F. and Van Bambeke, C. (1889). Le vestibule de la bouche chez les
têtards des batraciens anoures d’Europe; sa structure, ses caractères chez les diverses
espèces. Archives de Biologie, 9, 185–309.
Iwasawa, H. I. and Yamashita, K. (1991). Normal stages of development of a hynobiid
salamander, Hynobius nigrescens Stejneger. Japanese Journal of Herpetology, 14, 39–62.
Lannoo, M. J. (1985). Neuromast topography in Ambystoma larvae. Copeia, 1985,
535–9.
Lannoo, M. J. (2008). Malformed Frogs. The Collapse of Aquatic Ecosystems. University of
California Press, Berkeley, CA.
Lannoo, M. J., Townsend, D. S., and Wassersug, R. J. (1987). Larval life in the leaves:
arboreal tadpole types, with special attention to the morphology, ecology, and behavior
of the oophagous Osteopilus brunneus (Hylidae) larva. Fieldiana Zoology, 38, 1–31.
Lehtinen, R. M., Lannoo, M. J., and Wassersug, R. J. (2004). Phytotelm-breeding
anurans: past, present and future research. In R. M. Lehtinen (ed.), Ecology and
Evolution of Phytotelm-Breeding Anurans, pp. 1–9. Miscellaneous Publication 193.
Museum of Zoology, University of Michigan, Ann Arbor, MI.
Nieuwkoop, P. D. and Faber, J. (1956). Normal tables of Xenopus laevis (Daudin). North-
Holland Publishing Company, Amsterdam.
Pfingsten, R. A. and Downs, F. L. (1989). Salamanders of Ohio. Bulletin of the Ohio
Biological Survey, 7, 1–350.
Reilly, S. M., Wiley, E. O., and Meinhardt, D. J. (1997). An integrative approach to
heterochrony: the distinction between interspecific and intraspecific phenomena.
Biological Journal of the Linnean Society, 60, 119–43.
Relyea, R. A. and Auld, J. R. (2005). Predator- and competitor-induced plasticity: how
changes in foraging morphology affect phenotypic trade-offs. Ecology, 86, 1723–9.
Ruibal, R. and Thomas, E. (1988). The obligate carnivorous larvae of the frog,
Lepidobatrachus laevis (Leptodactylidae). Copeia, 1988, 591–604.
Sammouri, R., Renous, S., Exbrayat, J. M., and Lescure, J. (1990). Développement
embryonnaire de Typhlonectes compressicaudus (Amphibia, Gymnophiona). Annales de
Sciences Naturelles-Zoologie et Biologie Animale, Paris, 11, 135–63.
Sarasin, P. and Sarasin, S. (1887–1890). Zur Entwicklungsgeschichte und Anatomie der
Ceylonesischen Blindwühle Ichthyophis glutinosus. In Ergebnisse Naturwissenschaftlicher
Forschungen auf Ceylon in den Jahren 1884–1886. Band II. C. W. Kreidel’s Verlag,
Wiesbaden.
Thibaudeau, D. G. and Altig, R. (1999). Endotrophic anurans: development and evolu-
tion. In R. W. McDiarmid and R. Altig (eds), Tadpoles, the Biology of Anuran Larvae,
4
Larval sampling
David K. Skelly and Jonathan L. Richardson
4.1 Introduction
Most amphibian species are metamorphic, and among those, the majority have
larvae that are fully aquatic (Duellman and Trueb 1986). These larvae are found
in an enormous variety of contexts ranging from bromeliads and tree holes, to
brackish pools and the largest rivers and lakes (Duellman and Trueb 1986). Add
to this mix of environments the fact that amphibian larvae differ dramatically in
their microhabitat use, from the water surface to benthic mud, and researchers
sampling larval amphibians are confronted with a non-trivial challenge of match-
ing techniques with study goals and logistic limits. Fortunately, there is a wealth
of experience that can be tapped when making decisions about where, when, and
how to sample (Shaffer et al. 1994; Olson et al. 1997).
4.1.1 Why sample larvae?

The reasons to sample amphibian larvae are many, but most emphasize either
strategic or logistic considerations. From the strategic perspective, larvae represent
a critical life history stage. Many of the reasons that motivate scientists to study
amphibians are related to the dynamics and fate of their populations. Monitoring
of larval cohorts can provide critical information regarding the trajectory of a
population and the factors that may affect abundance and distribution. Larvae
also represent concrete evidence of breeding. While many studies of anurans use
male calling as an index of breeding distribution, males can call from wetlands
where no breeding takes place. Larval surveys are less likely to overestimate breed-
ing distribution.
For many species larvae are also convenient. Many adult amphibians are found
at low density spread across large areas of terrestrial habitat. In addition they are
often cryptic, arboreal, or fossorial, making it extremely difficult to sample their
populations systematically. For amphibians with fully metamorphic life histories

and aquatic larvae, breeding represents the point in the life cycle where a spe-
cies can be indexed in a relatively small, defined area. The techniques described
below offer a variety of systematic approaches to estimating population attributes
that would be extremely challenging to determine for the adult stages of many
amphibian species.
Because of their strategic and logistic advantages, amphibian larvae have been
used in a wide range of research aimed at addressing fundamental questions in
fields ranging from ecology and evolution to physiology and developmental biol-
ogy. Well before amphibian larvae were widely studied by scientists concerned
about declines and species extinctions, larval amphibians were held up as a model
system by biologists.
4.1.2 Target responses

Once the motivation for sampling larvae is known, the target response vari-
ables may be selected (Table 4.1). Information on species occupancy is frequently
sought in both basic and applied settings. Presence and absence data can be
important in determining species richness within a wetland as well as providing
evidence for changes in species range. Any of the techniques described in this
chapter can be used to estimate species presence and absence. However, if presence/
absence is the sole motivation for sampling, some techniques such as dip-netting
(described below) may be much more efficient than others.
Table 4.1 Larval amphibian sampling methods and resulting response variables
Response Metrics Sampling methods Inferences
Species Presence/absence All methods Species distribution,

occupancy richness
Relative Catch per unit Time- or Comparing
density effort (CPUE) area-constrained among populations
sampling, trapping, and species
litterbags
Density Individuals per Area- or Comparing among
unit area or volume-constrained populations and
volume sampling including species cohort
box and pipe survival
sampling
Population Total number of Extrapolation from area- Population dynamics
size individuals in or volume-constrained extinction risk
population samples, mark–recapture
4 Larval sampling | 57
Beyond estimating presence and absence, sampling can be directed at estimat-

ing density. Any technique which can be scaled by effort (sometimes expressed as
catch per unit effort, or CPUE) can offer estimates of relative density; that is, the
density of a given species relative to another. As one example, the number of larvae
from one species recovered during 30 person minutes of dip-netting may be com-
pared with a second species as an index of their relative density (both expressed
as individuals recovered per person minute). A smaller number of techniques,
identified below, allow direct estimates of the number of individuals per area, or
per volume. In general, these techniques operate on the principle that a defined
area or volume is sampled effectively. As one example, each use of a standard box
sampler covers 0.5 m2. After completing a set of 10 box samples within a wet-
land, the number of larvae of some species recovered from 5 m2 can be expressed
as the number of individuals per square meter. Finally, using mark–recapture
techniques, estimates of the total number of individuals within a larval cohort
may be made.
4.1.3 Timing
Amphibian larvae range from practically immobile, yolk-laden hatchlings, to
30-cm-long salamander larvae capable of moving extremely rapidly. Deciding
when to sample during the larval period is best done in conjunction with deci-
sions about how to sample. Small larvae move slowly and are often found in shal-
low areas. These attributes can make sampling relatively straightforward using
a variety of techniques. This is something to consider if your study aims allow
flexibility in the developmental stage sampled. It also suggests that effective lar-
val sampling depends on a close knowledge of the life histories of the species to
be sampled and their developmental progress within a given year. Late snowpack
melt, droughts, and a particularly cold or warm spring can all shift the timing of
larval development substantially. In tropical climates the time of year can be far
less important than the onset of wet season rains in triggering breeding and the
timing of larval sampling.
4.1.4 Sampling effort

How much is enough? In any sampling context, researchers are confronted with
the task of deciding how to allocate effort. As expected, a satisfactory answer
depends on study goal and tolerance of uncertainty. Any of the techniques
described below can be scaled in effort. The effort adopted, however, will require
direct estimates of how species detection, larval density estimates, or whatever
the target variable is, changes with increased effort. In our own research, we have
intentionally varied per wetland sampling effort (samples per visit, number of
visits, timing of visits) to determine the sensitivity of target responses (e.g. Werner
et al. 2007). There is no shortcut here; rules of thumb can be misleading resulting
in wasted effort and more collection than necessary or, more likely, incomplete
and inaccurate information (Skelly et al. 2003).
If you are new to a system, plan to learn during your initial sampling.
Intentionally trying multiple techniques and different degrees of sampling effort
will provide information that will ultimately save you time and produce more
reliable information. Before you step in the water, be prepared to spend more time
in each habitat if it is needed and to consider using multiple techniques to capture
the range of species and larval stages you intend to study.
4.2 Sampling techniques

4.2.1 Box/pipe sampler
4.2.1.1 Description
These area-based samplers are dropped rapidly through the water column in
areas of less than 1 m in depth. The earliest versions were sheet metal or ply-
wood boxes, 0.5 1.0 m, used in conjunction with a metal frame net designed
to fit snugly within the width of the box (Harris et al. 1988). Repeated sweeps
of the trapped volume of water are used to clear and count the amphibian lar-
vae within. A second form, the pipe sampler (Skelly 1996), ranges in size but is
typically constructed of an approximately 1 m length of aluminum pipe, 30 cm
or so diameter. Polyvinyl-coated aluminum pipe manufactured for air handling
in commercial heating and cooling applications is particularly effective. Pipe
samplers are used with dip nets constructed to measure half of the diameter of
the pipe. Entrapped larvae are cleared using repeated circular sweeps of the water
volume (Figure 4.1a).
Researchers have developed algorithms to ensure complete, or at least consist-
ent, clearance of animals. As one example, Werner et al. (2007) report sweeping
the water column of a pipe sampler a minimum of 10 times, and for at least 10 null
sweeps after the last collected animal has been removed from the net. For both
types of sampler, multiple samples are collected in a single wetland. Researchers
can use grids to lay out sampling points which can be distributed among habitat
types within a wetland. Alternatively, a minimum distance between sampling
points can be specified and placement of individual samples set haphazardly
within that constraint. If desired, data can be kept individually for each box or
pipe providing information on spatial distribution, variation in local density, pat-
terns of species coincidence, or association with particular microenvironments
(Freidenburg 2003).
(a)
(b)
(c)
(d)
Fig 4.1 Representative sampling techniques targeting larval amphibians in aquatic

habitats: (a) Pipe sampling in which a pipe is used to trap larvae that are then cleared
using a dip net (section 4.2.1); (b) leaf litterbags are used for sampling salamander
larvae in stream habitats (from Waldron et al. 2003; section 4.2.4); (c) a collapsible
funnel trap deployed along a pond edge (section 4.2.5); (d) a larval Ambystoma
talpoideum salamander marked with a lateral orange visible implant elastomer (VIE) tag
(photograph courtesy of Kristen Landolt of Murray State University; section 4.2.6).
4.2.1.2 Application
Box samplers are most effective within vernal ponds with an open water column and
simple bottom substrates such as decaying leaves. Pipe samplers were later developed
to capture the advantages of a box sampler while enabling the sampling of a wider
variety of environments including those where water is interspersed with emergent
vegetation (Skelly 1996). We are unaware of the use of area-based larval amphibian
samplers within stream environments, although the use of comparable samplers for
benthic macroinvertebrates suggests the potential for such an application.
4.2.1.3 Considerations
The major advantage of box and pipe samplers is that sampling of a known area
of wetland bottom provides direct estimates of larval density (individuals/m2). If
depth within each pipe is recorded, estimates of density per unit volume can be
determined. In either case, if wetland bottom area (or volume) is known, research-
ers have extrapolated pipe- and box-sample-based density estimates into estimates
of the size of entire larval cohorts (e.g. Werner et al. 2007).
A major disadvantage of these area-based samplers is their time-intensiveness.
Repeatedly placing and clearing a box or pipe is relatively slow and methodical
work even for experienced users. In addition, when sampling takes place in remote
areas that require hiking into and out of, box and pipe samplers can be inconveni-
ent because of their size.
4.2.2 Dip net

4.2.2.1 Description
Dip nets are probably the most common sampling tools used to collect amphibian
larvae. Dip-netting can be relatively unstructured if the goal is simply to capture
representatives of a particular species. Alternatively, dip-net surveys in which the
elapsed time (Werner et al. 2007) or number of sweeps (Gunzburger 2007) is
counted can be used to provide estimates of relative abundance of species. With
calibration from other methods such as pipe sampling, time-constrained dip-net
surveys have been used to provide per-unit-area density estimates (Werner et al.
2007). Effective dip-netting depends strongly on the species targeted and the
environment type sampled. However, most neophytes tend to collect a great deal
of substrate along with each sweep. With experience, users can learn how far into
the water column and substrate the net needs to pass to capture the larvae without
burying them in an overabundance of mud and vegetation.
Dip nets used in larval amphibian sampling range in size and shape from small
aquarium nets used in confined volumes of water (Thoms et al. 1997) to very large
nets approaching the size of some small seines (S. Cortwright, personal communi-
cation). Regardless of net size, it is common for amphibian biologists to construct
their own nets or to customize store-bought nets. Dip nets used for larval amphib-
ian sampling can have much shallower net bags and tend to require finer mesh
than those used for fish and other larger organisms. A shallow net bag facilitates
rapid processing of each sweep and can speed sampling significantly. Researchers
can also construct net dimensions that fit the structure of the environments they
sample (e.g. narrower nets may be appropriate for dipping out of marshes with
emergent vegetation, small stream pools, or tire ruts).
4.2.2.2 Application
Dip nets are used in most of the places where amphibian larvae are found. Nets
can be of lighter construction in vernal ponds (e.g. using mosquito mesh for the
net bag) compared with those used in streams and other places with abrasive
substrates.
Dip-netting is fast, requires a minimum of equipment, and can be performed in
a wide range of environments. Nets can be constructed inexpensively to perform
well in particular contexts. On the downside, as much as any technique, dip-
netting effectively relies on experience. An experienced user can vastly outper-
form a beginner working side by side. A second disadvantage relates to population
density estimates. While dip-netting can be used by itself to estimate relative
density in terms of catch per unit effort (where effort is often measured as time
spent sampling), estimates of area- or volume-based density must rely on other
techniques or through calibration from samples collected using other methods in
the same or comparable environments (Werner et al. 2007).
4.2.3 Seine
4.2.3.1 Description
Seines have long been used to collect amphibian larvae (Routman 1984). They are
particularly effective in sampling open, deeper areas that cannot easily be sampled
using box and pipe samplers or dip nets. Seining typically requires at least two
people. One person at each end sets the seine in a line and then begins moving in
an agreed direction until a position is reached where they move together and begin
gathering the ends of the seine up and out of the water, forcing the entrapped lar-
vae down into the middle. This is done most easily if the seine is being gathered
onto the shore. Sometimes this is not possible in which case, some sort of floating
platform such as a foam bucket float can be used. When the captured sample is
concentrated in the center of the seine, it is picked up and moved onto the shore
where the contents are sorted and the larvae are processed as desired. Most seines
used for amphibian larvae are relatively small, on the order of 3–5 m long and 1 m
or so deep (Shaffer et al. 1994). Larger seines often have a bag sewn in the middle.
The bag can greatly increase sampling effectiveness, but may also require a third
person to keep it from getting caught or rolled up in vegetation.
4.2.3.2 Application
Seines are typically used in open-water areas of ponds and lakes, although large
stream pools may also be sampled using seines in conditions where flow is not too
great.
Often seines are the only means of sampling larger amphibian larvae that live
in open-water regions of ponds (e.g. large Ambystomatid salamander larvae). A
major disadvantage of seines is that they are hard to handle in vegetation-choked
ponds and lakes often frequented by larval amphibians. In such an environment,
it can take over half an hour to sort through the vegetation and muck gathered
in a 5-min seine haul. During this sorting process, smaller amphibian larvae can
be hard to detect (and seine mesh is often coarse enough to enable their escape)
meaning that seines are most often used for large larvae. As with dip-netting,
seines are typically used to estimate catch per unit effort, although it is possible to
estimate number captured per unit area in some conditions (Shaffer et al. 1994).
4.2.4 Leaf litterbags

4.2.4.1 Description
Leaf litterbag sampling is a relatively new method for sampling stream habitats for
amphibians, particularly salamander larvae and adults (Pauley and Little 1998).
Litterbags create an artificial habitat refugium reproducing leaf packs commonly
found within streams, yet enclosed within mesh bags that can be easily removed
and sampled for salamanders. Leaf litterbags are constructed using plastic mesh
netting. Mesh gauge varies, but the 1.9 cm mesh (commonly found in deer-exclusion
netting and similar products) has proven useful. The mesh is cut into squares, with
50 50, 70 70, and 90 90 cm mesh sections representing small, medium,
and large bags, respectively. Several rocks are then placed on the netting to anchor
the bag, followed by leaves, needles, and other material likely to be found within
a particular stream environment. Once filled with litter, the corners of the mesh
are pulled together and cinched at the top using a cable tie (Figure 4.1b). Colored
flagging may be added to facilitate retrieval; this can be critical in contexts where
the bag may be displaced downstream by a high-flow event. Litterbags tend to

work best when deployed in locations where debris packs are likely to naturally
accumulate (e.g. pools, channel bends). In higher-flow environments, bags can be
secured by partially covering them with larger rocks or by tethering bags to roots
or pinning them using a stake driven into the substrate (Waldron et al. 2003;
Talley and Crisman 2007). Bags are sampled by placing a dip net underneath the
bag in the water column, quickly removing the bag from the water, and placing
it over a light-colored dishpan (Pauley and Little 1998; Jung et al. 2000). While
gently shaking the bag, salamander larvae and adults will often fall into the dish-
pan where species identification and enumeration can take place. The litter pile
should also be examined to ensure that all captured individuals are included. The
litterbag can then be reassembled and deployed again, although the leaf litter may
need to be replaced to compensate for decomposition over time.
4.2.4.2 Application
Litterbags have been used in stream environments where debris collects, and this
technique targets species known to utilize leaf packs. They are easy to deploy,
and quick and inexpensive to construct. They also manage to exploit an attribute
of species that makes them otherwise hard to sample. In the absence of litterbag
sampling, larvae of many stream-dwelling species are difficult to collect. They are
also able to capture more secretive or uncommon species that might be missed in a
dipnet survey. While litterbag sampling can produce species presence and relative
density data for a stream reach, it does not likely provide accurate estimates of abso-
lute population size or density, since it is unclear what exact area of the stream is
being sampled with each bag (Chalmers and Droege 2002; Waldron et al. 2003).
Litterbags are a highly specialized sampling tool. They are useful only for species
that dwell in streams and use leaf packs. But if such a species is being targeted, the
advantages of litterbags are substantial. Litterbag size is an important consider-
ation, as medium and large bags can capture more individuals and species; how-
ever, the size of the focal stream may only accommodate a smaller bag (Waldron
et al. 2003; Talley and Crisman 2007). Additionally, samples can be biased if
potential competitors or predatory individuals colonize the bag. Researchers
can discourage predatory (usually larger) adults and species by using finer mesh
or submerging bags in deeper water, leading to a preferential capture of larvae
(Waldron et al. 2003). Finally, the utility of litterbags may vary seasonally, as the
abundance of natural leaf pack habitats can vary throughout the year, depending
on the surrounding habitat cover.
4.2.5 Trapping
4.2.5.1 Description
Traps can be an effective means to capture amphibian larvae, requiring the
researcher to simply deploy the traps and check them after a time period sufficient
to have captured resident larvae. Most traps used by amphibian researchers are
of a funnel design, which channels larvae into a large holding section that can be
accessed by the researcher to recover captured animals. Commercially available
wire minnow traps have been used in many amphibian studies (e.g. Fronzuto and
Verrell 2000; Ghioca and Smith 2007). Home-made funnel traps using plastic
bottles (e.g. 2-L plastic drinks bottles) have also been used successfully (Calef
1973; Richter 1995). Collapsible traps made of fine nylon mesh and available
commercially (Promar, Gardena, CA, USA; Figure 4.1c), have capture rates equal
to or better than traditional wire minnow traps (Adams et al. 1997; C. Pearl,
personal communication). The finer mesh of these collapsible traps allows for the
retention of much smaller larvae, and the compact size and weight make them
suitable for backcountry work. Pyramid-shaped crayfish traps (Johnson and
Barichivich 2004) are an alternative to minnow and collapsible mesh traps that
can be particularly effective when it is important for part of the trap to extend
above the water surface. In all cases, traps are deployed by dropping each in a
predetermined area of the pond with a line attached and tied to a tree or float to
make locating and retrieving it easier.
4.2.5.2 Application
Traps can be effective in capturing amphibian larvae present within many aquatic
habitats with, perhaps, the exception of fast-moving water. Trapping is particu-
larly suited to detection of species presence, and to estimate catch per unit effort
(a metric of relative abundance). Trapping is sometimes the only suitable method
in deep water, steep-sided pond basins, or frozen ponds where wading in to con-
duct sampling is not feasible. Additionally, it may be easier to sample habitats
with structurally complex bottoms or vegetation-choked areas using traps, where
seining and dip-netting may be difficult. Lastly, funnel trapping will often cap-
ture rare, secretive, or more nocturnal species not detected using other methods
conducted in a short time period and during the daytime.
Whereas traps can be left overnight, this practice requires a sampling location to
be revisited within a short time period (usually 12–24 h) to avoid trap mortality,
especially of non-target species or life stages. It is especially important, when traps
are to be left for extended periods, to keep part of the trap above water to allow
access to water surface for trapped animals. Secondly, wire mesh size in commer-
cial minnow traps (around 6 6 mm) may be too large to effectively trap smaller
larvae, in which case sealing the trap with window screening composed of finer
mesh or the use of nylon collapsible traps can address this issue. Also, any trapping
methods and resulting data are based on an assumption of equal capture prob-
ability among individuals and populations. This can be violated if the presence
of conspecifics, competitors, or predators in the cage alters capture probabilities.
There could also be community-level biases if populations being sampled differ
in species composition. For instance, predators present in one habitat but not the
other may alter the behaviour of a target species and subsequent capture probabil-
ities. There can be other specific sampling biases in trapping certain species and
in some habitat types (e.g. playa wetlands; Ghioca and Smith 2007). Weather,
larval size and developmental stage, and resource availability can all affect capture
rates using traps (Adams et al. 1997). Some researchers use baits when trapping
amphibian larvae. However, for at least some species, it appears that baiting traps
does not increase trap effectiveness (Adams et al. 1997).
4.2.6 Mark–recapture
4.2.6.1 Description
Mark–recapture techniques are commonly used in studies of amphibian adult
populations, but can also be useful for the larval stage as well. However, rapid
growth often accompanied by a dramatic shift in body design can render marking
methods (often developed for juvenile and adults; see Chapter 8) unsuitable for
larval amphibians. Successful marking techniques for larvae include temporary
injectible organic dyes (Seale and Borass 1974) and externally staining dyes, which
typically stain amphibian larvae for less than 24 h (Pfennig 1999; Jung et al. 2002;
Harris et al. 2003), but can also slow growth rates (Travis 1981). More permanent,
yet onerous, marking techniques using paint sprays, stains, dimethyl sulfoxide,
and Super Glue (Ireland 1989) have largely been replaced by more convenient and
robust visible implant elastomer (VIE) tagging (Figure 4.1d). Passive integrated
transponder (PIT) tags may have limited use in all but the largest larval species due
to tag size (down to about 8 mm) and surgery required to implant, although the
development of smaller “injectable” PIT tags, applied using a hypodermic needle,
may expand the potential for this technique (Biomark, Boise, ID, USA).
VIE tagging appears to hold the most promise for amphibians, balancing the
ease of marking and longevity of the actual mark. Elastomer marks consist of a
silicone-based polymer material that is injected subcutaneously and cures into a
pliable and biologically inert solid (Northwest Marine Technology, Shaw Island,
WA, USA). Whereas some elastomer colors are visible to the naked eye when pre-
sent under translucent skin, fluorescent colors are often used and easily detected
using an ultraviolet light source (portable lights are available for field purposes).
Lowe (2003) used VIE to mark larvae of the spring salamander (Gyrinophilus
porphyriticus) and has indicated that marks can still be seen 10 years after mark-
ing (W. Lowe, personal communication). Fading of the elastomer does not
appear to be a common problem, although marks can migrate from the point
of injection or be lost altogether. Grant (2008) found that wood frog (Rana syl-
vatica) tadpoles can retain marks through metamorphosis and that larger marks
( 2 mm) were more likely to migrate in two larval stream salamander species.
Additionally, it was indicated that stream salamanders could be marked without
anesthesia, while wood frog tadpoles required anesthesia and also had poorer
mark retention (E. Grant, personal communication).
4.2.6.2 Application
Mark–recapture studies can be conducted in just about any habitat type, assuming
that the same method of capture is used for each sampling period. Assuming that
a sufficient proportion of the population is originally marked, mark–recapture
techniques can provide robust estimates of absolute population size, especially
when combined with robust capture–recapture estimation models (Chapter 24).
Jung et al. (2002) found that mark–recapture methods provided the most accurate
estimates of population size for two species of tadpoles in desert pool habitats.
Regardless of technique, small amphibian larvae are relatively difficult to mark.
The VIE technique will be more useful for species with larger larvae, or at least
with individuals farther along in development. Gyrinophilus porphyriticus larvae
as small as 2 cm in total length have been marked successfully, as well as R. syl-
vatica individuals down to 2.5 cm in total length. Additionally, any substantial
loss of tags within a marked cohort can seriously bias population size estimates.
Consider this when deciding which technique to use and for what exact purpose.
4.3 Other techniques

A number of additional techniques have been used in sampling amphibian lar-
vae. Because they have limited application or because they have been used rela-
tively infrequently, we mention them only briefly here and point readers to sources
where more detailed descriptions are available.
4.3.1 Bottom net

A bottom net is placed on the bottom in a set position such that, when triggered,
floats on an upper frame carry the net to the surface, entrapping larvae in open-
water areas of small ponds (Shaffer et al. 1994). This area-based sampling tech-
nique has been used relatively little for amphibians, but could be effective in water
where other techniques are not feasible.
4.3.2 Electroshocking
Electroshocking, developed initially to sample fish, is used primarily within
streams and rivers. However, it can stun and facilitate capture of amphibian lar-
vae as well (Brown and May 2007). It appears to be most effective for lentic species
found in slow-moving parts of rivers (Shaffer et al. 1994). Many stream-dwelling
amphibian species use retreats or bury themselves in the substrate in ways that
prevent their detection and capture even if stunned during electroshocking.
4.3.3 Visual encounter survey

Visual encounter surveys, discussed more thoroughly in Chapter 15, may also be
used, with some considerations, to sample larvae. Visual encounter surveys alone
will likely be insufficient to detect a significant proportion of larvae present in dark
or turbid water (such as a tannin-rich vernal pool, an algae-/vegetation-choked
pond, or a stream with turbulent waters). Larval behavior and microhabitat pref-
erence (e.g. aggregation, crypsis, water-column basking, hiding in the substrate,
preference for deep water or thick vegetation) can lead to species-specific and
highly variable rates of detection. However, where water conditions allow, and the
potential for confusing species is low, visual encounter surveys can be an efficient
technique.
4.4 Conclusions
As in many aspects of field biology, the techniques used for sampling amphib-
ian larvae are often passed without criticism or comment from one generation
of researchers to the next. In many cases, there has been little effort to ask why
one technique should be used as opposed to its alternatives or how a given
technique may be most effectively applied. The many techniques outlined in
this chapter are connected through the references listed below to an enormous
cumulative effort to understand the most effective and efficient means to esti-
mate the presence and density of larvae. Most of the techniques require little
equipment, and that equipment is typically relatively inexpensive. Neither are
the techniques difficult to master. Collectively, this means that there is little
reason not to try multiple techniques and to calibrate and understand the con-
sequences of altering the timing and intensity of sampling. The modest effort
to do so will greatly increase the reliability of the information gathered and, in
all probability, lead to unforeseen insights into the biology of the species being
studied.
4.5 Acknowledgments
We thank S. Cortwright, M. McPeek, E. Werner, and H. Wilbur for teaching us
about larval amphibian sampling. M. Adams, E. Grant, B. Hossack, W. Lowe,
and C. Pearl provided helpful insight and details into the techniques they use.
4.6 References
Adams, M.J., Richter, K.O., and Leonard, W.P. (1997). Surveying and monitoring
amphibians using aquatic funnel traps. In D.H. Olson, W.P. Leonard, and R. Bury
(eds), Sampling Amphibians in Lentic Habitats: Methods and Approaches for the Pacific
Northwest: Northwest Fauna Number 4, pp. 47–54. Society for Northwestern Vertebrate
Biology, Olympia, WA.
Brown, L.R. and May, J.T. (2007). Aquatic vertebrate assemblages of the Upper Clear
Creek Watershed, California. Western North American Naturalist, 67, 439–51.
Calef, G.W. (1973). Natural mortality of tadpoles in a population of Rana aurora. Ecology,
54, 741–58.
Chalmers, R.J. and Droege. S. (2002). Leaf litter bags as an index to populations of north-
ern two-lined salamanders (Eurycea bislineata). Wildlife Society Bulletin, 30, 71–4.
Duellman, W.E. and Trueb, L. (1986). Biology of Amphibians. John Hopkins University
Press, Baltimore, MD.
Freidenburg, L.K. (2003). Spatial Ecology of the Wood Frog (Rana sylvatica). PhD
Dissertation, University of Connecticut, Storrs, CT.
Fronzuto, J. and Verrell, P. (2000) Sampling aquatic salamanders: tests of the efficiency
of two funnel traps. Journal of Herpetology, 34, 146–7.
Ghioca, D.M. and Smith, L.M. (2007). Biases in trapping larval amphibians in playa
wetlands. Journal of Wildlife Management, 71, 991–5.
Grant, E.H.C. (2008). Visual implant elastomer mark retention through metamorphosis
in amphibian larvae. Journal of Wildlife Management, 72, 1247–52.
Gunzburger, M.S. (2007). Evaluation of seven aquatic sampling methods for amphibians
and other aquatic fauna. Applied Herpetology, 4, 47–63.
Harris, R.N, Alford, R.A., and Wilbur, H.M. (1988). Density and phenology of
Notophthalmus viridescens dorsalis in a natural pond. Herpetologica, 44, 234–42.
Harris, R.N., Vess, T.J., Hammond, J.I., and Lindermuth, C.J. (2003). Context-
dependent kin discrimination in larval four-toed salamanders Hemidactylium scutatum
(Caudata: Plethodontidae). Herpetologica, 59, 164–77.
Ireland, P.H. (1989). Larval survivorship in two populations of Ambystoma maculatum.

Journal of Herpetology, 23, 209–15.
Johnson, S.A. and Barichivich, W.J. (2004). A simple technique for trapping Siren lac-
ertina, Amphiuma means, and other aquatic vertebrates. Journal of Freshwater Ecology,
19, 263–9.
Jung, R.E., Droege, S., Sauer, J.R., and Landy R.B. (2000). Evaluation of terrestrial
and streamside salamander monitoring techniques at Shenandoah National Park.
Environmental Monitoring and Assessment, 63, 65–79.
Jung, R.E., Dayton, G.H., Williamson, S.J., Sauer, J.R., and Droege, S. (2002). An
evaluation of population index and estimation techniques for tadpoles in desert pools.
Lowe, W.H. (2003). Linking dispersal to local population dynamics: a case study using a
headwater salamander system. Ecology, 84, 2145–54.
Olson, D.H., Leonard, W.P., and Bury, R.B. (eds) (1997). Sampling Amphibians in Lentic
Habitats: Methods and Approaches for the Pacific Northwest: Northwest Fauna Number 4.
Society for Northwestern Vertebrate Biology, Olympia, WA.
Pauley, T.K. and Little, M. (1998). A new technique to monitor larval and juvenile sala-
manders in stream habitats. Banisteria,12, 32–6.
Pfennig, D.W. (1999). Cannibalistic tadpoles that pose the greatest threat to kin are most
likely to discriminate kin. Proceedings of the Royal Society of London Series B Biological
Sciences 266, 57–61.
Richter, K.O. (1995). A simple aquatic funnel trap and its application to wetland amphib-
ian monitoring. Herpetological Review, 26, 90–1.
Routman, E.J. (1984). A modified seining technique for single person sampling of deep or
cold water. Herpetological Review, 15, 72–3.
Seale, D. and Borass, M. (1974). A permanent mark for amphibian larvae. Herpetologica,
30, 160–2.
Shaffer, H.B., Alford, R.A., Woodward, B.D., Richards, S.J., Altig, R.G., and Gascon, C.
(1994). Quantitative sampling of amphibian larvae. In W.R. Heyer, M.A. Donnelly,
R.W. McDiarmid, L.C. Hayek, and M.S. Foster (eds), Measuring and Monitoring
Biological Diversity: Standard Methods for Amphibians, pp. 130–41. Smithsonian
Institution Press, Washington DC.
Skelly, D.K. (1996). Pond drying, predators, and the distribution of Pseudacris tadpoles.
Copeia, 1996, 599–605.
Skelly, D.K., Yurewicz, K.L., Werner, E.E., and Relyea, R.A. (2003). Estimating decline
and distributional change in amphibians. Conservation Biology, 17, 744–51.
Talley, B.L. and Crisman, T.L. (2007). Leaf litterbag sampling for larval plethodontid
salamander populations in Georgia. Environmental Monitoring and Assessment, 132,
509–15.
Thoms, C., Corkran, C.C., and Olson, D.H. (1997). Basic amphibian survey for inven-
tory and monitoring in lentic habitats. In D.H. Olson, W.P. Leonard, and R. Bury
(eds), Sampling Amphibians in Lentic Habitats: Methods and Approaches for the Pacific
Northwest: Northwest Fauna Number 4, pp. 35–46. Society for Northwestern Vertebrate
Biology, Olympia, WA.
Travis, J. (1981). The effect of staining on the growth of Hyla gratiosa tadpoles. Copeia,
1981, 193–6.
Waldron, J.L., Dodd, Jr., C.K., and Corser, J.D. (2003). Leaf litterbags: factors affecting
capture of stream-dwelling salamanders. Applied Herpetology, 1, 23–36.
Werner, E.E., Skelly, D.K., Relyea, R.A., and Yurewicz, K.L. (2007). Amphibian species
richness across environmental gradients. Oikos, 116, 1697–1721.
5
Dietary assessments of larval amphibians
Matt R. Whiles and Ronald Altig
5.1 Background
Diet analyses of larval amphibians provide information on foraging patterns,
nutritional requirements, and trophic interactions in aquatic food webs, infor-
mation that is critical for successful conservation and management. Numerous
amphibian species around the globe are declining or presumed extinct (Stuart
et al. 2004). Thus, there is an urgent need for detailed information on larval
diets so that we can understand whether food-related factors are limiting wild
populations and facilitate successful rearing of species in captivity.
Some species of all three orders of living amphibians have free-living, feeding,
larval forms. Larval caecilians and salamanders, both of which occur in a variety
of lentic and lotic habitats, are all carnivores that eat small aquatic organisms
by suction-feeding, and their diets and trophic status are thus relatively easily
assessed. Anuran larvae (tadpoles) occur in almost every conceivable type of
freshwater habitat and show great diversity in feeding modes. As such, morpho-
logical diversity of the oral structures, which glean materials from substrata,
and the buccopharyngeal apparatus, which selectively captures food particles,
of tadpoles is large.
Whereas a few tadpoles are macrocarnivores (e.g. Lepidobatrachus,
Leptodactylidae), most either rasp materials associated with substrata such as
biofilms, periphyton, and deposited organic particles (e.g. ranids and many
hylids), harvest naturally suspended particles in the water column (e.g. micro-
hylids, pipids, and rhinophrynids), capture particles in the surface film (i.e.
neustonic tadpoles with umbelliform oral structures), or consume conspecific
or heterospecific eggs. Most of these feeding modes need further study (Altig
et al. 2007). Although many tadpoles are classified as general herbivores or
detritivores, they likely obtain a fair amount of energy and nutrients from more
easily assimilated heterotrophic components of their diets (e.g. animal parts

and microbes; Altig et al. 2007).
Herein, we focus on descriptive and observational methods for assessing diet
and trophic status of larval amphibians. However, other approaches that are
beyond the scope of this review can also provide valuable insight into the trophic
ecology of larval amphibians. In particular, manipulative approaches such as
cafeteria experiments can be used to examine amphibian food preferences (e.g.
Kupferberg 1997) and field enclosure/exclosure experiments can illuminate
functional roles and trophic interactions (e.g. Solomon et al. 2004). Combined
approaches, such as experimental manipulations performed in conjunction with
gut content analyses (e.g. Ranvestel et al. 2004; Whiles et al. 2006) have been
used to examine amphibian ecology at multiple scales, from individual foraging
to roles of amphibians in ecosystem processes. Manipulative approaches for
studying various aspects of amphibian ecology are discussed in Chapters 6 and
12 in this volume.
5.2 Larval caecilians and salamanders

With some modifications (e.g. different foraging strategies; benthic versus nek-
tonic and day versus night feeding), general habitats (e.g. lentic versus lotic),
and morphologies (e.g. gape size and size and number of gill rakers), the method
of food capture is essentially the same in all larval caecilians and salamanders.
These groups form a negative pressure by rapidly depressing the throat as they
open the mouth so as to suck food items into the buccal cavity. Plant materials
found in the guts of these groups are generally considered as incidental. As such,
gut content analyses can be effective for assessing diet in these groups, although
robust sampling strategies that account for spatiotemporal foraging patterns
and feeding behaviors should be used.
5.3 Anuran tadpoles

Most tadpoles appear to be omnivorous, although this contention is debated
and the true trophic status of many tadpoles has not been accurately assessed
(Altig et al. 2007). When tadpoles consume a range of food types that vary
in digestibility and nutritional quality, simply quantifying the constituents of
their diet may provide limited information on the relative importance of spe-
cific foods to growth and development. Accurate assessments of the trophic sta-
tus of omnivorous tadpoles require information on assimilation which can be
obtained through isotopic analyses, fatty acid analyses, or by applying ecological
5 Dietary assessments of larvae | 73
efficiencies (e.g. assimilation and/or production efficiencies) to gut content data.

Consummatory diet information can be useful for studying feeding behaviors
and the types of materials that the oral apparatus and buccopharyngeal struc-
tures can harvest and capture.
5.4 Assessing food sources and diets

5.4.1 Category I: preparatory studies
Along with analyses of larval amphibians and/or their gut contents, many stud-
ies also include some assessment of food availability. In most freshwater habitats,
food resources are dynamic. For example, algal blooms can occur over short
time periods, and thus frequent sampling of potential food items is necessary.
Coupled with studies of foraging behavior, information on the seasonal phen-
ology of resources can guide the timing of sampling and aid in data interpret-
ation. On an even shorter diurnal time scale, samples of larvae and their prey
collected during daylight may not accurately reflect trophic interactions of sala-
manders active in the dark.
Sampling protocols, including number of samples collected at a given time,
apportioning samples among available micro- or mesohabitats (e.g. pools versus
riffles in a stream, macrophyte beds versus non-vegetated areas of a pond), will
vary with the specific system examined and associated questions. Preliminary
sampling can be used to obtain estimates of variability for a priori power ana-
lyses to assist with decisions regarding sampling intensities (e.g. Steidl et al.
1997).
Diets of many larvae change through development, and thus size and stage
should be accounted for when sampling. Larval size is important, but does not
always reflect age. For this reason, a staging table based on the sequence of acqui-
sition of discrete morphological markers should be used. Dünker and Wake
(2000) provide a staging table for caecilians, and those for salamander larvae
and tadpoles are summarized in Duellman and Trueb (1986).
Collection and preservation techniques for amphibians are often taxonomic-
ally specific (see Heyer et al. 1994). For collection of potential food items, Hauer
and Lamberti (2006) provide a wealth of information on proper collection tech-
niques of materials ranging from detritus to algae. Smith (2001), Thorpe and
Covich (2001), and Merritt et al. (2008) provide methods and taxonomic keys
for invertebrate prey. Prey items should be identified to the lowest taxonomic
level possible. For both amphibian larvae and prey, specimens should be placed
in a readily accessible, cataloged collection. This allows others to verify identifi-
cations and use materials for future studies.
5.4.2 Category II: gut contents

Most studies of amphibian diets have relied on examinations of gut contents.
However, gut-content data should be interpreted cautiously because the types
and amounts of materials present are a function of the time and place of collec-
tion (time of day, season, age of the individual, etc.) and euthanization and pres-
ervation methods. While gut contents acquired with robust sampling designs
that account for spatial and temporal variations can be reliable indicators of
foraging patterns and functional roles of larval amphibians, they may not be
good indicators of true trophic status because of differential assimilation of
materials (Altig et al. 2007). As noted above, we assert that larval caecilians and
salamanders are the only groups for which true trophic position can be assessed
via examination of gut contents alone. All methods have limitations, so the best
approach is to assess diets and trophic relations using multiple approaches (e.g.
examination of stomach contents combined with stable-isotopic analyses of tis-
sues). Further, approaches should be iterative to account for time lags between
ingestion of available foods and actual changes in isotopic signatures or fatty
acid profiles, which are a function of tissue turnover rates.
Gut contents can be removed from larger larval salamanders and caecilians
without killing the individual. Although gut contents may be removed with for-
ceps from large individuals that are anesthetized, the safest and most effective
techniques involve forced regurgitation. This is generally accomplished with
stomach flushing or gastric lavage, a technique that involves displacing materials
from the stomach and back out the mouth with a low-pressure stream of water
(Solé et al. 2005). Approaches vary and should be adjusted for an individual’s
size and mouth and gut morphologies, but generally involve gently forcing water
from a syringe through a lubricated (water or a water-based lubricant) catheter
tube that is inserted into the esophagus. During this procedure, the animal is
held with the head facing down and over a vessel where displaced materials and
fluid are collected. Contents can then be rinsed through an appropriately sized
sieve and preserved in ethanol or buffered formalin. Excess water should be
removed before preservation. If performed correctly, this technique is harmless
and very effective because fresh materials are collected from the stomach.
5.4.3 Procedures: anurans and small predators

For other amphibians (e.g. tadpoles, small salamanders), specimens are generally
sacrificed, in which case individuals should be collected as gently as possible, anes-
thetized with MS-222®, Orajel®, or similar materials, and quickly euthanized in
a humane manner to avoid vomiting and/or continued digestion of food. The
gut and its contents can be removed in the field or later from properly preserved
specimens. In cases where fatty acid or isotope analyses are also planned, speci-
mens may need to be frozen rather than placed in fluid preservatives. Freezing
leads to distortion of some materials, and removal of guts and processing or pres-
ervation should take place immediately after specimens are thawed.
5.4.4 Processing and analysis of gut contents

Processing methods for gut contents must account for differences in gut tracts
among taxa. Most tadpoles have long, coiled guts that are typically several times
their body length, whereas salamanders and caecilians have shorter guts with a
well-developed stomach that is typical of carnivores. These differences warrant
different approaches.
For salamanders and caecilians, prey items are removed from the gut, identi-
fied to the desired taxonomic level, and counted and measured. In cases where
guts contain many small prey items any number of subsampling techniques can
be used, so long as they are unbiased toward different types of prey. A simple
technique involves spreading the gut contents evenly in a Petri dish with num-
bered grids and then using a table of random numbers to select grids in which
contents are identified and measured until a representative subsample has been
examined. Protocols for degree of subsampling range from fixed counts (e.g.
100 individuals) to proportion of total examined (e.g. one-eighth of the total
grids or area examined). Methods should be calibrated by progressively exam-
ining larger portions of a subset of samples to assess the accuracy of various
subsample sizes. For most purposes, expressing prey in terms of mass or volume
is best, although simple numerical analyses are sufficient for some assessments
of foraging behaviors and prey availability. Mass analyses are more appropriate
for assessments of trophic status and nutrition because they can be converted
to organic content or caloric equivalents. Mass is generally estimated by actual
weighing (generally as dry mass (DM) or ash-free dry mass (AFDM), which
is a proxy for organic content; see methods below) or use of length/mass rela-
tionships, which are generally in the form: M aLb where M is the mass of the
organism, L is a linear dimension such as length, and a and b are regression con-
stants (reviewed by Benke et al. 1999). Length/mass relationships for a variety
of potential prey items can also be found in Rogers et al. (1976), Bottrell et al.
(1976), Dumont et al. (1975), and Culver et al. (1985). In many cases, prey items
are fragmented, and thus care must be taken to reassemble them or estimate
total lengths. Body length and head width, which are the most common meas-
ures used, are generally performed with a dissecting microscope equipped with
an ocular micrometer, graduated stage, or image-analysis capabilities. Mass is
usually expressed as DM or AFDM. For more in-depth analyses, caloric content
is very informative. Estimates of caloric contents for many types of prey are
available (e.g. Cummins and Wuychek 1971; Rodgers and Qadri 1977), but
these estimates tend to be quite variable. Depending on the level of precision
desired, analyses such as bomb calorimetry should be performed on actual prey
items or other materials from the guts or specimens collected from the same
habitats at the same time. As with DM and AFDM, size/caloric-content rela-
tionships can be developed.
While collection and identification of gut contents from salamanders and
caecilians is relatively straightforward, tadpoles present a challenge because
their long guts contain mixtures of materials of different digestibility that are at
different stages of digestion. In this case, the most recently ingested materials
from the foregut should be the foci of analyses. This is generally accomplished
by removing a section of the foregut that is a consistent proportion of the total
gut length (e.g. the anterior quarter of the gut) and preparing it for examination
under a microscope (for identification of individual food particles) or for ana-
lyses of caloric content, fatty acid profiles, or isotopic composition (see methods
for fatty acids and isotopes below). Obviously, method of collection and preser-
vation can limit the types of analyses that can be performed on gut contents.
For identification of foregut contents, materials are generally rinsed from
the gut, slide-mounted, and examined under a microscope. One rather simple
and effective method is to place materials in glycerin and gently homogenize
them. If needed, water can be added and the materials sonicated to further
break up contents. Depending on the abundance of particles, either the entire
sample or a subsample is filtered with low vacuum on to a 0.45 μm gridded
membrane filter to obtain a reasonable dispersion of particles across the filter
(e.g. 10–20 particles/grid). Filters should be allowed to dry at ambient tem-
perature for 2–3 h to remove excess water, and then two ot three drops of Type
A immersion oil, which clears the filter, are added. After approximately 24 h
the filters will clear and can then be placed on a slide and covered with a cover-
slip, which is sealed with enamel nail polish. Slide-mounted materials are gen-
erally subsampled (e.g. some fi xed number of fields of view examined, a fi xed
number of grids examined, or all particles along transects are identified and
measured). Measurements of individual particles can be made using methods
ranging from an ocular micrometer to sophisticated image analysis systems.
Data on identified food types are generally expressed as area or converted to
mass, or, in the case of algae, biovolume (Lowe and LaLiberte 2006; Steinman
et al. 2006). There are many variations for mounting and analyzing gut con-
tents (for recent examples using various taxonomic groups, see Evans-White et al.
2003; Ranvestel et al. 2004; Rosi-Marshall and Wallace 2002). Regardless of
modifications, the key points are to avoid damaging materials beyond recogni-
tion and to obtain an evenly dispersed, representative sample.
Masses of materials in guts can be estimated with methods ranging from
simply weighing materials to using size/mass relationships (e.g. body length/
mass relationships for prey; see above section on salamander and caecilian guts)
either developed by the investigator or gleaned from the literature. For direct
weighing, DM or AFDM estimates are best. Dry mass estimates are generally
developed by drying materials at 55–60°C to constant mass. For AFDM, dried
materials are weighed, then ashed in a furnace (≈500°C) for 1–2 h, and weighed
again; mass of remaining ash is subtracted from the original dry mass to obtain
AFDM. In either case, samples should be allowed to cool in a dessicator before
weighing so that moisture is not absorbed in the cooling process. Actual drying
and ashing times will vary with sample size. Small amounts of materials can
be processed in this manner on small pieces of heavy duty aluminum foil or
glass fiber filters. Development of wet mass/DM or AFDM relationships can
streamline procedures and decrease the number of samples that are destroyed.
Likewise, if caloric equivalents are derived through calorimetry, similar rela-
tionships can be developed.
Gut content data expressed as mass or area (or in some cases biovolume for
algae (see Lowe and LaLiberte 2006) can be combined with estimates of assimi-
lation efficiencies for more accurate assessments of trophic status. Procedures
vary, but a simple approach is to express food types as percent of total gut con-
tents and multiply each by its respective assimilation efficiency. Resultant values
for the food types are then summed and the percentage contribution of each to
assimilated materials can be calculated. If caloric content is estimated, indices of
prey or food-type importance can be calculated using basically the same basic
procedure; masses of each food type in the gut are multiplied by their corre-
sponding caloric density and percentage caloric contribution of each food type is
the total number of calories attributable to a given food type divided by the total
caloric content in the gut. Regester et al. (2008) provide examples of these types
of analyses and related procedures applied to pond-breeding ambystomatid com-
munities. Examples using various freshwater omnivores can be found in Bowen
et al. (1995) and Evans-White et al. (2003). Regester et al. (2008) and Skelly and
Golon (2003) provide examples of methods for estimating assimilation efficien-
cies of amphibian larvae.
In cases where algae are obviously consumed, analyses of photosynthetic pig-
ments may be performed on gut contents to estimate algal biomass. For this pro-
cedure, specimens and/or their gut contents must be frozen immediately upon
collection and kept dark and frozen prior to analyses. Chlorophyll a, which can
be used to estimate algal biomass, can be extracted from samples of gut contents
with an organic solvent and concentration is then estimated with spectropho-
tometry, fluorometry, or high-performance liquid chromatography. Steinman
et al. (2006) provide a review of methods to estimate algal biomass via photo-
synthetic pigment analyses. Although informative to a degree, pigment analyses
alone cannot account for the digestibility of algal materials, and some tadpoles
will re-ingest feces containing undigested algal materials. The ability of tad-
poles to digest algal materials can be assessed through microscopic examination
of cells in feces (e.g. whether full chloroplasts are present) or culturing algae
from feces can identify the quantity and types of materials that are still viable
after passing through guts (Peterson and Boulton 1999). Simple counts of prey
items or food types in guts may not provide reliable information on trophic sta-
tus, but can be valuable for assessing feeding preferences and interactions with
other consumers. A variety of indices, including Horn’s index and Morisita’s
index, can be applied to gut-content data to quantify diet overlap among species
(see Brower et al. 1998; Krebs 2001; Chapter 18). Likewise, if data on the abun-
dance of prey and/or food type in the environment are also available, feeding
selectivity can be assessed and quantified using a variety of available indices.
Relatively simple and commonly used indices include Ivlev’s index of electivity
(Ivlev 1961) and the selectivity index (Chesson 1983).
5.5 Category III: assimilatory diet

Analyses of isotopic composition and fatty acid profiles can provide reliable
information on diets and trophic status because they reflect assimilation over
long periods of time. However, these approaches can be more costly and compli-
cated than simple gut analyses because they require expensive analytical equip-
ment, and they also have their own limitations, some of which are discussed
below. In most cases, samples are simply collected and preserved in a manner
that does not alter isotopic or fatty acid composition, depending on the analyses
to be performed, and then sent to an analytical laboratory for analyses.
5.5.1 Stable-isotope analysis

Isotopes of carbon (13C) and nitrogen (15N) are commonly used in studies of
freshwater food webs because they allow for tracking transfers of organic car-
bon and nitrogen from living plant and detrital sources to primary consumers
and predators (Peterson and Fry 1987; Fry 2006). Other isotopes, including
deuterium (2H or D) and 18O, are gaining increasing attention from ecologists
for tracing movements of organisms. We focus on carbon and nitrogen, as they
can provide valuable information on assimilatory diets and methods for their
analyses are standardized. Stable-isotope analyses focus on ratios of heavier
(e.g. 15N and 14C) and lighter isotopes (14N and 13C). Ratios are expressed as ,
which is parts per thousand (‰) deviation from a standard, calculated as 13C
or 15N [(Rsample Rstandard)/Rstandard] 1000 where R 13C/12C or 15N/14N.
Common standards are Pee Dee belemnite limestone for C and atmospheric N
(‰ value set to 0). A positive (N) or less negative (C) value indicates the sample
is enriched with the heavy isotope.
Carbon isotope signatures of consumers are generally similar to those of their
food source or sources (i.e. you are what you eat), which can have 13C val-
ues ranging from slightly higher than atmospheric CO2 (−8‰) to −40‰ for a
variety of reasons that influence degree of fractionation (see Peterson and Fry
1987; Fry 2006; Hershey et al. 2006). In particular, in streams, water velocity
can influence carbon isotopic compositions of foods such as algae (Finlay et al.
1999). Differences in C3 and C4 photosynthetic pathways result in differences
among some plants and their detritus. Likewise, there are often measurable dif-
ferences in carbon isotope ratios between terrestrial and aquatic autotrophs and
their detritus.
Nitrogen is more often used to assess trophic position, as 15N shows a fairly
consistent change with each trophic step (≈3.4‰ increase in 15N with each
trophic step up from primary consumer to predator; Vander Zanden and
Rasmussen 1999; Post 2002). This fairly predictable fractionation results
from the tendency for organisms to excrete more of the lighter nitrogen isotope
than the heavy one. However, fractionation can vary and there is evidence that
the increase in 15N with each trophic step is considerably less than 3.4‰ in
some tropical stream food webs (Kilham and Pringle 2000). In general, basal
resources such as periphyton and detritus in freshwater habitats will have 15N
values ranging from near 0 to 3‰, primary consumers range from 3 to 6‰,
and predators from 6 to 12‰, but values can vary considerably for a variety
of reasons (Fry 2006). Whiles et al. (2006) and Verburg et al. (2007) provide
examples of isotopic analyses of stream food webs with abundant and diverse
tadpole assemblages.
Most studies examine naturally occurring isotopic ratios (natural abundance
studies) in consumer tissues and food resources, but additions of enriched
materials (e.g. tracer studies using 15N enriched ammonium or nitrate or 13C
enriched acetate) are also used, particularly when natural isotopic ratios among
food types are similar. Tracer studies can be informative for assessing diets and
trophic interactions, as well as examining roles of consumers in biogeochem-
ical cycling, because materials (e.g. N or C) are followed from basal resources
through top consumers. Stable-isotope tracer additions have been used to exam-
ine carbon cycling through consumers in streams (Hall 1995), nitrogen cycling
in stream food webs (Hall et al. 1998), and movement of nutrients from streams
to riparian habitats via consumers (Sanzone et al. 2003). Tracer studies could
be particularly informative for assessing roles of amphibians in freshwater eco-
system processes, as well as material and energy exchanges between aquatic and
terrestrial systems.
5.5.2 Procedures
Stable-isotope analyses can be performed on entire organisms or, in the case of
larger individuals, muscle tissue (generally 1–2 g wet weight). While muscle tissue
is the usual focus for larger animals, blood and other tissues such as liver or skin
are sometimes examined because they have different incorporation and turn-
over rates, and can thus be used to assess temporal patterns of food availability
and assimilation (Dalerum and Angerbjorn 2005). For example, while isotopic
analyses of muscle tissues may reflect assimilation over a period of months, liver
or blood plasma, which turnover much more rapidly, can reflect a period of days
or weeks. This isotopic memory is an important consideration, as diets of some
amphibians can change considerably through development and with changing
resource availability.
Muscle tissues from large individuals can be sampled non-lethally using ster-
ile biopsy punches, which are available from many surgical supply companies. If
biopsies are collected, wounds should be treated with an appropriate antibiotic
before releasing the specimen. Live amphibians and prey items that are to be proc-
essed in their entirety should be allowed to clear their guts as much as possible
by placing them in filtered water from the same habitat for approximately 12 h.
Alternatively, gut tracts can be removed before processing, but this can be a tedi-
ous process for small specimens. The best methods for preservation of samples for
isotopic analyses are freezing or drying at low temperature (55–60°C to constant
mass). In the field, samples should be placed on ice or in liquid nitrogen and then
transferred to a freezer (−80°C) or drying oven when available. Dried samples
should be stored in a desiccant such as drierite or silica. Cross-contamination
must be avoided throughout sampling and processing of all materials.
Comprehensive analyses of food webs require sampling of all potential food
items that are available for larval amphibians or their prey. Sampling of prey,
algae, biofilms, detritus, and other potential resources simply requires collection
of representative samples. These samples can be collected using standard proced-
ures (e.g. Hauer and Lamberti 2006), but care must be taken to avoid cross-
contamination. Sampling of fine particles from substrates or the water column
can be collected and processed on glass-fiber filters. As with any other sampling,
replicates should be collected to account for natural variability. In flowing waters,
areas of different velocities should be sampled, as current velocity can alter the
carbon isotope compositions of periphyton (Finlay et al. 1999).
Prior to isotopic analyses, all samples are dried to constant weight at 55–60°C,
ground into a fine powder, weighed, and packed into tin capsules. Samples of
particulates on glass-fiber filter can either be carefully scraped from the filters,
or a portion of the filter, with sampled material on it, can be packed into the
tin capsule. While muscle tissue is the usual focus for larger animals, for most
prey items (e.g. aquatic invertebrates) the entire organism is processed, or, in
the case of very small prey, individuals are combined. Samples for N isotope
analysis should contain at least approximately 100 μg of DM of N; samples
for C isotope analysis should contain approximately 2 mg of DM of C. Sample
sizes for analysis of both C and N isotopes vary with C:N ratio of the sample,
with larger sample sizes needed (e.g. 10–60 mg) for materials with high C:N
ratio (e.g. sediments with low %N). Samples of basal resources and sediments
from calcareous regions should be processed to remove carbonates before ana-
lysis. Carbonates can easily be removed by acidification via exposure to HCl
vapors (Yamamuro and Kayanne 1995). Isotopic analyses are performed using
a continuous-flow isotope-ratio mass spectrometer.
Analyses and interpretation of stable-isotope data range from simply com-
paring values of resources and consumers to assess relative contributions, to
quantifying trophic status (Vander Zanden et al. 2000; Post 2002) or niche
breadth (Layman et al. 2007). Depending on the nature of the system, stable-
isotope analyses will sometimes reveal clear relationships among consumers
and resources. In other cases, environmental complexities and high degrees
of omnivory can confound results. In cases where multiple food sources are
likely contributing to the isotopic composition of a consumer, as would be the
case with most tadpoles, mixing models can be used to assess relative contribu-
tions of different foods. Simple two-source mixing models can be used to esti-
mate relative contributions of two resources (e.g. periphyton and detritus; see
Hershey et al. 2006), whereas more complex models are necessary when more
than two food sources are likely.
5.5.3 Fatty acid analyses

Fatty acids are the essential foundations of energy storage lipids in consumers,
but they are not readily synthesized by animals. As such, they must be obtained
through their food. Differences in fatty acid content among food types can
contribute to many of the qualitative differences among larval food resources in
terms of their importance for growth and development. Given their nutritional
value, that they are conserved, and that various resources and consumers have
unique combinations of fatty acids, fatty acid profiles can be good indicators
of diet and trophic status. Analyses of fatty acid profiles have been successfully
used to assess food web structure in marine, lake, and stream systems (Muller-
Navarra et al. 2000; Stübing et al. 2003; Sushchik et al. 2003; as well as aquatic–
terrestrial food web linkages, Koussoroplis et al. 2008). In some cases, fatty acid
analyses can provide detailed taxonomic information, including specific types
of algae assimilated (e.g. Napolitano et al. 1997).
Samples of muscle tissues (large amphibians), whole organisms (small
amphibians, prey items), and basal resources (algae, detritus) can be collected
in the same manner as described above for isotopic analyses, except that all
materials should be frozen at −80°C as quickly as possible. Moisture content
of materials is quantified with freeze drying so that lipids are not altered or
volatilized. Lipids are extracted from homogenized materials using chloroform/
methanol solvent extraction (Bligh and Dyer 1957). Analyses involve separation
of polar and non-polar lipids with a solid-phase extraction system, conversion
into fatty acid methyl esters, and separation using a gas chromatograph with a
flame ionization or mass spectrometer detector.
Discriminant function analysis and similar approaches can be used to assess
trophic interactions among sampled resources and consumers (Budge et al.
2002). As with isotopic analyses, resolution will vary among systems, and in
situations where trophic interactions are complex multiple approaches (e.g. fatty
acid analyses combined with isotopic analyses and/or gut content analyses) will
obviously produce more robust results.
5.6 Summary
Understanding the food habits of larval amphibians is actually the first step in
understanding the interactions of amphibians and the communities in which
they live and ultimately their ecological significance (Altig et al. 2007). Given the
current status of many amphibian species and populations, and the likelihood
that past and ongoing declines have ecosystem-level consequences (Ranvestel
et al. 2004; Whiles et al. 2006), there is urgent need for detailed, quantitative
information on the myriad ecological roles of larvae. The methods reviewed
herein provide the basic tools for assessing diet, nutritional ecology, and trophic
interactions. Choosing the proper technique ultimately depends on the research
question and resource limitations. Because each approach has inherent limita-
tions, we recommend combined approaches.
5.7 References
Altig, R., Whiles, M. R., and Taylor, C. L. (2007). What do tadpoles really eat? Assessing
the trophic status of an understudied and imperiled group of consumers in freshwater
habitats. Freshwater Biology, 52, 386–95.
Benke, A. C., Huryn, A. D., Smock, L. A., and Wallace, J. B. (1999). Length-mass rela-
tionships for freshwater macroinvertebrates in North America with particular reference
to the southeastern United States. Journal of the North American Benthological Society
18, 308–43.
Bligh, E.G. and Dyer, W. J. (1959). A rapid method of total lipid extraction and purifica-
tion. Canadian Journal of Biochemistry and Physiology 37: 911–917.
Bottrell, H. H., Duncan A., Gliwicz, Z. M., Grygierek, E., Herzig A., Hillbricht-Ilkowska
A., Kurasawa H., Larsson P., and Weglenska, T. (1976). A review of some problems in
zooplankton production studies. Norwegian Journal of Zoology, 24, 419–56.
Bowen, S. H., Lutz, E. V., and Ahlgren, M. O. (1995). Dietary-protein and energy as
determinants of food quality—trophic strategies compared. Ecology, 76, 899–907.
Brower, J. E., Zar, J. H., and von Ende, C. N. (1998). Field and Laboratory Methods for
General Ecology. McGraw-Hill, Boston, MA.
Budge, S. M., Iverson, S. J., Bowen, W. D., and Ackman, R. G. (2002). Among- and
within-species variability in fatty acid signatures of marine fish and invertebrates on the
Scotian Shelf, Georges Bank, and southern Gulf of St. Lawrence. Canadian Journal of
Fisheries and Aquatic Sciences, 59, 886–98.
Chesson, J. (1983). The estimation and analysis of preference and its relationship to for-
aging models. Ecology, 64, 1297–1304.
Culver, D. A., Boucherle, M. M., Bean, D. J., and Fletcher, J. W. (1985). Biomass of
freshwater crustacean zooplankton from length-weight regressions. Canadian Journal
of Fisheries and Aquatic Sciences, 42, 1380–90.
Cummins, K. W. and Wuycheck, J. C. (1971). Caloric equivalents for investigations
in ecological energetics. Mitteilungen Internationale Vereinigung für Theoretische und
angewandte Limnologie, 18, 1–158.
Dalerum, F. and Angerbjorn, A. (2005). Resolving temporal variation in vertebrate diets
using naturally occurring stable isotopes. Oecologia, 144, 647–58.
Duellman, W. E. and Trueb, L. (1986). Biology of Amphibians. McGraw-Hill, New York.
Dumont, H. J., Van de Velde, I., and Dumont, S. (1975). The dry weight estimate of
biomass in a selection of Cladocera, Copepoda, and Rotifera from the plankton, per-
iphyton and benthos of continental waters. Oecologia, 19, 75–97.
Dünker, N., Wake, M. H., and Olson, W. M. (2000). Embryonic and larval develop-
ment in the caecilian Icthyophis kohtaoensis: a staging table. Journal of Morphology, 243,
3–34.
Evans-White, M. A., Dodds, W. K., and Whiles, M. R. (2003). Ecosystem significance
of crayfishes and central stonerollers in a tallgrass prairie stream: functional differences
between co-occurring omnivores. Journal of the North American Benthological Society,
22, 423–41.
Finlay, J. C., Power, M. E., and Cabana, G. (1999). Effects of water velocity on algal car-
bon isotope ratios: implications for river food web studies. Limnology and Oceanography,
44, 1198–1203.
Fry, B. (2006). Stable Isotope Ecology. Springer, Berlin.

Hall, Jr, R. O. (1995). Use of a stable carbon-isotope addition to trace bacterial carbon
through a stream food-web. Journal of the North American Benthological Society, 14,
269–77.
Hall, Jr, R. O., Peterson, B. J., and Meyer, J. L. (1998). Testing a nitrogen-cycling model
of a forest stream by using a nitrogen-15 tracer addition. Ecosystems, 1, 283–98.
Hauer, F. R. and Lamberti, G. A. (eds) (2006). Methods in Stream Ecology. Academic
Press, San Diego, CA.
Hershey, A. E., Fortino, K., Peterson, B. J., and Ulseth, A. J. (2006). Stream food webs. In
F. R. Hauer and G. A. Lamberti (eds), Methods in Stream Ecology, pp. 637–9. Academic
Press, San Diego, CA.
Heyer, W. R., Donnelly, M. A., McDiarmid, R. W., Hayek, L.-A. C., and Foster, M. S.
(eds) (1994). Measuring and Monitoring Biological Diversity. Standard Methods for
Amphibians. Biological Diversity Handbook Series, Smithsonian Institution Press,
Washington DC.
Ivlev, V. S. (1961). Experimental Ecology of the Feeding of Fishes. Yale University Press, New
Haven, CT.
Kilham, S. S. and Pringle, C. M. (2000). Food webs in two neotropical systems as revealed
by stable isotope ratios. Verhandlungen Internationale Vereinigung für theoretische und
angewandte Limnologie, 27, 1768–75.
Koussoroplis, A. M., Lemarchand, C., Bec, A., Desvilettes, C., Amblard, C., Fournier,
C., Berny P., and Bourdier, G. (2008). From aquatic to terrestrial food webs: decrease
of the docosahexaenoic acid/linoleic acid ratio. Lipids, 43, 461–6.
Krebs, C. J. (2001). Ecological Methodology, 2nd edn. Harper Collins Publishers, New York.
Kupferberg, S. (1997). Facilitation of periphyton production by tadpole grazing:
Functional differences between species. Freshwater Biology, 37, 427–39.
Layman, C. A., Quattrochi, J. P., Peyer, C. M., and Allgeier, J. E. (2007). Niche width
collapse in a resilient top predator following ecosystem fragmentation. Ecology Letters,
10, 937–44.
Lowe, R. L. and LaLiberte, G. D. (2006). Benthic stream algae: distribution and struc-
ture. In F. R. Hauer and G. A. Lamberti (eds), Methods in Stream Ecology, pp. 327–56.
Academic Press, San Diego, CA.
Merritt, R. W., Cummins, K. W., and Berg, M. B. (2008). An Introduction to the Aquatic
Insects of North America, 4th edn. Kendall/Hunt Publishing Company, Dubuque, IO.
Muller-Navarra, D. C., Brett, M. T., Liston, A. M., and Goldman, C. R. (2000). A highly
unsaturated fatty acid predicts carbon transfer between primary producers and con-
sumers. Nature, 403, 74–7.
Napolitano, G. E., Pollero, R. J., Gayoso, A. M., MacDonald, B. A., and Thompson, R. J.
(1997). Fatty acids as trophic markers of phytoplankton blooms in the Bahia Blanca
Estuary (Buenos Aires, Argentina) and in Trinity Bay (Newfoundland, Canada).
Biochemistry and Systematic Ecology, 25, 739–55.
Peterson, B. J. and Fry, B. (1987). Stable isotopes in ecosystem studies. Annual Review of
Ecology and Systematics, 18, 293–320.
Peterson, C. G. and Boulton, A. J. (1999). Stream permanence influences microalgal food
availability to grazing tadpoles in arid-zone springs. Oecologia, 118, 340–52.
Post, D. M. (2002). Using stable isotopes to estimate trophic position: models, methods,
and assumptions. Ecology, 83, 703–18.
Ranvestel, A. W., Lips, K. R., Pringle, C. M., Whiles, M. R., and Bixby, R. J. (2004).
Neotropical tadpoles influence stream benthos: evidence for the ecological conse-
quences of decline in amphibian populations. Freshwater Biology, 49, 274–85.
Regester, K. J., Whiles, M. R., and Lips, K. R. (2008). Variation in the trophic basis of
production and energy flow associated with emergence of larval salamander assem-
blages (Ambystomatidae) among forest ponds. Freshwater Biology 53, 1754–67.
Rodgers, D. W. and Qadri, S. U. (1977). Seasonal variations in calorific values of some lit-
toral benthic invertebrates of the Ottawa River, Ontario. Canadian Journal of Zoology,
55, 881–4.
Rogers, L. E., Hinds W. T., and Buschom, R. L. (1976). A general weight vs. length rela-
tionship for insects. Annals of the Entomological Society of America, 69, 387–9.
Rosi-Marshall, E. J. and Wallace, J. B. (2002). Invertebrate food webs along a stream
resource gradient. Freshwater Biology, 47, 129–41.
Sanzone, D. M., Meyer, J. L., Marti, E., Gardiner, E. P., Tank, J. L., and Grimm, N. B.
(2003). Carbon and nitrogen transfer from a desert stream to riparian predators.
Oecologia, 134, 238–50.
Skelly, D. K. and Golon, J. (2003). Assimilation of natural benthic substrates by two spe-
cies of tadpoles. Herpetologica 59, 37–42.
Smith, D. G. (2001). Pennak’s Freshwater Invertebrates of the United States, Porifera to
Crustacea. John Wiley and Sons, New York.
Solé, M., Beckmann, O., Pelz, B., Kwet, A., and Engels, W. (2005). Stomach-flushing for
diet analysis in anurans: an improved protocol evaluated in a case study in Araucaria
forests, southern Brazil. Studies on Neotropical Fauna and Environment, 40, 23–8.
Solomon, C. T., Flecker, A. S., and Taylor, B. W. (2004). Testing the role of sediment-
mediated interactions between tadpoles and armored catfish in a neotropical stream.
Copeia, 2004, 610–16.
Steidl, R. J., Hayes, J. P., and Schauber, E. (1997). Statistical power analysis in wildlife
research. Journal of Wildlife Management, 61, 270–9.
Steinman, A. D., Lamberti, G. A., and Leavitt, P. R. (2006). Biomass and pigments of
benthic algae. In F. R. Hauer and G. A. Lamberti (eds), Methods in Stream Ecology,
pp 352–79. Academic Press, San Diego, CA.
Stuart, S. N., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues, A. S. L., Fischman, D. L.,
and Waller, R. W. (2004). Status and trends of amphibian declines and extinctions
worldwide. Science, 306, 1783–6.
Stübing, D., Hagen, W., and Schmidt, K. (2003). On the use of lipid biomarkers in mar-
ine food web analyses: an experimental case study on the Antarctic krill, Euphausia
superba. Limnology and Oceanography, 48, 1685–1700.
Sushchik, N. N., Gladyshev, M. I., Moskvichova, A. V., Makhutova, O. N., and
Kalachova, G. S. (2003). Comparison of fatty acid composition in major lipid classes
of the dominant benthic invertebrates of the Yenisei River. Comparative Biochemistry
and Physiology, 134B, 111–22.
Thorpe, J. H. and Covich, A. P. (2001). Ecology and Classification of North American
Freshwater Invertebrates. Academic Press, San Diego, CA.
Vander Zanden, M. J. and Rasmussen, J. B. (1999). Primary consumer 15N and 13C and
the trophic position of aquatic consumers. Ecology, 80, 1395–1404.
Vander Zanden, M. J., Shuter, B. J., Lester, N. P., and Rasmussen, J. B. (2000). Within-
and among-population variation in the trophic position of a pelagic predator, lake
trout (Salvelinus namaycush). Canadian Journal of Fisheries and Aquatic Sciences, 57,
725–31.
Verburg, P., Kilham, S. S., Pringle, C. M., Lips, K. R., and Drake, D. L. (2007). A sta-
ble isotope study of a neotropical stream food web prior to the extirpation of its large
amphibian community. Journal of Tropical Ecology, 23, 643–51.
Whiles, M. R., Lips, K. R., Pringle, C. M., Kilham, S. S., Bixby, R. J., Brenes, R., Connelly, S.,
Colon-Gaud, J. C., Hunte-Brown, M., Huryn, A. D. et al. (2006). The effects of
amphibian population declines on the structure and function of Neotropical stream
ecosystems. Frontiers in Ecology and Environment, 4, 27–34.
Yamamuro, M. and Kayanne, H. (1995). Rapid direct determination of organic-carbon
and nitrogen in carbonate-bearing sediments with a Yanaco Mt-5 Chn Analyzer.
Limnology and Oceanography, 40, 1001–5.
6
Aquatic mesocosms
Raymond D. Semlitsch and Michelle D. Boone
6.1 Introduction
We have often told our students that any experiment can be easily criticized for
lack of realism because all experiments, whether in the laboratory or field, by
design are contrived by humans and are not natural. However, in spite of such
criticism the more challenging issue is coming up with new venues or innova-
tive designs to answer important ecological or conservation questions and to
balance trade-offs. As amphibian ecology has become more experimental, we
have been confronted with numerous trade-offs concerning the choice of venue,
design, realism, and replication. Each choice is often coupled with benefits as
well as limitations. As in many rapidly developing fields, criticism is an expected
and healthy process that pushes investigators to think about their choices and
develop solutions. The use of mesocosms in studies of aquatic ecology across
an array of taxonomic groups has been central in that debate (e.g. Kimball and
Levin, 1985; Carpenter 1996; Schindler 1998). Within the field of amphibian
biology, a number of reviews also have stimulated a healthy discussion on the use
of experiments (Hairston 1989a, 1989b), and in particular, use of mesocosms
(e.g. Jaeger and Walls 1989; Rowe and Dunson 1994; Skelly and Kiesecker
2001; Skelly 2002).
The purpose of our chapter is to provide background on the use of aquatic
mesocosms in ecological and conservation studies of amphibians. We describe
the history, various types, set-ups, and designs and present selected case studies
to emphasize their versatility. We also provide examples and comments on limi-
tations of the technique and strengths of inference that come with the results.
Our goal is to provide readers with an effective technique to answer a range of
ecological and conservation questions.
6.2 Historical background

The use of aquatic mesocosms in amphibian ecology was born out of the need
to move away from purely descriptive field studies and correlative analyses
and towards more manipulative studies that could test hypotheses to establish
cause and effect (see Wilbur 1987). Studies using aquatic mesocosms are often
viewed as the best compromise between the variability of natural wetlands, con-
founding factors, and the lack of control that is common in field studies and
the ability to manipulate single variables, replicate treatments, and randomly
assign treatments to experimental units in laboratory experiments. They have
sometimes been referred to as “hybrid” or “semi-field” experiments (Rowe and
Dunson 1994). Aquatic mesocosms differ from “microcosms” in that they are
self-sustaining once established and they do not need input of additional nutri-
ents or resources after initial set-up, thus allowing study organisms to complete
critical phases of their life cycle. To be effective, mesocosms must be capable of
rearing populations of amphibian larvae, simulate important components of the
aquatic ecosystem, and be reproducible and affordable. For purposes of this dis-
cussion, we refer primarily to two common types of aquatic mesocosm: cages or
pens (Figure 6.1) that are placed in a natural pond or stream to capture location
effects such as habitat variability and containers including plastic tubs, wading
pools, or cattle watering tanks (Figure 6.2) that are filled with water and have
components of aquatic ecosystems added to simulate natural variation. Artificial
streams fit in the category of container mesocosms but have a flow-through or
recirculating water source or system. Any of these types can be large or small and
contain few or many components of a natural system; in this way they form a
continuum between laboratory microcosms and natural field studies.
To the best of our knowledge, Warren Brockelman was the first to use
pens placed in natural ponds for experimental amphibian studies in 1966–7
in Michigan, USA (Brockelman 1969). He used wooden framed pens (0.61 m
wide 2.44 m long 0.6 m high) covered with fiberglass window screening
and with lids, half of which had screen bottoms and half had bottoms open to
sediments. His study addressed the effects of intraspecific competition on Bufo
americanus tadpole growth, development, and survival.
The first use of containers for experimental amphibian studies was by Henry
Wilbur and Joe Travis in 1977–9 in North Carolina (Wilbur and Travis 1984).
They used cattle watering tanks (galvanized steel, round, 3.58 m2 surface area,
0.61 cm deep) buried in the ground flush to the surface and filled with natural
pond water. Replicate tanks were positioned in different habitat types to test dif-
ferences in the colonization of aquatic habitats by invertebrates and amphibians,
6 Aquatic mesocosms | 89
Fig. 6.1 Examples of different cages used to rear amphibian larvae.
Full-Sun Treatment
High-Shade Treatment
Low-Shade Treatment
Fig. 6.2 Examples of different cattle tanks and pools used to rear amphibian larvae.
and the formation of community structure. In 1980, one of Henry Wilbur’s

graduate students, Peter Morin, subsequently developed a more efficient and
novel venue for experiments using cattle tanks to manipulate factors important
to community regulation (Morin 1981). He placed cattle tanks (epoxy-coated
galvanized steel, round, 1.52 m diameter, 1000 L) in a field, above ground, so
that they could be filled, manipulated, and easily drained to initiate new experi-
ments each year. Each tank was filled with tap water and then each received an
equal amount of dried litter collected from the edge of a natural pond (550 g)
and Purina Trout Chow (50 g) for substrate and a source of nutrients. Equal
amounts of inoculum of plankton derived from pooled plankton collections
from eight natural ponds were added to each tank to initiate a diverse food web
and prey base for larvae. This latter approach of using cattle tanks as aquatic
mesocosms that could be renewed for each question and study has been critical
in the advancement experimental amphibian ecology.
6.3 Why use mesocosms?

Studies in nature and in the laboratory allow researchers to answer important
questions, but both have limitations (Diamond 1986). While laboratory stud-
ies allow us to maximize control and manipulate factors so that cause–effect
relationships can be established, the laboratory lacks realism and complexity,
reducing its application in the real world. In contrast, field studies offer realism
and can encompass the complexity of natural food webs, but it can be difficult
to distinguish cause–effect relationships and there can be high levels of vari-
ation among replicates. Mesocosms are the middle ground. Proponents argue
that mesocosms offer the best of both worlds: control, the ability to determine
cause–effect relationships, natural food webs, and realism (Wilbur 1987; Rowe
and Dunson 1994; Drake et al. 1996; Boone and James 2005). Critics, however,
contend that mesocosms offer little gain over laboratory studies in that they are
still artificial and may have limited application to the real world, yet they require
more work and time, and they have greater variation among replicates than
laboratory studies (Carpenter 1996; Skelly 2002). Certainly, mesocosms would
not be mistaken for natural ponds, but they incorporate many natural elements
that allow them to mimic some but not all of the processes occurring in larger
natural ponds or lakes (Schindler 1998).
There are a number of reasons why mesocosms can be a useful technique in
your research program. First, mesocosms allow you to incorporate some compo-
nents of natural systems into your experiments—the changes in season, photo-
period, temperature, and habitat that animals experience in nature—something
that you typically cannot alter in the laboratory. Secondly, if you have evaluated
a mechanism within or between organism(s), or developed a theoretical model,
then a logical next step is to see if your predictions hold in a more complex and nat-
ural environment; mesocosm studies are a logical next step. Thirdly, if you want to
evaluate single factors in isolation, a mesocosm may be ideal. Fourthly, manipula-
tions in mesocosms can help you eliminate less important factors and find system
drivers, so that you can design a better, more efficient large-scale field study.
If you need a short-term answer to a basic experimental question, then a labora-
tory experiment can offer an expedient answer: can trematodes infect limb buds?
What concentration of an insecticide is lethal to tadpoles or zooplankton? Do
overwintered bullfrog tadpoles prey upon recently hatched tadpoles? If your ques-
tion is site- or habitat-specific or requires a vast area to encompass the entire ecosys-
tem, then field experiments may be the best way to go: can ponds in city parks and
golf courses support native fauna? What effects do forest-management practices
have on amphibian communities? Studies in mesocosms provide a useful surro-
gate for the field and allow us to evaluate single and multiple factors that may be
important at a larger scale, which allows us to better design larger field studies and
provides insight into the system: do changes in food webs from eutrophication
cause increases in number of trematodes and in the number of infected amphib-
ians? Does a sublethal pesticide used on golf courses cause changes in aquatic
community dynamics? Although mesocosms are often referred to as a “black box,”
suggesting that the mechanisms underlying the outcome are not clear, additional
studies can be done to discern mechanisms by behavioral observations, periodic
sampling of covariates (e.g. temperature, primary production, dissolved carbon),
or with a paired laboratory study (e.g. prey palatability).
Mesocosms are useful in addressing a variety of questions in ecology, evolu-
tion, behavior, and conservation (see case studies below). We surveyed the litera-
ture from 2004 to 2007 through the Ecological Society of America (including
the journals Ecology, Ecological Applications, Frontiers in Ecology and Evolution,
and Ecological Monographs), Oecologia, Evolution, Animal Behavior, and Behavior
by searching the terms “amphibian” or “frog” or “salamander” and then select-
ing all studies that used mesocosms. A total of 30 articles using mesocosms or
enclosures were found only in the journals Oecologia or of the Ecological Society
of America. The studies addressed a range of questions in ecology (e.g. species
interactions, population biology, community structure), evolution (e.g. pheno-
typic plasticity), behavior (e.g. oviposition selection), and conservation (e.g.
effects of pathogens, habitat fragmentation, pesticides). Whereas mesocosms
have been historically used for basic questions of ecology and evolution, they are
becoming more frequently used in the field of conservation and have enormous
potential for use in regulatory fields such as ecotoxicology (reviewed by Boone

and James 2005).
The appropriateness of a mesocosm as a representative of a natural system may
depend on many things, including the size of the mesocosm and the species or
community of interest. For studies with amphibian, invertebrate, and plankton
communities, aquatic mesocosms ranging from 200 to more than 1500 L (see
Figure 6.3a) were self-sustaining and maintained a similar temperature regime
(a)
14
Mesocosms
12 Enclosures
10
Number of studies
0
<120 200–450 750–1300 >1500
Volume (L)
(b)
12
10
Number of studies
0
0 0.1–0.3 0.875–1.8
Leaf litter (kg)
Fig. 6.3 Abundance of mesocosm studies conducted: (a) using different volumes and
(b) with different amounts of leaf litter (see text for methods and results).
to natural communities. This is an important limitation because mesocosms

of these sizes may not be appropriate for studies with larger species like wide-
ranging predatory fish or larger-scale ecosystem processes like nutrient cycling.
What is important for effective use is that the amphibians of interest, inverte-
brate prey, and plankton resources can complete major portions of their life cycle
and be self-sustaining in the mesocosm. Isolated, temporary pond communities
range in size from tire ruts to large shallow wetlands, and have characteristics
that differ from larger lake systems that may be deeper and experience turnover
and temperature inversions. Therefore, effective use of mesocosms is likely lim-
ited to species and processes that occur in shallow, temporary wetlands.
Additionally, there is evidence indicating that results generated from meso-
cosm experiments have high predictability to the field. Mesocosm studies evalu-
ating the role of competition, predation, and pond drying (e.g. Wilbur 1972,
1987; Morin 1981, 1983; Semlitsch 1987) have proven useful in understanding
the structure and regulation of amphibian communities. Long-term monitoring
at a natural wetland (Rainbow Bay) in South Carolina, USA, has shown that
community structure is determined by competitive and predatory interactions
occurring along a gradient of pond drying (Semlitsch et al. 1996), which was
predicted by mesocosm research. Manipulations with sublethal pesticide expos-
ure in mesocosms show zooplankton populations are reduced and periphyton
abundance increases for short periods of time, which can result in positive effects
on tadpoles despite the pesticide having sublethal but negative effects on tadpoles
directly (e.g. Boone and James 2003; Boone and Bridges-Britton 2006). These
studies suggest that changes in the food web may be more important than direct
negative effects of the contaminant alone (Mills and Semlitsch 2004). When
these studies were replicated under more natural field conditions in ponds, simi-
lar outcomes were found despite the fact that predators were not excluded and
that zooplankton populations could have recovered more quickly due to ephippia
present in the soil sediments (Boone et al. 2004). Studies of dragonfly preda-
tion on tadpoles have documented qualitatively similar results between natural
pools and mesocosms (Gascon 1992) and between different sizes of mesocosms
(Gascon and Travis 1992). Such examples lend strong support to the usefulness
of mesocosms, particularly given that mesocosm manipulations require less time,
money, and effort to conduct, relative to the large field studies.
6.4 Types of mesocosm

There are a number of types of mesocosm, including cattle watering tanks or wad-
ing pools (Figure 6.2) and field enclosures (Figure 6.1). Cattle tank ponds can be
purchased in galvanized steel (which must be sealed with an epoxy paint or lined)
and polyethylene plastic (Behlen or Rubbermaid produce tanks in blue or black).
The polyethylene ponds are most commonly used today because they are long-
lasting (10 years), are not prone to rusting, and do not require sealing as metal tanks
do. Wading pools are useful and cheap but short-lived (1–2 years) and usually made
of PVC/vinyl, which can leach phthalates, chemicals that are associated with endo-
crine disruption. Water tests found that Behlen polyethylene tanks did not leach
estrogenic compounds (S.I. Storrs, personal communication). However, wading
pools can be especially useful to address short-term behavioral questions related to
factors that may influence oviposition (Vonesh and Buck 2007) or plasticity in tad-
pole morphological development associated with predators and competition which
may occur in a matter of weeks (Relyea 2004; Relyea and Auld 2005).
Field caging studies enable us to study a discrete location or habitat while
moving toward greater realism. Field cages can range in size from small to large.
In our survey, most of the cages were relatively small (2.4–120 L; Garcia et al.
2004; Griffis-Kyle and Ritchie 2007; Ireland et al. 2007) or large (pond size or
sections; Scott 1990; Loman 2004; Boone et al. 2004). Aquatic field cages are
often made of fine screen mesh, such as fiberglass window screen, mosquito net-
ting, or plastic meshes, which are exposed to the larger matrix allowing for water
exchange and influx of aquatic life smaller than the mesh size. Alternatively,
field cages can also be made of polyethylene plastics that will prevent the flow
of materials into the enclosure after initial set-up and also prevent things like
pesticides or predators from moving into or out of the enclosure. These field
cages whether made of permeable or impermeable material can be attached to
untreated lumber into a box or rectangle to be placed in a pond (Figure 6.2). If
the screen, mesh, or fabric has rigidity, it is also possible to use it to construct
circular cages to hold animals (Figure 6.2). Alternatively, “bags” can be sewn or
fabricated that can be supported with an external structure like multiple stakes,
drinking water pipe, and/or floatation devices (Figure 6.2).
The advantage of field cages is that you can study sites of particular interest
and ask questions in the relevant environment giving your study greater realism
and the ability to manipulate factors of interest. In contrast, field cages often do
not represent independent observations (unlike wading pools or tanks) given
that there is the potential for exchange among enclosures. For example, if a
researcher was examining how a fish predator may influence growth and devel-
opment of tadpoles, the presence of fish in the pond or the presence/absence of
fish as a treatment could influence all the enclosures if there was water exchange.
However, these potential problems can be eliminated or minimized with choice
of mesh and some design considerations for the cages.
6.5 Setting up mesocosms

There are a number of ways to set-up mesocosms for larval amphibian studies,
with cattle tank ponds and wading pools generally established with a standard-
ized volume of water (tap or well water), a nutrient base (leaves or grass and/
or rabbit chow), and plankton from natural ponds. Larger volumes of water
(900–1500 L) have the advantage of being more stable environments, especially
in terms of temperature extremes in summer, and less prone to predator inva-
sion. Smaller volumes of water can be shaded and raised above the ground to
reduce likelihood of predators moving through. In our survey mentioned above,
most mesocosm studies were conducted in tanks holding 750–1300 L of water
(Figure 6.1). Water is typically allowed to sit between 2 days and 2 weeks so that
chlorine from the tap water will dissipate and so that algal and zooplankton
inoculum can establish.
The addition of leaf litter, grass, rabbit chow, and/or some other nutrient source
is a standard practice in mesocosm experiments. Natural litter substrates serve as
a slow-release nutrient source. Rabbit chow serves as a fast-release nutrient source.
Both of these will support the aquatic food web throughout the field season and
can influence other relevant factors like the amount of ultraviolet radiation pene-
tration of the water (by influencing abundance of dissolved organic compounds)
and dissolved oxygen. Researchers have used different types and amounts of
nutrients, which can impact the aquatic system. In our survey, we found that
roughly an equal number of studies used relatively low or high amounts of leaf lit-
ter (Figure 6.3). Most enclosure studies did not add nutrients unless it was being
explicitly tested, presumably because there were adequate flow of nutrients and
food into the enclosures. However, leaf litter or vegetation can be added to serve
as refugia for amphibian prey within the enclosures.
The amount and type of nutrient addition can have an effect on the aquatic
community because it sets the energy base for the system. A recent study by
Williams et al. (2008) using dead grass or deciduous leaf litter found higher
productivity in systems with grass, which resulted in greater tadpole growth. In
studies manipulating the amount of leaf litter, Rubbo et al. (2008) and Boone
(unpublished data) showed that increasing the amount of leaf litter increased
the productivity of the system, and hence the survival and growth of tadpoles.
Additionally, Rubbo and Kiesecker (2004) found the species from which the
leaf litter originated had an effect on amphibian biomass and survival, with oak
leaves having positive effects on amphibians compared to maple leaves. The type
of leaf litter or grass can influence the nutrient base of the community and influ-
ence the algal and bacterial communities that flourish there. Consideration of
the type of natural system you are attempting to represent (e.g. forested pond,
grassy pond, or marsh) is the best guide to selecting the appropriate nutrient
base to add.
It is important to establish the primary components of an aquatic food web in
mesocosm tanks in order for them to be self-sustaining. A complex community
of both phyto- and zooplankton is critical to help balance nutrient exchange,
bacterial growth, and nitrogenous waste production as well as serve as a food
source for amphibian larvae. Temporary wetlands are often an ideal source of
inoculum for mesocosms because there will be fewer insect predators and no
fish. All inoculum added to mesocosms should be strained to eliminate detri-
mental insect or fish larvae, and eggs or larvae of non-target amphibian species.
Several additions over time and from varying natural ponds will help ensure
adequate diversity and complexity.
6.6 Common experimental designs

Establishing cause–effect relationships is the hallmark of mesocosm approaches
and allows complex multifactorial designs. Designs using pools or tanks allow
clear tests of each factor independently. Mesocosm studies allow researchers to
address how incremental changes in a treatment level impact the study organ-
ism and how different treatments may interact with one another. This level of
manipulation would be difficult to achieve under most field conditions, but is
a critical tool for understanding how amphibian populations and communities
function.
Replicate tanks or pools are physically separate pond communities that can
be established with standardized procedures but diverge if treatment differences
occur. Each tank represents an independent replicate and experimental unit.
Cages located within ponds are typically not independent from one another;
therefore, the question of interest will determine how these field experiments
are replicated. If you are interested in how a particular type of habitat influences
an amphibian community, then you should replicate at the level of pond-habitat
type. For instance, to examine whether amphibians can survive larval develop-
ment in golf course ponds, Boone et al. (2008) used multiple golf course and
reference ponds with replicate cages both within and between ponds. Using
a single golf course and reference pond with replicate enclosures within each
pond would not have clarified whether golf-course management affected larval
amphibians; instead, it would only confirm that the two ponds are different
for some reason, perhaps due to golf course management, historically differ-
ent abiotic and biotic factors present in the pond, or chance. Replicate cages in
the latter case would constitute “pseudoreplicates” (Hurlbert 1984) and would
increase the probability of making a type I error or determining there is a sig-
nificant difference related to treatment when there is not. However, if you are
interested in how overwintered bullfrog tadpoles may influence the amphibian
community, then manipulating bullfrog abundance in replicate cages in a single
pond would be acceptable. Replication in field cage studies requires replicating
at the level of your question and factor of influence (e.g. golf course), which
will likely be at separate locations. Replication is also necessary within ponds to
determine variation among specific cage locations (e.g. aspect, slope, or shading
along a shoreline). Split-plot or nested designs are especially useful in studies
using cages in ponds and may allow the testing of hypotheses not testable with
less sophisticated designs (Gunzburger 2007).
The number of replicates you should use is also an important consideration.
Typically there is a trade-off between the number of replicates and the total
size of the study related to logistical and financial constraints. However, the
more divergent your replicates, the greater the variation within your treatments,
and the less your ability to detect significant differences between treatments.
Therefore, consideration of the level of variation inherent in your replicates will
influence the number of replicates that you need. While laboratory studies fre-
quently have five or more replicates because of the ease of replication, mesocosm
tank studies will often be limited to three to five replicates (Boone and James
2005), which is typically sufficient to detect significant differences in behavior,
morphology, development, and survival. You can estimate a priori how many
replicates you should use by determining your desired statistical power (1−;
a power of 0.8 or greater is desirable) based on your criteria and the effect
size (size of the difference). Using a statistical program like SAS (Proc Power)
or using online calculators (e.g. www.math.yorku.ca/SCS/Online/power/) you
can determine the ideal number of replicates to use based on the system you are
studying.
The random assignment of treatments is one of the most critical steps of all
true experiments and should be applied to mesocosm studies. Randomizing
your treatments is extremely important to avoid confounding variation
among replicate tanks or cages with variation among treatments. You can use
a random number chart or a random number generator in SAS (Proc Plan).
Treatments may be randomly assigned across all tanks or cages (completely
random design), or they may be blocked so that experimental units are grouped
along a known gradient that may cause variation (e.g. shading) not relevant to
the experimental question and which can be removed statistically (randomized
complete block).
6.7 Case studies

6.7.1 Community ecology
The study of community ecology has had a long history of experimental studies
to help disentangle the complex interactions among species. Morin (1981) used
three species of anuran tadpoles (Scaphiopus holbrookii, Pseudacris crucifer, and
Bufo terrestris) as competitors and adult newts (Notophthalmus viridescens) as a
predator to address the role of interspecific competition. Tadpoles were reared
together from hatching to metamorphosis in 16 cattle tanks (epoxy-coated gal-
vanized steel, 1.52 m diameter, 1000 L) at constant initial densities in all tanks.
Newts were added in four densities (none, two, four, and eight per tank) to vary
the intensity of predation and replicated four times each. The results showed
that in the absence of newt predators Scaphiopus (62%) and Bufo (33%) meta-
morphs were relatively abundant, followed by Pseudacris (5%). However, in
the presence of a high density of newts, Pseudacris (68%) metamorphs were
most abundant followed by Scaphiopus (20%) and Bufo (12%). The rarity of
Pseudacris in tanks without newts was best explained by intense interspecific
competition with Scaphiopus and Bufo. The increased survival and larger body
mass of Pseudacris in tanks with newts was best explained by release from inter-
specific competition due to newt predation on the preferred prey of Scaphiopus
and Bufo tadpoles. Morin suggested that newts acted as a ‘keystone’ predator
to prevent exclusion of competitively inferior species by reducing the survival
and density of competitively superior species. These results demonstrate that
predation and competition could interact to produce deterministic patterns of
community structure in amphibians.
In a subsequent experiment, Wilbur (1987) tested the interaction effects of
competition, predation, and disturbance on the structure of a similar aquatic
amphibian community. He used four species of anuran tadpoles (Rana sphe-
nocephala, S. holbrookii, P. crucifer, and B. terrestris) as competitors and adult
newts (Notophthalmus viridescens) as a predator. Tadpoles of the four species
were reared together in 36 cattle tanks (epoxy-coated galvanized steel, 1.52 m
diameter, 1000 L) at a constant high or low density. Four newts were added to
half the tanks and ponds were drained at three rates; either 50 days, 100 days,
or control tanks kept at 50 cm deep. All 12 treatment combinations were repli-
cated in three spatial blocks to account for physical differences across a field were
tanks were positioned. The results showed that Scaphiopus was least sensitive to
intraspecific density, metamorphosed quickly in fast-drying tanks, and was the
competitive dominant in tanks without newt predators, but suffered highest
predation from newts. Rana was most successful at low density but could not
grow fast enough to metamorphose in fast-drying tanks and suffered high pre-
dation by newts. Bufo were very sensitive to high density and produced few met-
amorphs without newt predators. This effect was reversed by newts selectively
preying on Scaphiopus and Rana tadpoles, especially in rapidly drying tanks.
Pseudacris performed poorly in all slow-drying tanks but showed moderate suc-
cess in high-density tanks where newts had removed most competitors. The
overall conclusion of this study was that all three factors interacted with each
other and with species’ life history to determine community structure. Further,
it suggested that competition dominates simple habitats of short duration and
predation becomes increasingly important in ponds with longer hydroperiods
that permit the establishment of diverse predators. Both of the studies described
above were influential in focusing attention on the complex interactions of pre-
dation, competition, and pond drying in aquatic communities rather than his-
torical single-factor arguments.
6.7.2 Evolutionary ecology

The “common-garden” experimental approach to evolutionary studies has been
important for clarifying genetic and environmental interactions by measuring
genotype performance in a simplified controlled environment (Clausen et al.
1947). Parris et al. (1999) examined the larval performance of hybrid and par-
ental genotypes of the southern and plains leopard frog (Rana sphenocephala
and Rana blairi) in a common-garden approach using cattle tanks. They estab-
lished 56 experimental populations in cattle tanks (polyethylene, round, 1.83 m
diameter) by adding 960 L of tap water, 1.0 kg of deciduous leaf litter, and mul-
tiple additions of plankton from natural ponds. Six genotypes of hatchling tad-
poles were added to tanks at a low (20 per tank) or high (60 per tank) density to
simulate a competitive gradient. The 12 treatment combinations were replicated
either three or five times in the 56 tanks (Parris et al. 1999). Survival was the
same among all genotypes, but decreased with increasing density. Body mass
at metamorphosis was the same among genotypes, but decreased with increas-
ing density. Tadpoles of R. sphenocephala had the shortest larval period while
the other parental species R. blairi had the longest larval period. Ponds with
hybrid genotypes produced a greater proportion of metamorphs than those with
R. blairi tadpoles but a smaller proportion than ponds with R. sphenocephala
tadpoles. Their results demonstrate equal or increased larval performance of
hybrids relative to parental genotypes in cattle tank environments. Studies such
as the one described above are critical for determining whether or not natural
hybridization could potentially lead to the production of recombinant geno-
types possessing novel adaptations.
6.7.3 Ecotoxicology
The field of toxicology historically has taken a reductionist approach by using
single species and chemical testing in the laboratory. However, there is growing
effort to develop multispecies, community, and ecosystem-level experimental
approaches that better mimic real-world problems (Kimball and Levin 1985;
Rowe and Dunson 1994; Boone and James 2005; Semlitsch and Bridges 2005).
A model study by Boone et al. (2007) examined the role that chemical con-
tamination, competition, and predation play singly and in combination in an
aquatic amphibian community. They established replicate aquatic communities
in 64 polyethylene cattle tanks (round, 1.85 m diameter, 0.6 cm height, 1480 L
volume) by adding 1000 L of tap water, 1.0 kg of leaf litter, plankton from nat-
ural ponds, 60 Bufo americanus tadpoles, 30 R. sphenocephala tadpoles, and 10
Ambystoma maculatum salamander larvae to each. Four factors were manipulated
and replicated four times in a fully factorial design: no or two bluegill sunfish,
no or six overwintered bullfrog tadpoles, 0 or 2.5 mg/L carbaryl (an insecticide),
and 0 or 10 mg/L ammonium nitrate fertilizer. The results showed that bluegills
had the largest impact on the community by eliminating B. americanus and A.
maculatum and reducing the abundance of R. sphenocephala. Chemical contam-
inants had the second strongest effect on the community with the insecticide
reducing A. maculatum abundance by 50% and increasing the mass of anurans
(R. sphenocephala and B. americanus) at metamorphosis; the fertilizer positively
influenced time and mass at metamorphosis for both anuran and salamander
larvae. Presence of overwintered bullfrog tadpoles reduced mass and increased
time to metamorphosis of the anurans. Although both bluegill and overwin-
tered bullfrog tadpoles had negative effects on the amphibian community, they
performed better in the presence of one another and in contaminated ponds.
These results indicate that predicting complex interactions from single-factor
effects may not be straightforward. Further, the research supports the hypoth-
esis that combinations of factors can contribute to population declines.
6.7.4 Land use and management

Only recently have ecologists begun to use mesocosms to address questions con-
cerning human land use and habitat fragmentation. Hocking and Semlitsch
(2007) tested the effects of timber harvest on oviposition-site selection by gray
treefrogs (Hyla versicolor) in four experimental forest treatments. Five plastic
wading pools (1.5 m diameter, 30 cm deep) were placed above ground in each of
the four forest treatments: (1) a clear-cut with high coarse woody debris (CWD),
(2) a clear-cut with low CWD, (3) a partial cut with 50% canopy removal,
and (4) a control. The five pools were placed at three distances from the nat-
ural breeding pond in each treatment (see Figure 1 in Hocking and Semlitsch
2007): one pool at 10 m, two pools at 64 m, and two pools at 115 m from
a natural breeding pond. The four forest treatments were positioned around
three replicate natural breeding ponds for a total of 60 wading pools. Pools
filled naturally with rainwater by April. Following the first oviposition event,
all pools were checked every 24–48 h for treefrog eggs. Eggs were removed
and counted. The results indicated that treefrogs laid significantly more eggs in
the clear-cuts (low-CWD 77 185 eggs; high-CWD 51 990 eggs) than in either
the partial (13 553 eggs) or the control (14 068 eggs) treatments. Further, the
interaction of forest isolation treatment showed that oviposition in both clear-
cut treatments decreased with increasing isolation whereas oviposition was the
same at all levels of isolation in the partial and control treatments. Hocking and
Semlitsch (2007) suggested that the strong preference for open-canopy pools
likely benefits the larval stage because of higher-quality food and warmer water
temperatures but that isolation of breeding pools beyond 50–100 m can inhibit
oviposition by female treefrogs. They also add that the possible benefits to one
stage (tadpoles) might be diminished by the risk of desiccation to the terrestrial
juvenile stage after metamorphosis. Thus, mesocosms can be used to address
land-use questions beyond the boundary of aquatic ecosystems.
6.8 Conclusion
Although we are both advocates for the use of mesocosms in amphibian ecology
and conservation, we also agree that a pluralistic approach to any question is
likely necessary for complete understanding. The use of laboratory experiments
to understand mechanisms as well as field studies to anchor mesocosm results
back to the real world are critical. The use of cattle watering tanks, wading
pools, and cages is a powerful technique in any researcher’s toolbox. The choice
of mesocosm should reflect knowledge of the natural history of the species and
system being studied. If used properly, at the correct scale for the questions one
is asking, many questions can be clearly answered and results can have strong
inferences back to natural systems. However, we caution readers that neither
type of mesocosm described here can address all ecological and conservation
questions alone. One has only to read Schindler (1998) to realize that most
large-scale ecosystem processes cannot be replicated in cattle tanks or cages.
But, for many population- and community-level studies of amphibians, meso-
cosms offer an excellent approach.
6.9 References
Boone, M. D. and Bridges-Britton, C. M. (2006). Examining multiple sublethal con-
taminants on the gray treefrog, Hyla versicolor: effects of an insecticide, herbicide, and
fertilizer. Environmental Contamination and Chemistry, 25, 3261–5.
Boone, M. D. and James, S. M. (2003). Interactions of an insecticide, herbicide, and
natural stressors in amphibian community mesocosms. Ecological Applications, 13,
829–41.
Boone, M. D. and James, S. M. (2005). Use of aquatic and terrestrial mesocosms in eco-
toxicology. Applied Herpetology, 2, 231–57.
Boone, M. D., Semlitsch, R. D., Fairchild, J. F., and Rothermel, B. B. (2004). Effects of
an insecticide on amphibians in large-scale experimental ponds. Ecological Applications,
14, 685–91.
Boone, M. D., Semlitsch, R. D., Little, E. E., and Doyle, M. C. (2007). Multiple stressors
in amphibian communities: interactive effects of chemical contamination, bullfrog
tadpoles, and bluegill sunfish. Ecological Applications, 17, 291–301.
Boone, M. D., Semlitsch, R. D., and Mosby, C. (2008). Suitability of golf course ponds
for amphibian metamorphosis when bullfrogs are removed. Conservation Biology, 22,
172–9.
Brockelman, W. Y. (1969). An analysis of density effects and predation in Bufo americanus
tadpoles. Ecology, 50, 632–43.
Carpenter, S. R. (1996). Microcosm experiments have limited relevance for community
and ecosystem ecology. Ecology, 77, 677–80.
Clausen, J., Keck, D. D., and Hiesey, W. M. (1947). Heredity of geographically and eco-
logically isolated races. American Naturalist, 81, 114–33.
Diamond, J. (1986). Overview: laboratory experiments, field experiments, and natural
experiments. In J. Diamond and T. Case (eds), Community Ecology, pp. 3–22. Harper
and Row Publishers, New York.
Drake, J. A., Huxel, G. R., and Hewitt, C. L. (1996). Microcosms as models for generat-
ing and testing community theory. Ecology, 77, 670–7.
Garcia, T. S., Stacy, J., and Sih, A. (2004). Larval salamander response to UV radiation
and predation risk: color change and microhabitat use. Ecological Applications, 14,
1055–64.
Gascon, C. (1992). Aquatic predators and tadpole prey in central Amazonia: field data
and experimental manipulations. Ecology, 73, 971–80.
Gascon, C. and Travis, J. (1992). Does the spatial scale of experimentation matter? A test
with tadpoles and dragonflies. Ecology, 73, 2237–43.
Griffis-Kyle, K. L. and Ritchie, M. E. (2007). Amphibian survival, growth, and devel-
opment in response to mineral nitrogen exposure and predator cues in the field: An
experimental approach. Oecologia, 152, 633–42.
Gunzburger, M. S. (2007). Habitat segregation in two sister taxa of hylid treefrogs.
Herpetologica, 63, 301–10.
Hairston, Sr, N. G. (1989a). Hard choices in ecological experimentation. Herpetologica,
45, 119–22.
Hairston, Sr, N. G. (1989b). Ecological Experiments. Cambridge University Press, Cambridge.
Hocking, D. J. and Semlitsch, R. D. (2007). Effects of timber harvest on breeding-site

selection by gray treefrogs (Hyla versicolor). Biological Conservation, 138, 506–13.
Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiment.
Ecological Monographs, 54, 187–211.
Ireland, D. H., Wirsing, A. J., and Murray, D. L. (2007). Phenotypically plastic responses
of green frog embryos to conflicting predation risk. Oecologia, 152, 162–8.
Jaeger, R. G. and Walls, S. C. (1989). On salamander guilds and ecological methodology.
Kimball, K. D. and Levin, S. A. (1985). Limitations of laboratory bioassays: the need for
ecosystem-level testing. Bioscience, 35, 165–71.
Loman, J. (2004). Density regulation in tadpoles of Rana temporaria: a full pond field
experiment. Ecology, 85, 1611–18.
Mills, N. E. and Semlitsch, R. D. (2004). Competition and predation mediate indirect
effects of an insecticide on southern leopard frogs. Ecological Applications, 14, 1041–54.
Morin, P. J. (1981). Predatory salamanders reverse the outcome of competition among
three species of anuran tadpoles. Science, 212, 1284–6.
Morin, P. J. (1983). Predation, competition, and the composition of larval anuran guilds.
Parris, M. J., Semlitsch, R. D., and Sage, R. D. (1999). Experimental analysis of the
evolutionary potential of hybridization in leopard frogs (Anura: Ranidae). Journal of
Evolutionary Biology, 12, 662–71.
Relyea, R. A. (2004). Fine-tuned phenotypes: tadpole plasticity under 16 combinations
of predators and competitors. Ecology, 85, 172–9.
Relyea, R. A. and Auld, J. R. (2005). Predator- and competitor-induced plasticity: how
changes in foraging morphology affect phenotypic trade-offs. Ecology, 86, 1723–9.
Rowe, C. L. and Dunson, W. A. (1994). The value of simulated pond communities in
mesocosms for studies of amphibian ecology and ecotoxicology. Journal of Herpetology,
28, 346–56.
Rubbo, M. J., Belden, L. K., and Kiesecker, J. M. (2008). Differential responses of aquatic
consumers to variations in leaf-litter inputs. Hydrobiologia, 605, 37–44.
Rubbo, M. J. and Kiesecker, J. M. (2004). Leaf litter composition and community struc-
ture: translating regional species changes into local dynamics. Ecology, 85, 2519–25.
Schindler, D. W. (1998). Replication versus realism: the need for ecosystem-scale experi-
ments. Ecosystems, 1, 323–34.
Scott, D. E. (1990). Effects of larval density in Ambystoma opacum: an experiment in
large-scale field enclosures. Ecology, 71, 296–306.
Semlitsch, R. D. (1987). Relationship of pond drying to the reproductive success of the
salamander Ambystoma talpoideum. Copeia, 1987, 61–9.
Semlitsch, R. D. and Bridges, C. M. (2005). Amphibian ecotoxicology. In M. Lannoo
(ed.), Amphibian Declines: the Conservation Status of United States Species, pp. 241–3.
University of California Press, Berkeley, CA.
Semlitsch, R. D., Scott, D. E., Pechmann, J. H. K., and Gibbons, J. W. (1996). Structure
and dynamics of an amphibian community: evidence from a 16-year study of a nat-
ural pond. In M. L. Cody and J. A. Smallwood (eds), Long-term Studies of Vertebrate
Communities, pp. 217–48. Academic Press, San Diego, CA.
Skelly, D. K. (2002). Experimental venue and estimation of interaction strength. Ecology,

83, 2097–2101.
Skelly, D. K. and Kiesecker, J. M. (2001). Venue and outcome in ecological experiments:
manipulations of larval anurans. Oikos, 94, 198–208.
Vonesh, J. R. and Buck, J. C. (2007). Pesticide alters oviposition site selection in gray
treefrogs. Oecologia, 154, 219–26.
Wilbur, H. M. (1972). Competition, predation, and the structure of the Ambystoma-Rana
sylvatica community. Ecology, 53, 3–20.
Wilbur, H. M. (1987). Regulation in complex systems: experimental temporary pond
communities. Ecology, 68, 1437–52.
Wilbur, H. M. and Travis, J. (1984). An experimental approach to understanding pattern
in natural communities. In D. R. Strong, Jr, D. Simberloff, L. G. Abele, and A. B. Thistle
(eds), Ecological Communities: Conceptual Issues and the Evidence, pp. 113–22. Princeton
University Press, Princeton, NJ.
Williams, B. K., Rittenhouse, T. A. G., and Semlitsch, R. D. (2008). Leaf litter input medi-
ates tadpole performance across forest canopy treatments. Oecologia, 155, 377–86.
7
Water-quality criteria for amphibians
Donald W. Sparling
7.1 Introduction
Most amphibian species spend portions or even their entire life cycle in water.
Whether their habitats are streams, wetlands, vernal ponds, farm ponds, or
larger bodies of water, amphibians are directly affected by several natural and
anthropogenic chemical and physical characteristics. Dissolved oxygen, tem-
perature, pH, salinity and water conductivity, organic carbons, and pollutants
are important factors of their habitats that can affect survival, growth, mat-
uration, and physical development. In addition to direct effects, these char-
acteristics interact with other factors such as predators, prey, parasites, and
competitors to affect populations. Just as one example, the incidence of infest-
ation of tadpoles by Ribeioria ondatrae, a trematode parasite that causes limb
malformations, has been linked to the nutrient status of ponds (Johnson and
Chase 2004). Field studies of amphibians should at least include a description
of naturally occurring environmental conditions. In some cases, this descrip-
tion should also include analysis of environmental contaminants. This chapter
discusses chemical and physical factors that are most important to aquatic life
stages of amphibians and directs the reader to reviews for more complicated
issues dealing with water quality.
Because of space limitations, the descriptions of these factors and the methods
used in analyzing them are abbreviated. More complete descriptions of physio-
logical requirements can be found in Feder and Bruggren (1992) and McDiarmid
and Altig (1999). I recommend US Geological Survey (USGS) (2008a) and
Eaton et al. (2005) for standardized protocols of water-quality analysis.
7.2 Dissolved oxygen

Whereas free oxygen in the atmosphere occurs as O2 and is a relatively abun-
dant gas, it must be dissolved in water to be useful to aquatic aerobic organisms.
These organisms take in oxygen through moist membranes such as gill epithe-
lium, dermis, or other structures. Aside from for rare exceptions, only aquatic
life stages of amphibians potentially face oxygen problems; atmospheric oxy-
gen is sufficient for terrestrial forms. Some species of amphibian, such as tailed
frogs (Ascaphus spp.), live in swift-moving lotic environments where water is
constantly mixed with air and oxygen concentrations are usually near satur-
ation. Others, however, inhabit quiescent, warm water, lentic environments
where oxygen concentrations can fluctuate or persist at low concentrations.
Under laboratory conditions, oxygen concentrations below 4 mg/L are deemed
stressful to amphibian larvae and other aquatic organisms (ASTM 1988) and
prolonged exposure to such hypoxic conditions in the field can be considered
adverse. However, there are differences among species and animals can accli-
mate over time to low oxygen concentrations.
Many amphibians have anatomical features, behaviors, and physiological
processes that allow survival under temporary hypoxic conditions. For example,
members of the families Ranidae and Ambystomatidae assimilate oxygen
through dermal uptake, gills, and lungs. Under normoxic conditions (non-
stressful oxygen concentrations), approximately 70% of oxygen uptake by these
larvae is through the dermis, 20% via the gills, and 10% through the lungs
(Gatten et al. 1984). In contrast, Bufo spp. and Ascaphus spp. do not inflate their
lungs until just before metamorphosis (Feder 1984) and Plethondontids lack
lungs entirely.
In hypoxic conditions, larvae with lungs often swim to the surface and gulp
air into the lungs and pulmonary uptake of oxygen becomes more important.
Surfacing, however, incurs metabolic costs and exposes tadpoles to predators such
as turtles (Feder 1983). However, in an experiment using aquariums, McIntyre
and McCollum (2000) found that the alternative behavior of not coming to the
surface could also incur risks. Young, pre-lunged bullfrog (Rana catesbeiana)
tadpoles experienced a higher degree of predation in high-oxygen tanks than in
low-oxygen ones because predacious tiger salamanders (Ambystoma tigrinum)
spent a greater proportion of time searching for tadpoles at the bottom of the
tanks under high-oxygen conditions than when the dissolved oxygen was low.
Hypoxia can induce similar physiological responses in amphibians as in other
vertebrates including changes in blood pH, build-up of lactate in muscles, leth-
argy, and, under severe conditions, mortality (Ultsch et al. 1999). Chronic and
intermittent hypoxia can reduce growth and developmental rates and delay
hatching of salamander embryos (Valls and Mills 2007).
Several factors affect the concentration and measurement of dissolved oxygen.
Oxygen concentrations can vary widely through the course of a day, especially
7 Water-quality criteria for amphibians | 107
in warm, eutrophic bodies of water. In ponds, a major source of oxygen comes

from photosynthesis by algae and other green plants. Because photosynthesis
is driven by water temperature and sunlight, dissolved oxygen concentrations
are frequently lowest at dawn, increase several-fold during the course of the day,
reach maximum concentrations in mid afternoon, and decrease sharply during
the night due to biological oxygen demand (Figure 7.1). Cloudy or cool weather
may reduce the extremes in concentrations. These daily variations demonstrate
the need to record measurements under similar times of day and sunlight when
comparing oxygen concentrations among ponds or through time. Measurements
taken in early morning may be most meaningful with regard to environmental
stress.
In lotic conditions, especially in shallow headwater streams inhabited by
many species of salamander, dissolved oxygen concentrations are relatively uni-
form throughout the water column. In lentic bodies, however, there is often a
declining gradient from the surface, where mixing with air is maximal, to the
bottom, where anaerobic decay is greatest. Because larvae of different species
may occupy different water depths, oxygen concentrations should be taken at
least several centimeters below the surface or at the depth commonly used by
the species of concern.
The measurement of dissolved oxygen is relatively straightforward with one
of the many commercially available electronic dissolved-oxygen meters. For
15
12
Dissolved oxygen (mg/L)
0
12:00 4 8 12:00 4 8 12:00
am am am
Time of day
Fig. 7.1 Hypothetical variation in dissolved oxygen concentration in a warm, semi-eutrophic,

lentic body of water.
example, YSI Instruments, Orion, Accumet, and Oakton are a few of the com-
panies that manufacture portable field meters. These almost always come with
temperature sensors and may be bundled with conductivity and pH sensors.
Colorimeter kits (e.g. LaMotte, Orion, Chemetrics) are available and tend to be
less expensive than electronic meters but require reagents for every test and may
be less sensitive. When collecting water for oxygen analysis it is important to fill
plastic or glass bottles to the very top without adding air and to keep the bottle
cool until analysis. Ideally, oxygen should be analyzed on site to obtain the most
accurate values.
7.3 Temperature
Because amphibians are poikilotherms they have limited ability to regulate
their body temperature and are greatly affected by the temperature of their sur-
rounding environment. Water temperature, therefore, is extremely important in
affecting metabolic rates, other physiological processes, and behavior.
As a class, amphibians demonstrate a wide tolerance to temperature regi-
mens: some species are able to withstand “temperature changes of 30°C on a
daily basis and up to about 35°C on a seasonal basis” (Rome et al. 1992, p. 183).
Wood frogs (Rana sylvatica) have the most northern distribution of any North
American species (Conant and Collins 1998) and live above the Arctic Circle.
Couch’s spadefoot (Scaphiophus couchii) inhabits very arid, hot desert areas in
the west where temperatures may exceed 45°C. Some species are eurythermic;
for example, American bullfrogs range from central Mexico to northern Maine
and Nova Scotia and tiger salamanders can be found from southern Texas into
Alberta and Saskatchewan. Given time to acclimate, adult and larval amphib-
ians in temperate and northern climes can survive near and even subzero tem-
peratures by becoming dormant and greatly reducing their metabolic rates
(Voituron et al. 2003). Those that inhabit dry, hot regions often burrow into the
ground, enmesh themselves in cocoons, and estivate (Pinder et al. 1992). Some
species, such as northern leopard frogs (Rana pipiens), can shunt water from vital
organs to prevent tissue damage from ice crystal formation and concentrate glu-
cose in organs to serve as a type of anti-freeze (Tattersall and Ultsch 2008).
In general, between 10 and 40°C, each 10° increase in ambient tempera-
ture increases metabolism by 1.4–2.4 times (Rome et al. 1992). Higher meta-
bolic rates require greater oxygen; however, oxygen concentrations decrease
as water temperatures increase (see below). At temperatures near 0°C most
amphibians are very sluggish. At high temperatures, say above 30–35°C for
some species but as low as 25–30°C for less tolerant animals, thermal stress can
result in reduced mobility, abnormally high heart rates, and eventually death.
Subtle interactions between temperature and other stressors can also occur. For
example, Anderson et al. (2001) demonstrated complex relationships between
predators and temperature in Pacific treefrogs (Pseudacris regilla). At 25.7°C
tadpoles grew faster and exceeded the size limitation of their insect predators
more quickly than tadpoles raised at 9.9°C. However, tadpoles of similar size
encountered greater predation rates at the warmer temperature. Temperature
can have a strong influence on life history progress (Camp and Marshall 2000)
and microhabitat use (Crawford and Semlitsch 2008) in both aquatic and ter-
restrial amphibians.
There is a direct relationship between water temperature and dissolved oxygen
concentrations. At sea level, where the partial pressure of oxygen, PO, is highest,
the concentration of dissolved oxygen is around 14 ppm at 0°C and declines as
water temperatures increase (Figure 7.2). The same general relationship exists
at higher elevations except that PO saturation concentrations decline. Because
oxygen demand increases as water increases in temperature, physiological stress
can be very high in warm, low oxygenated waters.
Measuring water temperature is straightforward with any digital or analog
thermometer. Many electronic meters used for other water chemistry parameters
such as pH, conductivity, or dissolved oxygen come equipped with thermom-
eters. Some thermometers automatically record minimum and maximum tem-
peratures between readings so that a range can be ascertained. More sophisticated
systems include data loggers that keep a continuous record of temperatures.
15
Dissolved oxygen (mg/L)
13
11
5
0 5 10 15 20 25 30
Temperature (°C)
Fig. 7.2 Relationship between water temperature and dissolved oxygen concentration
at sea level.
7.4 pH
pH, the negative log of the hydrogen ion concentration in water, is another key
characteristic in describing amphibian habitats. The pH scale ranges from 0 to 14
which corresponds to a solution of 1 M H (100) to 1 M OH with all H bound
with oxygen as a hydroxide. A pH of 7 defines neutral solutions, those with lower
pH values are acidic, and those with higher values are alkaline. For ecological
purposes, pH ranges of 6.0–7.5 are generally considered to be circumneutral or
within a range that should present no harm to most aquatic organisms. A corre-
sponding term, alkalinity, is the ability of water to buffer changes in pH.
A key factor in determining pH is the type of bedrock found within the
watershed. Sedimentary rock such as sandstone, limestone, and dolomite are
high in carbonates and bicarbonates that bind with hydrogen ions and form
natural buffers. Alkalinity is typically high in these waters. Watersheds with a
bedrock of igneous rock such as granite or basalt typically have low alkalinity
due to the absence of carbonates.
Another natural factor affecting pH is the concentration of organic acids
including tannic, fulvic, and humic acids. These organic compounds are weak
acids in that they have low dissociation constants compared to mineral acids.
Organic acids can be important in maintaining a low pH in isolated water bod-
ies with high organic matter such as fens and bogs, but they also serve to some
degree as buffers against anthropogenic sources of acidity.
Two major sources of anthropogenic acidity are acid deposition and acid
mine drainage. Acid deposition comes from the production of nitrous oxides
(NOx) and sulfates (SO42) during the combustion of fossil fuels. These mol-
ecules combine with hydrogen in the air and are deposited on to land and water
either with rain or attached to dust and organic particles that precipitate in
dry form. Once in water the hydrogen and anions dissociate and free hydrogen
ions increase acidity. Acid deposition can be of major concern in wetlands and
streams that lie downwind of major industrial centers or cities and in water-
sheds with low alkalinity. The US Environmental Protection Agency (USEPA)
(2008) is a good reference for current aspects of acid rain and deposition. Acid
mine drainage is most prevalent in unreclaimed mined areas. Inversion of soil
during mining places layers high in sulfides towards the surface and oxidation
of these soils results in acidic seeps and leaching. The USGS (2008b) provides a
good review of acid mine drainage.
Many review articles have been written on the adverse effects of acid depos-
ition and anthropogenic acidity on amphibians (e.g. Sparling 1995; Rowe
and Freda 2000). Variation in sensitivity to low pH occurs at the life stages,
population, and species levels. At pH values equal to or below 4.5, embryonic

development may cease entirely. At somewhat higher pH values (≈4.5–5.0)
development continues but hatching is curtailed (Dunson and Connell 1982).
Interspecific differences to low pH have been found in several studies (e.g.
Gosner and Black 1957; Sadinski and Dunson 1992). The critical pH or that
which can cause significant increases in mortality for amphibian embryos
ranges from 5.0 to 3.5 (Freda 1990). Tolerant species include carpenter’s frogs
(Rana virgatipes), wood frogs, and pine barren’s treefrog (Hyla andersoni).
Tolerance to low pH may be acquired in that animals living in acidic envi-
ronments for many generations may have higher tolerance to reduced pH
than those inhabiting circumneutral waters (Andren et al. 1989). The pri-
mary mechanism of toxicity due to low pH is interference with ion transport.
Other effects include compromised immune systems (Brodkin et al. 2003), an
inability of embryos to hatch, reduced growth, and delayed metamorphosis.
Acidic waters can be treated to increase pH. In watersheds with low alkalinity
the addition of limestone to streams can substantially increase pH. Hard water with
high calcium and magnesium concentrations can affect the toxic effects of acidity
on embryos and larvae. Calcium can increase amphibian tolerance to low pH by
reducing the loss of plasma ions. For example, adding 10 mg/L dissolved calcium
to water increased hatching of Jefferson salamander (Ambystoma jeffersonianum)
embryos at pH 4.5 compared with 1 mg/L (Freda and Dunson 1985). Conversely,
calcium reduced hatching in wood frogs at pH 4.5. Freda (1990) explained that the
difference in response to calcium was related to a curling effect seen in embryos
that is elicited by H and by calcium. Embryos that can hatch even though curled
are benefited by calcium whereas those that cannot hatch when curled suffer.
Metals are more soluble with reduced pH and may interact with acidity to
influence toxicity. Aluminum is particularly of concern because it is the most
common metal in the Earth’s crust, is soluble with low pH, and changes its spe-
ciation form as pH varies (reviewed by Sparling and Lowe 1996). Aluminum
interacts with acidity in complex ways (e.g. Skei and Dolmen 2006) and a com-
plete review of this interaction is beyond the scope of this chapter.
A common way of measuring pH is through one of the many commercially
available pH meters. A typical meter will accurately measure pH to 0.01 pH
units. If budgets are limited, however, pH papers with different ranges can be
purchased and they are accurate to about 0.2 pH units. As with other water-
quality measurements, it is usually preferable to assess water several centimeters
below the surface. Sealed probes can be submerged to the desired depth or water
can be collected in narrow-mouthed bottles for open or pH papers. The meas-
urement of pH is temperature sensitive and most meters are set to 25°C for
optimal accuracy. For field measurements, therefore, I recommend meters with

automatic temperature compensation (ATC); these simultaneously record tem-
perature and adjust the pH values accordingly.
7.5 Conductivity, hardness, and salinity

Conductivity is the ability of water to carry an electrical current and is due to
the total concentration of anions and cations in water, collectively called total
dissolved solids (TDS) in water. It is measured in units of micro-Siemens per
centimeter or μS/cm. For comparison, ultrapure water has a conductivity of
5.5 103 μS/cm, drinking water around 5–50 μS/cm and sea water around
5000 μS/cm. Waters that have conductivity ranging between 150 and 500 μS/cm
are adequate for aquatic organisms (USEPA; www.epa.gov/volunteer/stream/
vms59.html).
Hardness can be measured as either the concentration (mg/L) of calcium ions
or the sum of calcium and magnesium ions in water. These cations help buffer
the effects of acidity and alter the toxic effects of dissolved metals (Horne and
Dunson 1995b). Hard water is traditionally defined as a calcium-equivalent
concentration of 80–120 mg/L and soft water as 0–20 mg/L. Moderate to hard
water tends to be better for amphibians than soft water because calcium and
magnesium can ameliorate the toxic effects of metals and pH, improve osmo-
regulation, and are required for proper formation of bones and other physio-
logical processes (Ultsch et al. 1999).
Salinity is the concentration of chloride salts in water and is usually measured
in parts per thousand (g/L or ‰), molarity (mM), or osmoles (mOsm). Most
amphibians have low tolerance to salinity because they are not good osmoregula-
tors (Gomez-Mestre et al. 2004). Balinksy (1981) listed 64 species of anurans and
13 species of urodeles that are tolerant to brackish water, although none are truly
marine. Among the most tolerant species are cane toads (Bufo marinus), crab-
eating frog (Rana cancrivora), African clawed frog (Xenopus laevis), European
green toad (Bufo viridis), and some Batrachoseps spp. salamanders (Shoemaker et al.
1992). Euryhaline species can live in water that is 600–800 mOsm (70–93 ppm
NaCl) but adults of most others do best at 200–300 mOsm (23–34 ppm NaCl;
Shoemaker et al. 1992). Relatively little data are available on tadpoles but, in gen-
eral, the lower the salinity the better. Road salts may pose problems for amphib-
ians inhabiting ponds located near roads (Karraker et al. 2008).
As with other water-quality factors mentioned above, there are electronic
meters and colorimeter kits available to measure these parameters.
7.6 Total and dissolved organic carbon

Organic carbon comes from organic molecules as contrasted to inorganic car-
bon that is derived from carbon dioxide and carbonates. Total organic car-
bon includes all sizes of organic molecules, both dissolved and non-dissolved.
Dissolved organic carbon consists of organic molecules smaller than 0.45 μm
in diameter. This carbon comes from many sources including surface run-off,
aerial deposition, and decomposition of aquatic organisms. Elevated concen-
trations of organic carbon, along with suspended soil particles, are principal
constituents of total suspended solids (TSS) and indicate potential problems
with run-off into streams and lakes. Organic carbons are used as nutrients by
bacteria and can increase biological oxygen demand. However, moderate levels
of organic carbon are important because these molecules are a source of food
for microorganisms. They also bind with certain pollutants such as metals and
reduce their availability to aquatic organisms.
Dissolved organic carbon (DOC) is ecologically important to amphibians.
Banks et al. (2007) showed a positive relationship between DOC concentration
in wetlands of Acadia National Park (range 2–18 mg/L) and the concentration of
methylated mercury (MeHg) in sediments and northern green frog (Rana clami-
tans) and bullfrog tadpoles. Because MeHg is more bioavailable and toxic than
inorganic mercury, DOC in this situation can be detrimental to amphibians. In
contrast, some forms of DOC, such as tannins, cause a brownish coloration to
the water. These DOCs reduce light penetration into water which may curtail
primary productivity (Carpenter et al. 2001). They also significantly suppress
penetration of ultraviolet-B radiation which has been linked to malformations.
Diamond et al. (2005) determined that DOC reduced ultraviolet-B penetration
in the top 1 cm of water up to 87% and Calfee et al. (2006) determined that DOC
protected salamander embryos from harmful ultraviolet-B exposure. Freda (1991)
demonstrated that DOC can mitigate the toxicity of aluminum and other metals
in acidified waters. However, humic and fulvic acids in excess of 10–20 mg/L can
be toxic to amphibians (Freda et al. 1989; Horne and Dunson 1995a).
The analysis of TOC and DOC is a bit more complex than the other ana-
lytes so far described. Briefly, the first step is to separate inorganic and organic
carbons in the water. This can be done by acidifying the water, thus converting
carbonates to carbon dioxide which is then sparged from the water. The remain-
ing carbon is considered organic. An aliquot of this is filtered through a 0.45 μm
membrane and the carbon in the filtrate is DOC. Both TOC and DOC can be
measured on a carbon analyzer or other instruments.
Colored DOCs (tannic, humic, and fulvic acids) can be measured indir-
ectly by comparing the color of water to a set of standards available in kit form.
Turbidity, which is affected by the total amount of inorganic and organic sus-
pended particles, can be measured with a turibidimeter that measures light
penetration through a standardized vial in units called nephelometric turbidity
units (NTUs).
7.7 Pollutants
The number of types of pollutants or contaminants that can affect amphibian
populations is enormous and beyond the scope of this brief chapter. Sparling
(2003) surveyed the potential effects of contaminants on amphibians and
Sparling et al. (2000) provided an extensive review of contaminant effects on
amphibians and reptiles. Here I list a brief summary of contaminant classes and
their significance.
7.7.1 Fertilizers and nitrogenous compounds

There are several ways of measuring ammonia and nitrates in water, includ-
ing: (1) ion chromatography (for nitrates); (2) ion-specific probes coupled with
a pH meter that registers millivolt output; (3) colorimetry; (4) test papers;
and (5) other methods for sewage treatment operations and large volumes of
material. Of the first four methods, ion-specific probes are less expensive than
purchasing an ion chromatograph and are more sensitive and accurate than col-
orimetry or test papers. Test papers, although least expensive, are more limited
in their resolution and detection limits than the other methods.
7.7.2 Pesticides
Pesticides are chemicals that are used to control plant and animal species con-
sidered to be noxious or unwanted by humans. The types of pesticides can be
distinguished by intended target and by chemical class, which usually corres-
ponds to primary mode of action. By target the primary classes are insecticides,
fungicides, herbicides, and rodenticides.
Very roughly and with exceptions, insecticides are most acutely toxic to
amphibians and other aquatic organisms of these classes. However, insecti-
cides, fungicides, and herbicides exert a wide variety of sublethal or chronic
effects that affect individual and population health. Rodenticides are not of
much concern to amphibians, simply because of how they are used. There is a
large and growing number of chemical families used as pesticides (see Cowman
and Mazanti 2000 for a more complete review). Of greatest concern due to
their widespread usage are organophosphates, carbamates, pyrethrins, and a few

organochlorine insecticides. These are neurotoxins and either function at the
synapse between neurons (organophospates and carbamates) or by affecting the
transmission of action potentials down axons. In addition to direct mortality,
pesticides can reduce growth, inhibit development, affect escape and predatory
behavior, cause genotoxicity, affect gonad development, and alter other physio-
logical processes (see Cowman and Mazanti 2000; Sparling 2003) and inter-
act with other stressors (e.g. McIntyre and McCollum 2000; Relyea and Mills
2001). Aerial spraying of pesticides results in airborne particles that can travel
many kilometers from the point of application to affect remote populations of
amphibians (Davidson et al. 2001; Sparling et al. 2001).
There is no field method to determine the presence or concentration of pesti-
cides in that their analyses are complex and often expensive. Matrix samples
(sediment, water, or organisms) need to be collected and brought to a laboratory
where the pesticides are extracted and analyzed using high-performance liquid
chromatography (HPLC), gas chromatography, or other methods, all following
recognized protocols.
7.7.3 Metals
Metals and metalloids such as zinc, copper, lead, chromium, arsenic, cadmium,
mercury, selenium, and others are naturally occurring elements but are also
released through many different industrial processes at concentrations that can
be toxic to amphibians. Linder and Grillitsch (2000) reviewed the ecotoxicol-
ogy of metals on amphibians and reptiles and showed that effects can range
from direct mortality to a host of sublethal and potentially debilitating effects.
Toxicity is affected by pH in that metal solubility increases with acidity and sol-
uble forms of metals are more bioavailable than non-soluble forms.
Ion-specific probes are available for lead, silver, and copper, but the method
for measuring is more complicated than inserting the probe directly into a water
sample and should be done under laboratory conditions. Fortunately, metals are
very stable, and once samples have been acidified to pH
2 they can be stored
for several months before analysis. Colorimetric methods exist for iron, man-
ganese, aluminum, molybdenum, and copper, but most metals are analyzed
through atomic absorption spectrophotometry or ion-coupled spectrophotom-
etry (ICP) in a laboratory.
7.7.4 Organic pollutants and halogenated hydrocarbons

For this chapter, I include polycyclic aromatic hydrocarbons (PAHs), poly-
chlorinated biphenyls (PCBs), dioxins, furans, and chlorinated pesticides
(e.g. DDT, toxaphene, dieldrin) in this category. All of these chemicals are
organically based and some have chlorine or other halogens incorporated into
the molecular structure. PAHs include benzene, toluene, benzo[a]pyrene, and
hundreds of other molecules that form through natural processes of decompos-
ition, but also come from combustion of fossil fuels, processing of petroleum,
and other sources. PCBs were used as lubricating fluids and in electrical trans-
formers until the 1970s when their use in North America was banned. Dioxins
and furans are formed during forest fires but also through the combustion
of fossil fuels and are strong carcinogens. Similarly, many of the chlorinated
hydrocarbon pesticides have been banned because they are extremely persist-
ent in the environment and cause cancer, genotoxicity, endocrine disruption,
and other harmful effects to humans and wildlife. Nevertheless, because they
are extremely persistent, they can still be found throughout the world. At most
environmentally realistic concentrations these chemicals produce a variety of
sublethal effects including genotoxicity, inhibition of growth and develop-
ment, endocrine disruption, skeletal defects and malformations, and reduced
hatching success (reviewed by Sparling 2000). Sophisticated laboratory methods
including HPLC or gas chromatography are required for the analysis of these
chemicals.
7.7.5 Pharmaceuticals
Ecotoxicologists are just becoming aware of a myriad of chemicals that pass
through our waste-treatment plants and enter natural bodies of waters essen-
tially intact. These include prescription and non-prescription medicines, anti-
biotics, caffeine, and other products that pass through human bodies or are
disposed of into toilets and down drains. These chemicals can exert numerous
effects such as endocrine disruption and many physiological and behavioral
changes that have yet to be defined. Because our awareness of these substances is
so new, methods for determining their concentrations in field-collected matrices
are often not available. Others can be determined through gas chromatography/
mass spectrophotometry.
7.8 Summary and conclusions

In summary, students of amphibian ecology and behavior must be aware of
various chemical and physical factors. These, in addition to vegetation, sub-
strate composition, structural elements in the environment, water depth, stream
flow, and the presence of other vertebrate and invertebrate species constitute the
basic elements of habitat assessment for amphibians. There are many different
methods of measuring these factors and the choice of method should be based
on desired accuracy and precision as well as economics.
7.9 References
Anderson, M. T., Kiesecker, J. M., Chivers, D. P., and Blaustein, A. R. (2001). The direct
and indirect effects of temperature on a predator-prey relationship. Canadian Journal
of Zoology, 79, 1834–41.
Andren, C., Henrikson, L. Olsson, M., and Nilson, G. (1989). Effects of pH and alu-
minium on embryonic and early larval stages of Swedish brown frogs Rana arvalis, R.
temporaria and R. dalmatina. Holoarctic Ecology, 11, 127–35.
ASTM (1988). Standard Practice for Conducting Acute Toxicity Tests with Fishes,
Macroinvertebrates, and Amphibians. ASTM, West Conshohocken, PA.
Balinsky, J. B. (1981). Adaptation of nitrogen metabolism to hyperosmotic environment
in Amphibia. Journal of Experimental Zoology, 215, 335–50.
Banks, M. S., Crocker, J., Connery, B., and Amirbahman, A. (2007). Mercury bio-
accumulation in green frog (Rana clamitans) from Acadia National Park, Maine, USA.
Environmental Toxicology and Chemistry, 26, 118–25.
Brodkin, M., Vatnick, I., Simon, M., Hopey, H., Butler-Holston, K., and Leonard, M.
(2003). Effects of acid stress in adult Rana pipiens. Journal of Experimental Zoology,
298A, 16–22.
Calfee, R. D., Bridges, C. M., and Little, E. E. (2006). Sensitivity of two salamander
(Ambystoma) species to ultraviolet radiation. Journal of Herpetology, 40, 35–42.
Camp, C. D. and Marshall, J. L. (2000). The role of thermal environment in determining
the life history of a terrestrial salamander. Canadian Journal of Zoology, 78, 1702–11.
Carpenter, S. R., Cole, J. J., Hodgon, J. R., Kitchell, J. F., Pace, M. L., Bade, D.,
Cottingham, K. L., Essington, T. E., Houser, J. N., and Schindler, D. E. (2001).
Trophic cascades, nutrients, and lake productivity: whole-lake experiments. Ecological
Monographs, 71, 163–86.
Conant R. and Collins, J. T. (1998). Reptiles and Amphibians. East/Central North America.
Peterson Field Guides. Houghton-Mifflin, Boston, MA.
Cowman, D. F. and Mazanti, L. E. (2000). Ecotoxicology of “New Generation” pesticides
to amphibians. In D. W. Sparling, G. Linder, and C. A. Bishop (eds), Ecotoxicology of
Amphibians and Reptiles, pp. 233–68. SETAC Press, Pensacola, FL.
Crawford, J. A. and Semlitsch, R. D. (2008). Abiotic factors influencing abundance and
microhabitat use of stream salamanders in southern Appalachian forests. Forest Ecology
and Management, 255, 1841–7.
Davidson, C., Shaffer, H. B., and Jennings, M. R. (2001). Declines of the California red-
legged frog: climate, UV-B, habitat, and pesticides hypotheses. Ecological Applications,
11, 464–79.
Diamond, S. A., Trenham, P. C., Adams, M. J., Hossack, B. R., Knapp, R. A., Stark, S. L.,
Bradford, D., Corn, C. S., Czarnowski, K., Brooks, P. D. et al. (2005). Estimated ultra-
violet radiation doses in wetlands in six national parks. Ecosystems, 8, 462–77.
Dunson, W. A. and Connell, J. (1982). Specific inhibition of hatching in amphibian
embryos by low pH. Journal of Herpetology, 16, 314–16.
Eaton, A. D., Clesceri, L. S., Rice, E. W., and Greenberg, A. E. (2005). Standard Methods
for the Examination of Water and Waste Water, 21st edn. American Public Health
Association Press, Washington DC.
Feder, M. E. (1983). The relation of air breathing and locomotion to predation on tad-
poles, Rana berlandieri, by turtles. Physiological Zoology, 56, 522–31.
Feder, M. E. (1984). Consequences of aerial respiration for amphibian larvae. In
R. S. Seymour (ed.), Respiration and Metabolism of Embryonic Vertebrates, pp.
71–86. Junk, Dordrecht.
Feder, M. E. and Burggren, W. W. (1992). Environmental Physiology of the Amphibians.
Freda, J. (1990). Effects of acidification on amphibians. In J. P. Baker (ed.), State of Science
and State of Technology, SOS/T report no. 13, Biological Effects of Changes in Surface
Water Acid-Base Chemistry, pp. 2-185–2-202. National Acid Precipitation Assessment
Program, Washington DC.
Freda, J. (1991). The effects of aluminum and other metals on amphibians. Environmental
Pollution, 71, 305–28.
Freda, J. and Dunson, W. A. (1985). The influence of external cation concentration on
the hatching of amphibian embryos in water of low pH. Canadian Journal of Zoology,
63, 2649–56.
Freda, J., Cavdek, V., and McDonald, D. G. (1989). Role of organic complexation in the
toxicity of aluminum to Rana pipiens embryos and Bufo americanus tadpoles. Canadian
Journal of Fisheries and Aquatic Sciences, 47, 217–24.
Gatten, Jr, R. E., Caldwell, J. P., and Stockard, M. E. (1984). Anaerobic metabolism dur-
ing intense swimming by anural larvae. Herpetologica, 40, 164–9.
Gomez-Mestre, I., Tejedo, M. R., and Estepa, J. (2004). Developmental alterations and
osmoregulatory physiology of a larval anuran under osmotic stress. Physiological and
Biochemical Zoology, 77, 267–74.
Gosner, K. L. and Black, J. H. (1957). The effects of acidity on the development and
hatching of New Jersey frogs. Ecology, 38, 256–62.
Horne, M. T. and Dunson, W. A. (1995a). The interactive effects of low pH, toxic metals,
and DOC on a simulated temporary pond community. Environmental Pollution, 89,
155–61.
Horne, M. T. and Dunson, W. A. (1995b). Effects of low pH, metals, and water hardness on
larval amphibians. Archives of Environmental Contamination and Toxicology, 29, 500–5.
Johnson, P. T. J. and Chase, J. M. (2004). Parasites in the food web: linking amphibian
malformations and aquatic eutrophication. Ecology Letters, 7, 521–6.
Karraker, N. E., Gibbs, J. P., and Vonesh, J. R. (2008). Impacts of road deicing salt on the
demography of vernal pool-breeding amphibians. Ecological Applications, 18, 724–34.
Linder, G. and Grillitsch, B. (2000). Ecotoxicology of metals. In D. W. Sparling,
G. Linder, and C. A. Bishop (eds), Ecotoxicology of Amphibians and Reptiles,
pp. 325–408. SETAC Press, Pensacola, FL.
McDiarmid, R. W. and Altig, R. (1999). Tadpoles: The Biology of Anuran Larvae.
McIntyre, P. B. and McCollum, S. A. (2000). Responses of bullfrog tadpoles to hypoxia
and predators. Oecologia, 125, 301–8.
Ortiz-Santaliestra, M. E., Marco, A., Fernandez, M. J., and Lizana, M. (2006). Influence
of developmental stage on sensitivity to ammonium nitrate of aquatic stages of amphib-
ians. Environmental Toxicology and Chemistry, 25, 105–11.
Pinder, A. W., Storey, K. B., and Ultsch, G. R. (1992). Estivation and hibernation. In
M. E. Feder and W. W. Burggren (eds), Environmental Physiology of the Amphibians,
Relyea, R. A. and Mills, N. (2001). Predator-induced stress makes the pesticide carbaryl
more deadly to gray treefrog tadpoles (Hyla versicolor). Proceedings of the National
Academy of Sciences USA, 98, 2491–6.
Rome, L. C., Stevens, E. D., and John-Alder, H. B. (1992). The influence of temperature
and thermal acclimation on physiological function. In M. E. Feder and W. W. Burggren
(eds), Environmental Physiology of the Amphibians, pp. 183–205. University of Chicago
Press, Chicago, IL.
Rowe, C. L. and Freda, J. (2000). Effects of acidification on amphibians at multiple levels
of organization. In D. W. Sparling, G. Linder, and C. A. Bishop (eds), Ecotoxicology of
Amphibians and Reptiles, pp. 545–72. SETAC Press, Pensacola, FL.
Sadinski, W. J. and Dunson, W. A. (1992). A multilevel study of effects of low pH on
amphibians of temporary ponds. Journal of Herpetology, 26, 413–22.
Shoemaker, V. H., Hillman, S. S., Hillyard, S. D., Jackson, D. C., McClanahan, L. L.,
Withers, P. C., and Wygoda, M. L. (1992). Exchange of water, ions, and respiratory
gases in terrestrial amphibians. In M. E. Feder and W. W. Burggren (eds), Environmental
Physiology of the Amphibians, pp. 125–50. University of Chicago Press, Chicago, IL.
Skei, J. K. and Dolmen, D. (2006). Effects of pH, aluminum, and soft water on larvae of the
amphibians Bufo bufo and Triturus vulgaris. Canadian Journal of Zoology, 84, 1668–77.
Sparling, D. W. (1995). Acidic deposition: a review of biological effects. In D. Hoffman,
B. A. Rattner, G. A. Burton, and J. Cairns Jr (eds), Handbook of Ecotoxicology,
pp. 301–29. Lewis Publishers, Boca Raton, FL.
Sparling, D. W. (2000). Ecotoxicology of organic contaminants to amphibians. In
D. W. Sparling, G. Linder, and C. A. Bishop (eds), Ecotoxicology of Amphibians and
Reptiles, pp. 461–94. SETAC Press, Pensacola, FL.
Sparling, D. W. (2003). A review of the role of contaminants in amphibian declines.
In D. Hoffman, B. A. Rattner Jr, and J. Cairns (eds), Handbook of Ecotoxicology,
pp. 1099–1128. Lewis Publishers, Boca Raton, FL.
Sparling, D. W. and Lowe, T. P. (1996). Environmental hazards of aluminum to plants, inver-
tebrates, fish, and wildlife. Reviews Environmental Contamination Toxicology, 145, 1–127.
Sparling, D. W., Linder, G., and Bishop, C. A. (2000). Ecotoxicology of Amphibians and
Reptiles. SETAC Press, Pensacola, FL.
Sparling, D. W., Fellers, G. M., and McConnell, L. L. (2001). Pesticides and amphibian
population declines in California, USA. Environmental Toxicology and Chemistry, 20,
1591–5.
Tattersall, G. J. and Ultsch, G. R. (2008). Physiological ecology of aquatic overwintering
in ranid frogs. Biological Review, 83, 119–40.
Ultsch, G. R., Bradford, D. F., and Freda, J. (1999). Physiology: coping with the environ-
ment. In R. W. McDiarmid and R. Altig (eds), Tadpoles. The Biology of Anuran Larvae,
USEPA (US Environmental Protection Agency) (2008). Acid Rain. www.epa.gov/

acidrain/.
USGS (United States Geological Survey) (2008a). National field manual for the collection of
water-quality data: U.S. Geological Survey Techniques of Water-Resources Investigations,
book 9, chapters A1–A9. http://pubs.water.usgs.gov/twri9A.
USGS (United States Geological Survey) (2008b). Acid mine drainage: introduction.
http://energy.er.usgs.gov/health_environment/acid_mine_drainage/.
Valls, J. H. and Mills, N. E. (2007). Intermittent hypoxia in eggs of Ambystoma macu-
latum: embryonic development and egg capsule conductance. Journal of Experimental
Biology, 210, 2430–5.
Voituron, Y., Eugene, M., and Barré, H. (2003). Survival and metabolic responses to
freezing by the water frog (Rana ridibunda). Journal of Experimental Zoology, 299A,
118–26.
Part 3
Juveniles and adults
8
Measuring and marking
post-metamorphic amphibians
John W. Ferner
8.1 Introduction
In this chapter, I review marking and measuring techniques for post-metamorphic
amphibians; larvae are covered in Chapter 4 in this volume. Measuring the size
of individuals is important for determining age, categorizing life history stages
(juvenile or adult), and calculating growth rates (larval measurements are covered
in Chapter 3). More field and behavioral studies requiring marking are now being
conducted than in the past, and often investigators need very specific marking
techniques and non-invasive ways of identifying individuals. Additionally, ethical
issues relative to certain techniques, such as radioactive tags and mutilation pro-
cedures, are now more commonly considered as, for example, when considering
toe-clipping. Some older techniques are used much less often than they used to
be. Major advancements in biotelemetry (see Chapter 11) and passive integrated
transponder (PIT) tags have added needed flexibility to the choice of marking
methods.
Recent reviews of marking and identification techniques include Baker and
Gent (1998) and Ferner (2007) for amphibians and reptiles, and Donnelly et al.
(1994) for amphibians. Additional references are found in a number of papers
published in a special edition of Mertensiella (Henle and Veith 1997).
Some criteria for ideal marks or tags are as follows (after Ferner 2007):
• they should not affect the survivorship or behavior of the organism;
• they should allow the animal to be as free from stress or pain as possible;
• they should identify the animal as a particular individual, if desirable;
• they should last indefinitely, or at least the duration of the study;
• they should be easily read and/or observable;
• they should be adaptable to organisms of different sizes;

• they should be easy to use in both laboratory and field, and use easily
obtained material at minimal cost.
There are no techniques that satisfy all of these criteria. Selecting a technique
requires deciding which are the most feasible and scientifically accurate in meet-
ing the objectives of the particular study. Two or more marks often are used to
meet multiple study objectives or to serve as a back-up to the primary mark (e.g.
simultaneous toe-clipping and photographing of dorsal patterns). Techniques
Table 8.1 Sources of materials for marking amphibian species (as reviewed in Ferner
2007 unless otherwise indicated).
Technique item Source of materials
Adhesive (small wounds) Vetbond®

®
Bandage (New Skin ) Medtech (www.newskinproducts.com)
Beads, balls, and elastic Ball Chain Manufacturing Co. (www.ballchain.com)
for waistbands Bead Industries (www.beadindustries.com)
B. Toucan (www.btoucan.com)
Fluorescent pigments for Scientific Marking Materials, PO Box 23122,
branding Seattle, WA 98124, USA
Radiant Color Company, 2800 Radiant Ave.,
Richmond, CA 94804, USA (+1 800 Radiant;
fax +1 510 233 9138)
Fluorescent elastomer dyes; Northwest Marine Technology, PO Box 427, 976
visual implant elastomers (VIEs) Ben Nevis Road, Shaw Island, WA 98286, USA
(www.nmt.us)
PIT tags Destron Fearing Corporation/Digital Angel,
St. Paul, MN, USA (www.destronfearing.com,
www.digitalangel.com)
Also see www.biomark.com
PIT pack Battery powered Destron Fearing transceiver
(model FS 2001 A-ISO) and custom-built
antenna (Blomquist et al. 2008)
Scanner (for PIT tags) AVID Marketing, Norco, CA USA
Fluorescent tags (VIAlpha Northwest Marine Technology
Numeric Tags) (see address above)
Ultraviolet lamp for UVP, San Gabriel, CA, USA (www.uvp.com)
detecting fluorescents
PIT, passive integrated transponder.
8 Working with post-metamorphic amphibians | 125
described for use with one group of amphibians might be adapted for a dif-
ferent group. Therefore, investigators might benefit by considering all of the
techniques available, regardless of how originally used. Further, there may be a
need for standardizing marking techniques in comparative studies, or those that
might be carried on by other researchers in the future.
In the following text, I emphasize the use of marking techniques for adult
amphibians which have proved most successful. When available, information
on the advantages and disadvantages of each technique has been given. This
chapter provides information adequate for considering techniques without
consulting the original source. However, once a technique has been tentatively
selected, it is advisable to consult the primary literature if time and facilities
permit. The use of complex techniques, such as telemetry or PIT tags, requires
careful review of the original sources. Sources for materials used with various
techniques are given in Table 8.1, although availability and contact information
are likely to change constantly.
8.2 Toe-clipping
8.2.1 Anurans
Although several toe-clipping methods are proposed in the literature, that of
Martof (1953) is most widely used. He cut toes of Rana clamitans with scissors,
trying to avoid damage to webbing between them. Little blood loss or loss of
swimming power was observed and no regeneration of the digits was found.
This system assigns serial numbers to the digits of the feet. No more than two
toes were removed from any one foot. The left hind foot indicated units, and
for those larger than five, combinations of two digits (the fifth and one other)
were excised. For example, clipping the fifth and third toes designated tens and
the front feet denoted hundreds. With this arrangement, 6399 individuals can
be marked in series; the greatest number can be indicated by removing the two
outer digits on each foot. The only concern Martof expressed about this pro-
cedure was that confusion might arise where only a single toe on a single hand
foot was removed with recapture of a frog with an injured foot. He suggested
alleviating this problem with a special designation, such as a zero marking by
removal of the second and fourth toes on the other hind foot. Waichman (1992)
also proposed a numbering system for toe-clipping that letters the four limbs
(A through D) and numbers the toes. This scheme makes all 959 combinations
available using two and three toes, but not removing more than two toes for each
limb (see Table 8.2).
Table 8.2 Alphanumeric code for toe-clipping amphibians and reptiles from Waichman (1992)
One toe Two toes A2A5 A3D2 A5B5 B1D4 B4C1
A2B1 A3D3 A5C1 B1D5 B4C2
A2B2 A3D4 A5C2 B2B3 B4C3
A1 A1A2 A2B3 A3D5 A5C3 B2B3 B4C4
A2 A1A3 A2B4 A4A5 A5C4 B2B4 B4C5
A3 A1A4 A2B5 A4B1 A5C5 B2B5 B4D1
A4 A1A5 A2C1 A4B2 A5D1 B2D1 B4D2
A5 A1B1 A2C2 A4B3 A5D2 B2D2 B4D3
B1 A1B2 A2C3 A4B4 A5D3 B2D3 B4D4
B2 A1B3 A2C4 A4B5 A5D3 B2D4 B4D5
B3 A1B4 A2C5 A4C1 A5D4 B2D5 B5C1
B4 A1B5 A2D1 A4C2 A5D5 B3B4 B5C2
B5 A1C1 A2D2 A4C3 B1B2 B3B5 B5C2
C1 A1C2 A2D3 A4C4 B1B3 B3C1 B5C3
C2 A1C3 A2D4 A4C5 B1B4 B3C2 B5C4
C3 A1C4 A2D5 A4D1 B1B5 B3C3 B5C5
C4 A1C5 A3A4 A4D2 B1C1 B3C4 B5D1
C5 A1D1 A3A5 A4D3 B1C2 B3C5 B5D2
D1 A1D2 A3C1 A4D4 B1C3 B3D1 B5D3
D2 A1D3 A3C2 A4D5 B1C4 B3D2 B5D4
D3 A1D4 A3C3 A5B1 B1C5 B3D3 B5D5
D4 A1D5 A3C4 A5B2 B1D1 B3D4 C1C2
D5 A1A3 A3C5 A5B3 B1D2 B3D5 C1C3
A1A4 A3D1 A5B4 B1D3 B4B5 C1C4
C1C5 C5D3 A1A2C3 A1A4B2 A1A5D1 A2A4B5 A2A5D4
C1D1 C5D4 A1A2C4 A1A4B3 A1A5D2 A2A4C1 A2A5D5
C1D2 C5D5 A1A2C5 A1A4B4 A1A5D3 A2A4C2 A3A4B1
C1D3 D1D2 A1A2D1 A1A4B5 A1A5D4 A2A4C3 A3A4B2
C1D4 D1D3 A1A2D2 A1A4C1 A1A5D5 A2A4C4 A3A4B3
C1D5 D1D4 A1A2D3 A1A4C2 A2A3B1 A2A4C5 A3A4B4
C2C3 D1D5 A1A2D4 A1A4C3 A2A3B2 A2A4D1 A3A4B5
C2C4 D2D3 A1A2D5 A1A4C4 A2A3B3 A2A4D2 A3A4C1
C2C5 D2D4 A1A3B1 A1A4C5 A2A3B4 A2A4D3 A3A4C2
C3C4 D2D5 A1A3B2 A1A4D1 A2A3B5 A2A4D4 A3A4C3
C3C5 D3D4 A1A3B3 A1A4D2 A2A3C1 A2A4D5 A3A4C4
C3D1 D3D5 A1A3B4 A1A4D3 A2A3C2 A2A5B1 A3A4C5
C3D2 D4D5 A1A3B5 A1A4D4 A2A3C3 A2A5B2 A3A4D1
C3D3 A1A3C1 A1A4D5 A2A3C4 A2A5B3 A3A4D2
C3D4 Three toes A1A3C2 A1A5B1 A2A3C5 A2A5B4 A3A4D3
C3D5 A1A3C3 A1A5B2 A2A3D1 A2A5B5 A3A4D4
C4C5 A1A3C4 A1A5B3 A2A3D2 A2A5C1 A3A4D5
C4D1 A1A2B1 A1A3C5 A1A5B4 A2A3D3 A2A5C2 A3A5B1
C4D2 A1A2B2 A1A3D1 A1A5B5 A2A3D4 A2A5C3 A3A5B2
C4D3 A1A2B3 A1A3D2 A1A5C1 A2A3D5 A2A5C4 A3A5B3
C4D4 A1A2B4 A1A3D3 A1A5C2 A2A4B1 A2A5C5 A3A5B4
C4D5 A1A2B5 A1A3D4 A1A5C3 A2A4B2 A2A5D1 A3A5B5
C5D1 A1A2C1 A1A3D5 A1A5C4 A2A4B3 A2A5D2 A3A5C1
C5D2 A1A2C2 A1A4B1 A1A5C5 A2A4B4 A2A5D3 and so on
Codes A5 and B5 are available only for animals with five foretoes.
Regardless of the numbering system used, the potential for toe regeneration
must be considered (see Table 8.3). In addition, the thumbs of the forefeet are
required by males during amplexus, and they should not be clipped (for example,
Briggs and Storm 1970). Another potential problem is understanding how toe-
clipping may stress an animal and affect resource acquisition. Daugherty (1976)
indicated a problem with weight loss in toe-clipped Rana pipiens. The most ser-
ious criticism of toe-clipping anurans was raised by Clarke (1972), who noted
that the probability of recapturing Bufo fowleri decreased as the number of digits
excised increased; researchers today do not clip as many toes as Clarke, however.
Halliday (1995) reported that studies of Atelopus elegans, Atelopus carbonerensis,
Bufo marinus, Bufo granulosus, and Bufo bufo provided little or no evidence for
any impact on survival after toe-clipping. An assessment and concern about use
of toe-clipping of Hyla labialis was provided by Lüddecke and Amezquita (1999).
Fewer than 1% of toe-clipped Rana pretiosa showed any sign of infection, and no
mortality was found (Reaser and Dexter 1996). They stressed that hygienic, ster-
ile techniques must be used, and that impacts should be carefully monitored in
Table 8.3 Digit-regeneration occurrence and times for selected toe-clipped amphibian
species
Species Time interval for regeneration
Anurans
Bufo hemiophrys Several months but recognizable
Hyla regilla Only slight after 1 year
Hyperolius viridiflavus ferniquei Complete in newly metamorphosed
Rana catesbeiana None in 3-year study
Rana erythraea Slow or nonexistent over 4 years
Caudates
Ambystoma opacum Some in short period
Taricha spp. Several years
Taricha granulosa Indefinitely, if kept below or at 10C
Triturus cristatus Slow compared to tail clips
Triturus vulgaris Up to 10 months; little winter regeneration
(Griffiths 1984)
Plethodon cinereus 7 months in laboratory
Plethodon glutinosus At least 2 years
Plethodon wehrlei 50% after 100 days; regenerated digits
identifiable by lack of pigmentation
Cryptobranchus alleganiensis 1 year
Batrochoseps spp. Slow; regenerated toes recognizable
After review by Ferner (2007) unless otherwise indicated.
studies using toe-clipping. Despite these potential problems, toe-clipping is still

the most common marking technique used for anurans.
8.2.2 Salamanders
Toe-clipping has been the common method of marking salamanders in recent
years, and in most cases something similar to Martof ’s (1953) system of mark-
ing anurans is used. Another scheme (as above) was provided by Waichman
(1992) (Table 8.2). The potential of toe regeneration needs to be considered in
salamanders, but whether or not it is a deterrent to using toe-clipping depends
upon the species and study objectives. Table 8.3 summarizes reports in the lit-
erature concerning the regeneration of digits in salamanders.
Some salamanders, such as juvenile (
2.5 cm snout–vent length, or SVL)
dusky salamanders (Desmognathus fuscus), have toes too small for precise clip-
ping. In such cases, animals have been marked successfully by clipping small
pieces of the tail at either right angles to the longitudinal axis of the body or in
the transverse plane at different angles (Orser and Shine 1972). This technique
allowed groups of marked juveniles to be distinguished for at least 1 month
before regeneration became a problem.
Mutilation techniques, such as toe-clipping, may use some form of local or
general anesthesia. Peterman and Semlitsch (2006) stressed the importance of
properly buffering solutions of the anesthetic MS-222 in plethodontid salaman-
ders, and the effect of various concentrations on the time to become anesthe-
tized and recover. They found induction time usually decreased, and the time
for recovery increased, with increasing concentrations of MS-222. This study
should be consulted before using this anesthetic.
Davis and Ovaska (2001) found that toe-clipping in Plethodon vehiculum
resulted in less weight gain than animals marked with fluorescent tags or used
as controls. These authors used unique combinations of three clipped toes per
salamander, removing no more than one toe per foot. The number of ambigu-
ous marks in toe-clipped individuals increased dramatically in frequency com-
pared with the fluorescent tagged animals after 50 days, but the majority of both
retained useful marks throughout the 87-week field study.
8.2.3 Ethical issues related to toe-clipping of amphibians

In recent years there has been increasing concern over the possible impact of vari-
ous marking techniques on animals. Mutilation techniques such as toe-clipping
have been reviewed in several studies to determine their impact on survival. In
addition, ethical issues relative to the possibility of inflicting pain on organisms
are a concern. University Institutional Animal Care Committees (IACUC) in
the USA may request documentation on marking techniques and consideration

of several alternatives. For guidelines see www.ssarherps.org/pages/ethics.php.
In Europe legislation is often even more strict and a license is required before any
invasive surgical procedure including scientific studies using PIT tagging.
Parris and McCarthy (2001) reported evidence that toe-clipping anurans
decreases the rate of return of animals in mark–recapture studies. They encour-
aged clipping as few toes as possible, and stated that researchers must control for
the impact of clipping on the rate of recapture. They further pointed out that
these results have “important implications for the ethical treatment of animals,
the continued use of toe-clipping to mark species of conservation concern, and the
removal of multiple toes from an individual frog or toad” (McCarthy and Parris
2004). The debate on toe-clipping anurans continues (summarized by Ferner
2007 and Phillott et al. 2008).
8.3 Branding
8.3.1 Anurans
Kaplan (1958) described a branding technique for frogs involving the incorp-
oration of India ink into scarified skin of the venter. Numbers were etched in
the skin with a hypodermic needle and then filled with India ink. Erythema
disappeared quickly, no infection was observed, and the numerals remained
distinguishable for more than 3 months. Kaplan also described an electric tat-
tooing technique as an improvement upon the scarification method. Higgins
India Ink™ was found most effective if mixed with a drop of glycerin to aid its
spreading into the skin. Surplus ink was wiped away, leaving a clear, permanent
mark. Some initial inflammation was observed in the frogs, but was only rarely
prolonged. Additional branding techniques are provided in Table 8.4.
Table 8.4 Branding techniques used for marking amphibian species (as reviewed in
Ferner 2007)
Species Description of branding technique
Anurans Scarring with needle or electric tattoo marker then

filled with India ink
Heated chromium nickel wire
Silver nitrate (75%), potassium nitrate (25%)
Ascaphus truei Freeze branding with copper wire cooled in dry ice
Bufo bufo Alcian Blue dye solution using a “Panjet®” innoculator
Eleutherodactylus podiciferus Pressurized fluorescent granular powder
8.3.2 Salamanders
Salamander branding techniques involve heat or freeze brands, or the application
of pigments or dyes into scarified skin. Taber et al. (1975) branded Cryptobranchus
alleganiensis with heated 1.5-cm-high numerical brands of thin nickel-chrome
wire. These remained clear through a 2-year study, but some faded and required
rebranding. Bull et al. (1983) tested freeze-branding application times while mark-
ing Ambystoma macrodactylum. Letter brands of 5 mm high made with silver-tipped
copper rods were immersed for 5 min in liquid nitrogen. A 0.75-s application pro-
duced the best results, with maximum clarity occurring at 3 months. Woolley
(1962) used a black Carter’s felt ink pencil to mark salamanders. Dilute acetic acid
or ammonium hydroxide was used to remove mucus on the salamander’s tail before
application. These marks lasted at least a month.
Taylor and Deegan (1982) injected a fluorescent pigment into red eft
(Notophthalmus viridescens) dermis using a compressed air spray gun with minimal
mortality. After 39 days, individuals began succumbing to water mold (Saprolegnia)
infections, but these could not be linked to the use of fluorescent branding.
Nishikawa and Service (1988) described a technique using dry fluorescent
dust applied with pressurized air on Plethodon jordoni and Plethodon glutino-
sus. The technique required four to 10 colours of inert fluorescent pigments,
canisters, a spray gun with 6.35 mm nozzle, a hose, a single-stage regulator,
and pressurized air. Animals were placed on a dry enamel pan, and the spray
gun was held 1 cm away from the skin surface. Marks were applied using lower
pressures (25 psi) for smaller animals and greater pressures (40 psi) for larger
animals, depending on the size of the fluorescent particles used. Three to four
marks could be applied to each animal on either side of the body (cranial and
caudal to the forelimb, mid-body, and cranial and caudal to the hind limb) for
a maximum of 10 possible locations. Nishikawa and Service (1988) reported
higher recapture rates than for any salamander toe-clipping studies performed
to that time.
8.4 Tagging and banding

The use of any external tag or band must be considered carefully, as the device may
impede the movement of the animal or snag on the substrate and vegetation.
8.4.1 Anurans
Kaplan (1958) used aluminum toe bands to tag frogs (“butt-end” bird band,
#1242, size 2½). These numbered, cylindrical bands were placed around a toe, and
the two ends were pressed together with pliers. The bands were tightened so as not
to restrict circulation, but they pierced the webbing of the foot. These remained
fixed indefinitely and caused no apparent hindrance in the movement of the frog.
A variety of tags and bands used for anurans is summarized in Table 8.5.
McAllister et al. (2004) recommended against the knee tags used by Watson
et al. (2003) in R. pretiosa due to skin and muscle lacerations in 33% of the
recaptured animals. Emlen (1968) reported no differences in behavior, mor-
tality, emigration rates, or weight loss between tagged and untagged American
bullfrogs (Rana catesbeiana). Due to soiling and staining, seasonal replacement
of waistbands was necessary. However, Robertson (1984) reported a negative
impact using Emlen’s (1968) waistband design on 10 species of Australian hylids
and leptodactylids.
Windmiller (1996) suspended yarn tags in a plastic bag filled with fluorescent
pigment which was pressed into the yarn, and attached to the dorsal integument
of juvenile ranid frogs. The pigment trails were followed at night by illuminat-
ing them with a longwave (366 nm) ultraviolet lamp (Blak-Ray ® model ML-49)
while wearing safety/ultraviolet-enhancing glasses. Buchan et al. (2005) used
soft VIAlpha Numeric Tags (US $1.00 per tag) and an injector (US$120) to
insert the tag. Buchan et al. (2005) found high survivorship, retention, and
readability of these tags in both the laboratory and field.
Table 8.5 Tagging and banding techniques used for marking amphibian species (as
reviewed in Ferner 2007 unless otherwise specified)
Species Description of tagging/banding technique
Anurans Eppendorf microcentrifuge tubes filled with Cyalume™ luminescent

fluid glued to dorsum
Soft VIAlpha Numeric Tags® (fluorescent) with imprinted letters and
numbers inserted into sartorius muscle
Bufo bufo Knee-tagging (Elmberg 1989; Kuhn 1994)
Rana catesbeiana Painted nylon waistband around pelvic girdle
Rana catesbeiana and Fluorescent acrylic yarn held between squares of self-adhering
Rana clamitans verterinary bandage and glued to dorsum
Rana clamitans Waistband using surgical thread sewn through piece of surveyor’s
flagging
Rana pretiosa Numeric-coded fingerling tags attached over knee with elastic
thread
Uperoleia rugosa 1-mm squares Scotchlite® reflective sheeting attached with
cyanoacrylate tissue cement
Xenopus laevis Glass beads attached to forelimb with surgical wire
Visual implant elastomers (VIEs), as described for salamanders below, are

a very useful mark for anuran larvae. Their potential for use in marking adult
anurans may not be fully appreciated.
8.4.2 Salamanders
Woolley (1973) tagged Eurycea lucifuga and Eurycea longicauda with a sub-
cutaneous injection of two parts Liquitex acrylic polymer to one part distilled
water. This mixture was injected into the lateral proximal caudal region with
a 22-gauge hypodermic needle, leaving a mark 7–10 mm in diameter. Woolley
observed no adverse effects with this procedure and found slight fading in very
few individuals. The author suggested that a series of acrylic colours be used to
differentiate individuals.
Subcutaneous injections of fluorescent, elastomer dyes were used in Plethodon
vehiculum by Davis and Ovaska (2001). Red, orange, or yellow dyes (VIEs)
were injected into six locations in various combinations to create 816 unique
marks. The elastomer was mixed with a hardener following the manufacturer’s
instructions, and 0.1 mL was placed in a 0.3 mL syringe. The authors recom-
mend fluorescent marking or pattern mapping over toe-clipping in studies of
plethodontid salamanders.
Bailey (2004) evaluated VIE marking in Eurycea bislineata by documenting
mark retention, salamander survivorship, and growth rates. No adverse effects
and no loss of marks were found over the 11-month study. Misreading marks
in the field by observers was found to be a factor in some cases, so a training
period for investigators is recommended to minimize observer bias. Rittenhouse
et al. (2006) used the same technique with powdered fluorescent pigments with
newly metamorphosed Ambystoma maculatum, as described for Rana sylvatica
(see Table 8.4). No temporal effects of powder-dusting were detected.
8.4.3 Caecilians
Gower et al. (2006) reviewed the use of alphanumeric fluorescent tags (VIAlpha
tags) on the Indian caecilian Gegeneophis ramaswamii. They found that making
an incision with a scalpel blade before using the injector increased the efficiency
of tagging, and that anesthesia was needed to quiet specimens. Equipment ster-
ilization between uses minimizes the risk of spreading pathogens.
8.5 Trailing devices

As with bands and tags, the use of trailing devises must be carefully considered
because of the potential to impair the movement of animals. They may be very
Table 8.6 Trailing devices used for marking amphibian species (as reviewed in Ferner
2007 unless otherwise indicated)
Species Description of trailing technique
Ambystoma tigrinum Aquatic forms tagged with trailing plastic float on

monofilament line sutured in tail
Bufo bufo Sewing thread on bobbin fastened around waist with
elastic band (Sinsch 1988)
Bufo valliceps Cotton thread bobbin on rigid plastic tubing holder on
elastic band
Leptodactylus labyrinthicus Cotton thread spool-and-line quilting cocoon on 1 cm
inguinal elastic band
Rana pipiens Nylon sewing thread on rigid plastic tubing holder on
elastic band
Rana sylvatica Metamorphs dipped in fluorescent powder pigments
leaving trail
difficult to attach, and they may snag on the substrate and vegetation. Still, they
are useful in certain applications. Table 8.6 reviews these techniques.
8.6 Pattern mapping

The patterns and markings of many species are distinctive and yet have a high
degree of individual variation. Photographing or sketching these patterns may be
useful for individual recognition on subsequent recaptures (Meyer and Grosse
1997; Streich et al. 1997; Winkler and Heunisch 1997).
8.6.1 Anurans
While more commonly used in salamanders, the patterns of anurans may be
unique enough to make individual recognition possible. Wengert and Gabrial
(2006) used photographs of chin-spot patterns in Rana mucosa and found a high
success rate in reidentification over a 3-month study; changes in spot pattern
over time were observed so regular recaptures might minimize misidentifica-
tions using this technique. The effectiveness of using photographic identifica-
tion on long-term studies of adult anurans has yet to be determined.
Pointing out concerns in using artificial or invasive marks with the protected
Leiopelma archeyi (Archey’s frog) in New Zealand, Bradfield (2004) developed a
highly accurate photographic identification technique using natural markings.
This study is a model on how to develop and test a photographic pattern mapping
technique. A series of six digital photographs were taken from dorsal, ventral,
cranial (facial), caudal, right lateral, and left lateral views. Photographs using a
flash were most successful in discerning usable characteristics of the various black
markings. A filing system for the photos used subgroupings which substantially
reduced the amount of time needed to identify recaptures. Intra- and interobserver
consistency of identification was tested and found to be high. Overall, 99.2% of
identifications were successful once recaptures were assigned to their subgroups.
8.6.2 Salamanders
Naturally occurring variation in morphology was used for Triturus cristatus and
Triturus vulgaris by photographing belly patterns (Hagstrom 1973). With this
technique, one must be sure that the patterns are both recognizably different
and constant through time. Ontogenetic belly pattern variation of T. cristatus
was documented by Arntzen and Teunis (1993). Healy (1974) found the vari-
ation in dorsal spot patterns of Notophthalmus viridescens useful in identifying
individuals. This technique requires recapture and handling which may not be
consistent with the animals’ behavior.
Digital photographs of dorsal spot patterns were tested for their effectiveness
in the identification of Eurycea bislineata by Bailey (2004). Observers were given
color printouts of the dorsal view of each animal with the SVL written below
the image for size reference. The observers compared these photographs to test
animals and were timed as to how long it took them to identify each animal.
Photo-identification rates were very high (about 94% correct), but were poten-
tially lower if observers were also given un-marked (photographed) individuals
in the attempted comparisons. Use of Adobe Photoshop® was effective in com-
paring qualities of integument patterns of vertebrates in digital photographs
and may be useful in matching images (L. Rifai, personal communication).
Sweeney et al. (1994) describe the use of computer analysis for the belly patterns
of T. cristatus. Computer applications for animal pattern analysis can now be
found on websites such as www.conservationresearch.co.uk/.
Loafman (1991) photographed the natural variation of spot patterns in adult
Ambystoma maculatum to successfully identify about 97% of the recaptures.
Adult A. opacum were photographed in a box and successfully reidentified
using their distinctive bar pattern over a 1-year period; using head patterns
alone, 80% of the adults could be distinguished from one another (Doody
1995).
The use of spot patterns for identification of A. maculatum is reviewed by
Grant and Nanjappa (2006) with special attention to possible errors in this
technique. They quantified error rates in specific mapping techniques, modified
techniques to address sources of error, compared methods relative to observer

bias in both the field and laboratory, and determined the effort needed to search
databases for specific individuals. They provided very specific suggestions for
researchers dealing with large samples of specimens identified by patterns.
8.7 Passive integrated transponder (PIT) tags

An extensive review and evaluation of a microchip marking system was given by
Camper and Dixon (1988). The 10 mm 2.1 mm PIT tags are encased in glass
and encoded with an alpha-numeric code which is read by a portable reader with
a hand wand. The authors tested tags on 95 individuals from a wide range of
species, and implanted tags with a metal syringe having a 12-gauge needle. This
baseline study experienced only one failed PIT due to a cracked glass cover, a 6%
error rate in the readings, a reading success on the first pass of the wand of 92%
and a migration rate of a tag within the animal of 36% (with possible reduction
of first-pass reading-rate success) (Camper and Dixon 1988). Examples of spe-
cies of amphibians studied with PIT tags are given in Table 8.7.
Table 8.7 Use of PIT tags for marking amphibian species (as reviewed in Ferner 2007
unless otherwise indicated)
Species Description of technique
Anurans
Bufo bufo Needle injection into pinched dorsum, massaged
caudally to base of spine
Limnodynastes peronii and Inserted with needle behind front limb along side of
Litoria aurea body
Rana catesbeiana, Rana clamitans, Needle injected into pinched ventral side
and Rana pipiens
Rana pretiosa Inserted through 2 mm incision on dorsum 1–2 cm
caudal to eyes, then massaged back along spine
Rana sylvatica Inserted through 2 m incision above the scapula;
“PIT pack” used for detection (Blomquist et al. 2008)
Rana temporaria Needle injection into pinched dorsum, massaged
caudally to base of spine
Caudates
Ambystoma opacum Inserted into body cavity at 3 mm incision 5 mm
cranial to hind limb
Taricha torosa Needle insertion into abdomen
8.7.1 Anurans
All anurans in the Camper and Dixon (1988) project had the PIT tags implanted
intra-abdominally using the implanter syringe with the needle dipped in 70%
ethanol before implantation. Wounds were cleaned after injection with 70% etha-
nol and sealed with Krazy Glue®. Synthetic liquid bandages can also be used.
Ireland et al. (2003) found that the tag migrated on its own to the caudal
lymph space between the hind legs in all but 5% of the frogs they studied. A
field scanner was used for identification and the authors believed the technique
to be the best available for marking medium-to-large-sized anurans. The pri-
mary drawback is the cost of the AVID microchips (≈US$8.00) and scanner
(≈$1000). McAllister et al. (2004) inserted 12-mm tags (125 kHz model) on
frogs 42 mm and massaged them back along the body to the base of the spine
to minimize the chance of their being lost through the incision prior to healing
(about 2 weeks).
Pyke (2005) provides a good review on the use of PIT tags and reports on the
marking of 3000 individuals of nine species. The tags were supplied by Trovan
and are “individually-packaged needles inside hermetically sealed packages.”
The resulting small wound was sealed with Vetbond® (n-butyl cyanoacrylate
adhesive) and fewer than 1% of the animals exhibited any sort of distress call
during the procedure. As with previous studies, Pyke (2005) found little imme-
diate impact or effect on reproduction and long-term survival.
8.7.2 Salamanders
Ott and Scott (1999) anesthetized salamanders (see Table 8.6) for about 10
min with 2-phenoxy ethanol (30 drops/500 ml distilled H2O) and sealed the
incision with New Skin® liquid bandage. Recovery from anesthesia was in con-
tainers of distilled H2O with the head above water for a 3–4-h period until the
animals appeared active when prodded gently. Where no adverse impacts of PIT
tags were reported, Ott and Scott (1999) pointed out the disadvantage of cost
and marking time.
8.7.3 Caecilians
PIT tags probably hold the most promise for marking caecilians, but it is only
appropriate for larger specimens and requires perfecting the injection technique.
Inasmuch as the long salamanders Siren and Amphiuma have been marked suc-
cessfully over a period of more than 5 years (C. K. Dodd Jr, personal commu-
nication), PIT tagging holds promise for studies of caecilians. Since caecilians
may be less likely to be recaptured than some other amphibians, the cost of lost
tags should be considered.
8.8 Taking measurements

It is important to age specimens, place them in a class size (i.e. juvenile or
adult), and document growth rates that precise SVL and tail length measure-
ments be taken. Salamanders and caecilians may be placed on flat transparent
surfaces or restrained in clear plastic tubes with a millimeter scale or metric
stick used for measuring. Snout–vent distances are taken from the tip of the
snout to the caudal end of the cloacal aperture. Tail lengths are taken from the
caudal end of the cloacal aperture to the tip of the tail. Any regenerated por-
tion of the tail should be noted and measured separately. Anurans are usually
measured from the tip of the snout to the caudal projection of the urostyle as
the vent is often not easily found. Calipers are useful in measuring the lengths
of anurans.
Body mass should also be measured in grams after the specimen is patted dry
of any surface moisture. A specimen bag or customized bin on a scale may be
used to contain the animal. Use of an electronic or spring scale (such as Pesola®)
should be selected to obtain a weight to the nearest 0.10 g if possible. As many
amphibians evacuate the bladder or cloaca upon capture, investigators should
allow this to happen before weighing the specimen.
8.9 Recommendations
After extensive review of the literature and considering the criteria for selecting
a marking technique listed in the Introduction to this chapter, the following are
recommended.
• Toe-clipping remains the method of choice for most studies of anurans

with necessary consideration of the impact on the animals and long-term
retention of the clip.
• The emerging technique of choice with caudates is the use of VIE which
may also be used with anurans in the future. Non-invasive techniques,
particularly pattern recognition, are to be considered as a possibility for all
amphibians.
• When financial considerations and practicality dictate, PIT tagging is an
increasingly popular method for long-term studies.
In summary, it is essential that all techniques be considered and understood

thoroughly before use (the advantages, disadvantages, sources of bias), be prac-
ticed with specimens in captivity using good hygiene, and be monitored for the
impact of the mark on the survival and behavior of the amphibian.
8.10 References
Arntzen, J. W. and Teunis, S. F. M. (1993). A six year study on the population dynamics of
the crested newt (Triturus cristatus) following the colonization of a newly created pond.
Herpetological Journal, 3, 99–110.
Bailey, L. L. (2004). Evaluating elastomer marking and photo identification methods for
terrestrial salamanders: marking effects and observer bias. Herpetological Review, 35,
38–41.
Baker, J. and Gent, T. (1998). Marking and recognition of animals. In T. Gent and
S. Gibson (eds), Herpetofauna Worker’s Manual, pp. 45–54. Joint Nature Conservation
Committee, Peterborough.
Blomquist, S. M., Zydlewski, J. D., and Hunter, M. L. (2008). Efficacy of PIT tags for
tracking the terrestrial anurans Rana pipiens and Rana sylvatica. Herpetological Review,
39, 174–9.
Bradfield, K. S. (2004). Photographic Identification of Individual Archey’s Frogs, Leiopelma
archeyi, from Natural Markings. DOC Science Internal Series 191. New Zealand
Department of Conservation, Wellington.
Briggs, J. L. and Storm, R. L. (1970). Growth and population structure of the cascade
frog, Rana cascadae Slater. Herpetologica, 26, 283–300.
Buchan, A., Sun, L., and Wagner, R. S. (2005). Using alpha numeric fluorescent tags for
individual identification of amphibians. Herpetological Review, 36, 43–44.
Bull, E. L., Wallace, R., and Bennett, D. H. (1983). Freeze-branding: a long term mark-
ing technique on long-toed salamanders. Herpetological Review, 14, 81–2.
Camper, J. D. and Dixon, J. R. (1988). Evaluation of a microchip marking system for
amphibians and reptiles. Texas Parks and Wildlife Department, Research Publication,
7100–59, 1–22.
Clarke, R. D. (1972). The effect of toe clipping on survival in Fowler’s toad (Bufo wood-
housei fowleri). Copeia, 1972, 182–5.
Daugherty, C. H. (1976). Freeze branding as a technique for marking anurans. Copeia,
1976, 836–8.
Davis, T. M. and Ovaska, K. (2001). Individual recognition of amphibians: effects of
toe clipping and fluorescent tagging on the salamander Plethodon vehiculum. Journal of
Herpetology, 35, 217–25.
Donnelly, M. A., Guyer, C., Juterbock, J. E., and Alford, R. A. (1994). Techniques for mark-
ing amphibians. In W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. C. Hayek, and
M. S. Foster (eds), Measuring and Monitoring Biological Diversity, Standard Methods for
Amphibians, appendix 2, pp. 277–284. Smithsonian Institution Press, Washington DC.
Doody, J. S. (1995). A photographic mark-recapture method for patterned amphibians.
Herpetological Review, 26, 19–21.
Elmberg, J. (1989). Knee-tagging: a new marking technique for anurans. Amphibia-
Reptilia, 10, 101–4.
Emlen, S. T. (1968). A technique for marking anuran amphibians for behavioral studies.
Ferner, J. W. (2007). A Review of Marking and Individual Recognition Techniques for
Amphibians and Reptiles. Herpetological Circular No. 35. Society for the Study of
Amphibians and Reptiles, Salt Lake City, UT.
Gower, D. J., Oommen, O. V., and Wilkinson, M. (2006). Marking amphibians with
alpha numeric fluorescent tags: caecilians lead the way. Herpetological Review, 37, 302.
Grant, E. H. C. and Nanjappa, P. (2006). Addressing error in identification of Ambystoma
maculatum (Spotted Salamanders) using spot patterns. Herpetological Review, 37, 57–60.
Griffiths, R. A. (1984). Seasonal behaviour and intrahabitat movements in an urban
population of smooth newts Triturus vulgaris (Amphibian: Salamandridae). Journal
of Zoology, 203, 241–51.
Hagstrom, T. (1973). Identification of newt specimens (Urodela, Triturus) by recording
the belly pattern and a description of photographic equipment for such registrations.
British Journal of Herpetology, 4, 321–6.
Halliday, T. (1995). More on toe-clipping. Froglog, 12, 2–3.
Healy, W. R. (1974). Population consequences of alternative life histories in Notophthalmus
v. viridescens. Copeia, 1974, 221–9.
Henle, K. and Veith, M. (eds) (1997). Naturschutzrelevante Methoden der
Feldherpetologie. Mertensiella, 7.
Ireland, D., Osbourne, N., and Berrill, M. (2003). Marking medium to large sized
anurans with Passive Integrated Transponder (PIT) tags. Herpetological Review, 34,
218–20.
Kaplan, H. M. (1958). Marking and banding frogs and turtles. Herpetologica, 14,
131–2.
Kuhn, J. (1994). Methoden der Anuren-Marierung für Frielandstudien: Übersicht-
Knie- Ringetiketten-Erfahrunger mit der Phalangenamputation. Zeitscrift für
Feldherpetologie, 1, 177–92.
Loafman, P. (1991). Identifying individual spotted salamanders by spot pattern.
Lüddecke, H. and Amezquita, A. (1999). Assessment of disc clipping on the survival and
behavior of the Andean frog Hyla labialis. Copeia, 1999, 824–30.
Martof, B. S. (1953). Territoriality in the green frog, Rana clamitans. Ecology, 34,
165–74.
McAllister, K. R., Watson, J. W., Risenhoover, K., and McBride, T. (2004). Marking
and radiotelemetry of Oregon spotted frogs (Rana pretiosa). Northwestern Naturalist,
85, 20–5.
McCarthy, M. A. and Parris, K. M. (2004). Clarifying the effect of toe clipping on frogs
with Bayesian statistics. Journal of Applied Ecology, 41, 780–6.
Meyer, F. and Grosse, W.-R. (1997). Populationsökologische Studien an Amphibien
mit Hilfe der fotografischen Individualerkennung: Übersicht zur Methodik und
Anwendung bei der Kreuzkröte (Bufo calamita). Mertensiella, 7, 79–92.
Nishikawa, K. C. and Service, P. M. (1988). A fluorescent marking technique for individ-
ual recognition of terrestrial salamanders. Journal of Herpetology, 22, 351–3.
Orser, P. N. and Shine, D. J. (1972). Effects of urbanization on the salamander
Desmognathus fuscus fuscus. Ecology, 53, 1148–54.
Ott, J. A. and Scott, D. E. (1999). Effects of toe-clipping and PIT-tagging on growth and
survival in metamorphic Ambystoma opacum. Journal of Herpetology, 33, 344–8.
Parris, K. M. and McCarthy, M. A. (2001). Identifying effects of toe clipping on anuran
return rates: the importance of statistical power. Amphibia-Reptilia, 22, 275–89.
Peterman, W. E. and Semlitsch, R. D. (2006). Effects of tricaine methanesulfonate

(MS-222) concentration on anesthetization and recovery in four plethodontid sala-
manders. Herpetological Review, 37, 303–4.
Phillott, A., Skerratt, L., McDonald, K., Lemckert, F., Hines, H., Clarke, J., Alford, R.,
and Speare, R. (2008). Toe clipping of anurans for mark-recapture studies: acceptable
if justified. That’s what we said! Herpetological Review, 39, 149–50.
Pyke, G. H. (2005). The use of PIT tags in capture-recapture studies of frogs: a field
evaluation. Herpetological Review, 36, 281–5.
Reaser, J. K. and Dexter, R. E. (1996). Rana pretiosa (Spotted Frog). Toe clipping effects.
Rittenhouse, T. A. G., Altnether, T. T., and Semlitsch, R. D. (2006). Fluorescent powder
pigments as a harmless tracking method for ambystomids and ranids. Herpetological
Review, 37, 188–91.
Robertson, J. G. M. (1984). A technique for individually marking frogs in behavioral
studies. Herpetological Review, 15, 56–7.
Sinsch, U. (1988). Seasonal changes in the migratory behaviour of the toad Bufo bufo:
direction and magnitude of movements. Oecologia, 76, 390–8.
Streich, W. J., Beckmann, H., Schneeweiss, N., and Jewgenow, K. (1997).
Computergestützte Bildanalyse von Fleckenmustern der Rotbauchunke (Bombina
bombina). Mertensiella, 7, 93–102.
Sweeney, M., Oldham, R. S., Brown, M., and Jones, J. (1994). Analysis of belly pat-
terns for individual newt recognition. In T. Gent and R. Bray (eds), Conservation and
Management of Great Crested Newts Triturus cristatus. Symposium Proceeding at Kew
Gardens. English Nature Science Series 20, pp. 75–7. English Nature, Peterborough.
Taber, C. A., Wilkinson, R. F., and Topping, M. S. (1975). Age and growth of hellbenders
in the Niangua River, Missouri. Copeia, 1975, 633–9.
Taylor, J. and Deegan, L. (1982). A rapid method for mass marking amphibians. Journal
of Herpetology, 16, 172–3.
Waichman, A. V. (1992). An alphanumeric code for toe clipping amphibians and reptiles.
Watson, J. W., McAllister, K. R., and Pierce, D. J. (2003). Home ranges, movements,
and habitat selection of Oregon spotted frogs (Rana pretiosa). Journal of Herpetology,
37, 292–300.
Wengert, G. M. and Gabrial, M. W. (2006). Using chin spot patterns to identify individ-
ual mountain yellow-legged frogs. Northwestern Naturalist, 87, 192.
Windmiller, B. (1996). Tracking techniques useful for field studies of anuran orientation
and movement. Herpetological Review, 27, 13–15.
Winkler, C. and Heunisch, G. (1997). Fotografische Methoden der Individualerkennung
bei Bergmolch (Triturus alpestris) und Fadenmolch (T. helveticus) (Urodela,
Salamandridae). Mertensiella, 7, 71–7.
Woolley, H. P. (1973). Subcutaneous acrylic polymer injections as a marking technique
for amphibians. Copeia, 1973, 340–1.
Woolley, P. (1962). A method of marking salamanders. Missouri Speleology, 4, 69–70.
9
Egg mass and nest counts
Peter W. C. Paton and Reid N. Harris
9.1 Background: using egg mass and

nest counts to monitor populations
There is a critical need to develop statistically sensitive monitoring programs to
assess changes in the distribution and abundance of amphibians and determine
underlying mechanisms affecting populations (Alford and Richards 1999;
Marsh 2001; Storfer 2003; Petranka et al. 2004, 2006). However, assessing
trends is challenging because amphibian populations may experience dramatic
fluctuations (Pechmann et al. 1989) and all species have imperfect detection
probabilities (Bailey et al. 2004). Biologists interested in monitoring long-term
amphibian population trends have a number of methods available (Heyer et al.
1994), but selecting the most appropriate survey and analysis techniques can be
difficult (Williams et al. 2002).
Techniques used to assess amphibian populations during the breeding season
include surveys of calling anurans (Weir et al. 2005), dip-net sweeps to cap-
ture larvae in aquatic systems (Werner et al. 2007), drift fences with pit falls
(Pechmann et al. 1989), egg-mass counts (Adams et al., 1998; Lips et al. 2001;
Dodd 2003; Rödel and Ernst 2004), and nest counts (Corser and Dodd 2004;
Harris 2005). Advantages of calling surveys include the ability to survey many
breeding wetlands rapidly and most anurans can be detected, but urodeles are
not sampled because they do not vocalize. Dip-net surveys assess the occurrence
and density of larval amphibians, but are labor-intensive, require knowledge of
local breeding phenology, and require expertise in larval identification. Drift-
fence arrays can accurately assess changes in adult population size and prod-
uctivity for individual ponds; however, they are expensive because fences are
difficult to install and traps need to be checked often.
Many amphibians oviposit spawn in clumped aggregations, rather than
individual eggs. These spawn clumps are usually called “egg masses” in the
North America literature (e.g. Paton and Crouch 2002) and in the European
literature as “spawn clumps” (Griffith and Raper 1994), “batches” (Barton and
Rafinski 2006), “clutches” (Ficetola et al. 2006), or “egg masses” (Sofianidou
and Kyriakopoulou-Sklavounou 1983; Waringer-Löschenkohl 1991; Hartel
2008).
Nest counts can monitor populations in a similar manner to egg-mass counts.
Both types of count estimate annual breeding effort of populations and provide
an index of population size. In addition, both counts can test for effects of cli-
mate on population-level breeding effort (Corser and Dodd 2004), habitat rela-
tionships (Chalmers and Loftin 2006), population presence, and life-history
information (e.g. relationships between egg size and clutch size, and between
egg size and embryonic survival). Comparisons of egg size and clutch size among
populations can investigate maternal investment patterns as adaptations to local
environmental conditions. For nest counts, behavioural information can often
be gained, such as frequency of solitary and communal nesting and patterns of
nest attendance or brooding.
In the following sections we describe characteristics that make a species suit-
able for egg mass or nest counts. We also discuss factors that need to be consid-
ered when designing studies to collect and analyze these types of survey data.
Both egg mass and nest counts can be used to study habitat use over a variety of
spatial scales, thus are good candidates for landscape and metapopulation stud-
ies designed to manage and conserve at-risk species.
9.2 Oviposition strategies

Amphibians exhibit up to 39 different reproductive modes based on factors
such as oviposition site (e.g. aquatic, terrestrial, live-bearing), egg type, and
mode of development (Wells 2007). Worldwide, at least 10 families of anurans
and some urodeles have terrestrial oviposition (Wells 2007; see Appendix 9.1).
Although terrestrial oviposition strategies for many species have been described,
using egg counts to monitor their populations typically are not considered due
to low detection probabilities and potential negative impacts on populations by
disturbing eggs in nests. However, we review several case studies below where
nest-count surveys were used to monitor populations because detection prob-
abilities were high and negative impacts were low, which suggests that a similar
approach might be used for other species.
About 80% of anuran families deposit eggs in standing water (Wells 2007).
Species that oviposit in smaller wetlands, such as vernal pools (e.g. Ranidae,
Microhylidae), are probably best suited for egg-mass counts because biologists
9 Egg mass and nest counts | 145
can search sites completely. For example, Egan and Paton (2004) found that
most pond-breeding amphibian sites were less than 0.05 ha in southern New
England, USA. Oviposition sites selected by some tropical species are extremely
small bodies of water, such as treeholes, aerial plants, or water-filled seedpods
(Wells 2007). However, although their oviposition habitat has been described,
to our knowledge no one has used egg-mass counts to monitor tropical amphib-
ian populations. Searching for egg masses of species that breed in large lakes is
probably not feasible due to low detection probabilities, and thus other monitor-
ing strategies should be employed. For stream-breeding species, egg masses can
be detected (e.g. Ascaphidae), although stream monitoring programs typically
search for juveniles and adults (Adams and Bury 2002).
Some anurans lay eggs in foamy nests on the water’s surface or in water-filled
basins (e.g. members of the subfamily Leptodactylinae; Wells 2007), which can
be highly visible on the surface of shallow, temporary pools. Also, rhacophorid
and hyperoliid frogs build foam nests in trees and other substrates. Thus, one
possible technique to monitor their populations might be to attempt to count
foam nests, but given the difficulties of finding the nests of most foam-nesting
species, this is probably not a cost-effective technique for most species.
There are many arboreal tropical species that deposit eggs on leaves, branches,
or tree trunks above standing water (see summary in Wells 2007). Neotropical
hylids oviposit on upper surfaces of leaves and develop rapidly, where tadpoles
fall into the water below to complete metamorphosis. Phyllomedusine hylid
frogs from Central and South America deposit eggs on leaves and other sub-
strates. Although monitoring schemes have not incorporated egg mass or nest
counts of these arboreal breeding species, it is something that tropical biologists
could consider.
9.3 Egg-mass counts

Many species of pool-breeding amphibians oviposit egg masses attached to
submerged vegetation or on the pond bottom, with adults not attending eggs
during incubation. For these species, biologists have used egg-mass counts
to assess population sizes and trends of anurans (Table 9.1) and urodeles
(Table 9.2). The best long-term studies come from Europe, where several
biologists have monitored Rana temporaria egg counts for decades (Cooke
1985; Meyer et al. 1998; Loman and Andersson 2007). In North America,
egg-mass counts were initiated for Bufo canorus (Sherman and Morton 1993),
Rana sevosa (Richter et al. 2003), Rana sylvatica, and Ambystoma maculatum
(Petranka et al. 2004).
Table 9.1 Examples of studies where anuran populations were assessed by counting
egg masses for more than 2 years.
Species Country Habitat Years Source
Hyla regilla USA (Washington) Ponds 4 Adams et al. (1998)

Bufo calamita England Ponds 20 Buckley and Beebee (2004)
Bufo canorus USA (California) Lakes 20 Sherman and Morton
(1993)
Rana sevosa USA (Mississippi) Seasonal 13 Richter et al. (2003)
pools
Bufo boreas USA (Colorado) Lakes 12 Muths et al. (2003)
Rana capito USA (Alabama) Pond 13 Jensen et al. (2003)
Rana sylvatica USA (Wyoming) Seasonal 10 Corn and Livo (1989)
pools
R. sylvatica USA (southeastern Ponds 10 Petranka et al. (2004)
Kentucky)
Rana aurora USA (Washington) Ponds 4 Adams et al. (1998)
Rana temporaria England Ponds 14 Cooke (1985)
R. temporaria Sweden Ponds 17 Loman and Andersson (2007)
R. temporaria Switzerland Fens 28 Meyer et al. (1998)
R. temporaria Austria Ponds 7 Gollmann et al. (2002)
R. temporaria England Ponds 23 Williams (2005)
Rana arvalis Sweden Ponds 17 Loman and Andersson (2007)
Rana dalmatina Germany Pond 7 Gollmann et al. (2002)
R. dalmatina Romania Pond 11 Hartel (2008)
R. dalmatina Greece Pond 3 Sofianidou and Kyriakopoulou-
Sklavounou (1983)
R. dalmatina Austria Pond 7 Waringer-Loschenkohl (1991)
R. capito, Rana USA (Mississippi) Seasonal 3 Richter (2000)
sphenocephala pools
Rana pipiens USA (Wyoming) Ponds 10 Corn and Livo (1989)
Rana japonica, Japan Lenthic 1 Kuramoto (1978)
Rana tsushinensis, waters
Rana
nigromaculata
Hyperolius spp. Uganda Arboreal 1 Vonesh (2000)
Table 9.2 Examples of studies where urodele populations were assessed by counting
egg masses.
Species Country Habitat Years Source
Ambystoma USA (Rhode Island) Seasonal pools 2 Egan and Paton

maculatum (2004)
A. maculatum USA (Maine) Seasonal pools 2 Calhoun et al.
(2003)
A. maculatum USA (Ohio) Semipermanent 4 Brodman (1995)
pond
A. maculatum USA (southeastern Ponds 10 Petranka et al.
Kentucky) (2004)
Ambystoma. USA (Ohio) Semipermanent 4 Brodman (1995)
jeffersonianum pond
Ambystoma gracile USA (Washington) Lenthic 4 Adams et al.
(1998)
Ambystoma USA (Washington) Lenthic 4 Adams et al.
macrodactylum (1998)
Ambystoma barbouri USA (Kentucky) Stream 2 Kats and Sih
(1992)
Taricha torosa USA (California) Stream 2 Gamradt and
Kats (1997)
In contrast to drift-fence arrays, egg-mass counts only require biologists to

survey a pond several times annually to accurately estimate annual breeding
effort, and thus are much less labor-intensive (hence less expensive) than drift-
fence arrays. In addition, many sites can be monitored annually using egg-mass
counts, for example, Loman and Andersson (2007) monitored Rana arvalis and
R. temporaria populations using egg-count surveys in 120 ponds in Sweden over
17 years. In Australia, Martin and Cooper (1972) conducted egg-mass counts of
Crinia victoriana, and egg-mass counts are proposed for several vulnerable spe-
cies in including Geocrinia alba, Litoria littlejohni, and Litoria verreauxii alpina
(University of Canberra 2003).
9.4 Amphibian nests and nest counts

We use the term nest to mean an oviposition site associated with at least one of
the following characteristics: habitat alteration, including reduction of patho-
genic fungi, nest construction, and nest attendance. An amphibian nest is a
specific location selected by a female for oviposition and is often associated with
nest attendance by a parent. In some species, males or females may conduct

some habitat alteration or nest construction. For example, male frogs construct
basins in Hyla rosenbergi (Hobel 1999) and females do so in Litoria lesueuri
(Richards and Alford 1992). Female four-toed salamanders Hemidactylium scu-
tatum appear to compress soil or mat vegetation at their nesting site. In other
species, and perhaps most species, males or females select specific oviposition
sites, such as cavities in rotten logs or slight depressions in a rock, but probably
do not obviously modify the location or engage in nest construction. However,
nest attendants can modify the nest site to reduce the presence of pathogenic
fungi, which requires specific techniques to detect (Banning et al. 2008).
Before initiating nest counts, we recommend studying published accounts
and consulting with herpetologists who work or have worked on the species
of interest. Detailed studies of four-toed salamander nests by three different
research groups illustrate how nest counts can be incorporated into a study
of population dynamics (Corser and Dodd 2004; Harris 2005) and habitat
requirements (Chalmers and Loftin 2006; Wahl et al. 2008). The nests of this
species are fairly easy to detect; however, nest-site detection probabilities vary
among species. Salamanders that nest deep within talus, such as Plethodon
punctatus, will probably prove impossible to monitor by nest-count surveys.
Female four-toed salamanders lay eggs in moss and grass on the banks of
ponds and bogs. After hatching, gilled larvae disperse into aquatic habitats,
where they metamorphose after a relatively short larval period. Females attend
nests, communal nesting can be very common, and males are not associated
with nests. Because nests are restricted to banks near the water, they are rela-
tively easy to locate by gently separating mosses and grasses (Figure 9.1).
Corser and Dodd (2004) used nest counts to track population sizes of several
populations in the Great Smoky Mountain National Park, Tennessee, USA, and
determine that the population was not declining. Harris (2005) used nest counts
to reach similar conclusions for populations in montane Virginia. Since Harris
(2005) observed no decline in two populations over a 14-year period and he used
the same nest-location techniques each year, nests can be located and embryos
counted without harming populations. Chalmers and Loftin (2006) and Wahl
et al. (2008) used nest counts to quantify habitat characteristics of ponds and
bogs associated with successful nesting, which can be incorporated into habitat-
restoration efforts. In addition, nest counts are the primary technique used to
assess communal nesting and solitary brooding behaviour (Harris 2008).
Corser (2001) used nest counts to monitor population trends of green sala-
manders Aneides aeneus in the Appalachian Mountains, USA, by conducting
Fig. 9.1 Nests of the four-toed salamander (Hemidactylium scutatum) can be

discovered or checked by carefully spreading grasses and reeds along pond edges.
Nests were best located at this site by standing in the pond and carefully checking
the margin of the pond for nests.
extensive surveys of nests and attending females in the same populations begin-
ning in the early 1970s and then again in the 1990s. These data were critical in
showing that at least three large populations had severely declined, assuming
that each pond surveyed represented a separate population or deme.
We know of no studies that have used anuran nest counts to track changes
in population size. However, nest sites of Eleutherodactylus coqui have been
studied extensively in Puerto Rico (Stewart and Pough 1983; Townsend 1989;
Townsend and Stewart 1994). Counting nests may track E. coqui population
trends because researchers were able to confidently check all suitable potential
nest sites. For E. coqui, preferred nest sites are elevated, enclosed, and often occur
in curled dead leaves of Cecropia peltata and fronds of the palm Prestoea mon-
tana. In an experimental study, adding artificial nest sites increased population
density, including the number of clutches, indicating that preferred nest sites
are limited (Stewart and Pough 1983). One advantage of diurnal nest-count
surveys is that they can be conducted under most weather conditions, whereas
frog-call surveys in many areas are best done at night under rainy, moist, or
humid conditions.
9.5 Clutch characteristics

Females of some explosive-breeding species deposit a single jelly spawn annu-
ally that can contain from 750 to 1750 eggs (e.g. R. sylvatica; Crouch and Paton
2000; Figure 9.2). Both R. sylvatica and R. temporaria are cold-water breeders that
deposit egg masses in large communal aggregations near the surface (Seale 1982).
For relatively short-lived species, such as R. sylvatica, females may only breed once
or twice in their lifetime and breed every year. Crouch and Paton (2000) docu-
mented a 1:1 ratio between the number of egg masses counted in ponds and the
number of female R. sylvatica entering ponds based on drift-fence captures. Thus,
egg-mass counts can represent a precise estimate of annual breeding effort or the
total number of females depositing spawn at the breeding site that year.
Fig. 9.2 Egg masses of spotted salamanders (Ambystoma maculatum; upper panel)
and wood frogs (Rana sylvatica; lower panel) are commonly found in seasonal pools
in eastern North America. Photographs by Scott Egan.
In other explosive-breeding species, such as A. maculatum, females may ovi-

posit two to four clutches annually at a single wetland, with total reproductive
production of 100–300 eggs annually (Petranka 1998; Figure 9.2). Because
there is not a 1:1 ratio between number of egg masses and number of females
ovipositing for these multiple clutch breeders, quantifying long-term popula-
tion trends is more difficult. In fact, many temperate (12 species) and tropical
anurans (nine species) can deposit from two to 12 clutches annually (Wells
2007, p. 501).
Many species are not suited for egg-mass counts. Some species deposit clutches
of more than 1 m wide on the surface, but only one egg thick (e.g. Rana cates-
beiana can deposit 12 000 eggs in 12 min). Because eggs in warm water tend to
hatch rapidly (
5 days), detection probabilities during occasional surveys are
low. Therefore species that breed in warm water are probably best surveyed by
techniques other than egg-mass counts. Toad (Bufo) and some pelobatid egg
masses are often not distinctive because they are deposited in long intertwined
strings, thus it is difficult to count individual egg masses (but see Muths et al.
2003). Finally, many explosive-breeders oviposit single eggs or small packets at the
bottom of ponds rather than in aggregated clutches (e.g. Pseudacris, Bombina,
Discolglossus, and myobatrachid frogs; Parmelee et al. 2002; Wells 2007), thus
using egg counts to monitor those species is not practical. We suggest that alter-
native techniques such as calling surveys need to be used to monitor species with
these oviposition strategies.
Another issue is reproductive frequency, as females of species that live longer,
such as many Ambystoma salamanders (e.g. A. maculatum may live to be 20–30
years old), may only breed every 2–3 years (Petranka 1998). Females vary
reproductive frequency based on their ability to accumulate adequate resources
necessary for oviposition, which may vary with any number of factors, such
as climatic conditions (Harris and Ludwig 2004). In four-toed salamanders,
Corser and Dodd (2004) noted a positive correlation between the number
of nests and precipitation during the nesting season. A similar relationship
between rainfall and nests was found in frog E. coqui (Townsend and Stewart
1994). Females may skip reproduction when conditions do not favor migration
to nesting sites, even if they have accumulated enough resources to oviposit.
Environmental influences are probably common among amphibian species,
suggesting that egg mass and nest counts could be used to reveal important
life history and behavioural correlates but must be used cautiously to estimate
population size. Thus egg and nest counts may represent an estimate of annual
breeding effort, but not the total number of females in the vicinity of a breed-
ing pond.
9.6 Spatial distribution of eggs

Knowledge of the spatial distribution of oviposition sites within a wetland is
helpful when searching during monitoring programs. Anurans exhibit a variety
of spatial distribution patterns when ovipositing in lentic habitats (still-water
habitats such as ponds and lakes). Wells (2007) characterized nine different
social organization strategies among anurans, with a combination of clumped
oviposition sites (e.g. R. sylvatica, Rana lessonae. R. catesbeiana, Bufo calam-
ita, Scaphiopus couchii), dispersed oviposition sites across the entire wetland
(Bombina bombina, Hyla versicolor), and uniformly dispersed oviposition sites
in shallow sections of wetlands (Rana porosa, R. clamitans).
Spatial distribution patterns of oviposition sites can vary within a species
depending on a variety of characteristics (light, temperature, vegetation, preda-
tion). These may lead to shifts within habitats. For example, Crouch and Paton
(2000) documented a clumped aggregation of R. sylvatica in ponds with more
than 100 egg masses, whereas smaller aggregations were dispersed throughout
a pond where fewer than 100 egg masses were detected. In addition, although
communal aggregations of R. sylvatica egg masses often are located in the north-
east corner of ponds where water temperatures are highest (Seale 1982), this can
vary among ponds (P. Paton, unpublished results). Thus observers often need to
search an entire pond to accurately count all egg mass at a site.
9.7 Breeding phenology

Amphibians exhibit an array of reproductive modes during oviposition that can
be classified as either explosive (most individuals breeding in a single night) or
prolonged (some tropical species breed year-round). The length and timing of
the oviposition period is primarily related to abiotic factors such as temperature,
precipitation, and hydroperiod at aquatic breeding sites (reviewed by Wells 2007).
One of the most important considerations when conducting egg-mass counts is to
conduct surveys when the number of clutches reaches peak abundance, although
obviously for prolonged breeders this means multiple visits to breeding sites.
There is considerable interspecific and intraspecific variation in egg-deposition
dates; thus, determining breeding phenology in your local area is critical. You
should refer to local field guides or experts before initiating egg counts, or if no
quantitative data are available for your region, then you have to gather initial breed-
ing phenology data. Biologists in North America can use the North American
Amphibian Monitoring Program (NAAMP; Weir et al. 2005) to help determine
local breeding phenology.
Species in temperate areas typically breed in warmer months, with many

ranids, bufonids, and ambystomatids breeding explosively in early spring. Some
explosive breeders, such as R. sylvatica in North America (Crouch and Paton
2002) or R. temporaria in Europe, breed synchronously once annually within
2–3-week period (Elmberg 1990), whereas other species, such as southern leop-
ard frog (R. sphenocephala) exhibit a more prolonged breeding season, with short
breeding bouts separated by 1–2 months (Doody and Young 1995). Thus, local
knowledge of the length of the oviposition period and hatching rates is critical
when designing an egg-count monitoring program.
Tropical and subtropical species often breed throughout the year during long
rainy seasons, with the number of breeding individuals varying as a function of
rainfall (reviewed by Wells 2007). Some species that breed in permanent water
may be continuously active throughout the rainy season (e.g. Scinax boulengeri),
whereas species that breed in temporary aquatic habitats may oviposit explo-
sively only after heavy rainfall and then undetectable for the rest of the year (e.g.
Hyla pseudopuma). Finally, species that breed in the desert or savanna are explo-
sive breeders that take advantage of ephemeral seasonal wetlands, thus ovipos-
ition and egg development in these species tends to be rapid, hence monitoring
schemes for these species also have to be flexible.
9.8 Number of surveys needed

The goal of egg mass and nest counts is to determine the total number of females
that deposited clutches that year at a particular site. Species that are explosive
breeders are better suited for egg-mass counts. Species that breed relatively syn-
chronously (e.g. R. sylvatica, Crouch and Paton 2000; R. temporaria, Meyer et al.
1998) are ideal because a pond only need to be surveyed a few times to determine
the maximum number of egg masses. In Sweden, Loman and Andersson (2007)
studied R. temporaria populations by conducting three to six surveys annually
at each pond, with surveys separated by 3–7-day intervals. They stopped sur-
veying ponds only after no new (
5-day-old) spawn was detected. In contrast,
species with prolonged breeding seasons may require multiple surveys to a site to
estimate total annual reproduction (e.g. R. sevosa; Richter et al. 2003).
In addition, species whose eggs do not hatch rapidly ( 10 days from depos-
ition to hatching) are also best for egg-mass counts because they have higher
detection probabilities. In North America and Europe, those species that breed
early in the breeding season tend to be the best species to monitor with egg-
mass counts because their eggs can take over 3 weeks to hatch (e.g. R. sylvatica,
R. temporaria, R. arvalis, Rana dalmatina, A. maculatum); thus, biologists have
a longer survey window. For example, in Rhode Island, R. sylvatica eggs may
take over 20 days to hatch in cold ponds (Crouch and Paton 2000). In contrast,
R. catesbeiana eggs that are deposited in June may hatch in under 3 days.
9.9 Estimating egg-mass detection probabilities

Herpetologists have been criticized for assuming perfect detection probabilities
when conducting surveys for individuals, egg masses, and nests (Mazerolle et al.
2007). Even trained observers often do not count all egg masses present at a site
(Williams et al. 2002; Grant et al. 2005). Not adjusting counts for the proportion
of egg masses that are not detected can lead to a bias in trend estimates. Therefore it
is critical for observers to estimate detection probabilities of egg masses, which is the
proportion of the true number of egg masses present at a site that are detected.
Grant et al. (2005) suggested that estimating detection probability using a
double-observer sampling procedure or another capture–recapture method was
essential if the primary objective was to estimate changes in population size. In
this technique, two observers survey sites together, with the primary observer ini-
tially detecting egg masses and reporting them to the recorder. The recorder docu-
ments any egg masses that the primary observer missed. After half of a pond has
been surveyed, the two biologists switch observer and recorder roles (Nichols et al.
2000). Grant et al. (2005) did not detect any variables (e.g. observer variation,
pool area or depth, or vegetation structure) or covariates (e.g. observer experience
or differences in visual abilities, variation between amphibian species, or egg mass
density) that explained variation in detection probability of egg masses.
There are alternative methods that biologists could use to estimate detection
probabilities and estimate population sizes. For example, one observer could
count egg masses in a pond on multiple occasions within a short time interval,
making sure to mark detected egg masses (e.g. Regester and Woosley 2005).
These types of data could be modeled using a capture–recapture framework
(Williams et al. 2002; Grant et al. 2005). Another potential technique is an
independent double-observer method, where a second observer conducts a
count separately from the primary observer without sharing information until
after the survey has been completed.
9.10 Variation in counts among observers

Grant et al. (2005) found strong evidence of variation among observers in their
ability to detect and count egg masses. Windmiller (1996) found less than 10%
difference between surveys during independent double-observer counts of
A. maculatum egg masses in four seasonal ponds. Crouch and Paton (2000) esti-
mated about 12% variation by two independent observers surveying R. sylvatica
egg masses. Egan (2001) found that independent double-observer counts of A.
maculatum egg masses varied by 25%, while R. sylvatica egg-mass counts varied
by 11%. All these species had relatively high detection probabilities (84–100%,
Grant et al. 2005). Based on work by Grant et al. (2005) and Williams et al.
(2002), studies using multiple observers need to incorporate an observer param-
eter when modeling trend estimates.
9.11 Marking eggs

Some methods require that egg masses are marked so that observers can tell
which egg masses have or have not been detected. Egg masses can be marked
with colored flagging tied to adjacent vegetation, but this can be confusing
and difficult if no vegetation is close to the eggs (P. Paton, personal observa-
tion). Alternatively, Regester and Woosley (2005) tested the field practicality
and effects of visible fluorescent elastomer (visible implant elastomer; VIE) on
embryo development and retention in the jelly matrix. There was no effect on
embryo viability, and marks remained intact for more than 35 days and could be
detected up to 0.45 m deep. Because VIE comes in multiple colors, egg masses
could be injected with a different color on each survey.
9.12 Situations in which nest counts are not practical

9.12.1 Nest destruction
Nests of many salamander and frog species are cryptic or difficult to detect
(e.g. underground, inside rotting logs, deep in talus, under logs). Nest discov-
ery may result in damage to the nest, the nest environment, the embryos, and
to the nest attendant. For example, Peterson (2000) found that 40 red-backed
salamander Plethodon cinereus nests were damaged, while another 15 nests that
were not damaged. Merely turning a log to check for a nest can sometimes dam-
age the nest. One can minimize nest destruction by altering search techniques.
However, if nests or nest attendants are typically damaged upon discovery, then
nest counts cannot be used to monitor changes in population size. Only use nest
counts when there is no harm to nests or nest attendants.
9.12.2 Desertion of attendant

Nest discovery may cause the nest attendant to desert the nest. Breitenbach (1982)
suggested that nest disturbance caused the desertion of female H. scutatum,
although Harris et al. (1995) found the females could be disturbed daily in meso-
cosms without causing desertion. Some species, such as the anuran Cophixalus
ornatus, appear to treat humans as predators and will lunge toward human fin-
gers and not desert (Felton et al. 2006). Preliminary studies should be conducted
to assess the effects of nest discovery and degree of nest disturbance. Quickly
checking for nest presence will be less likely to cause desertion than temporar-
ily removing and measuring the attendant and counting embryos. In many spe-
cies, desertion by the nest attendant is associated with reduced embryonic success
(Harris et al. 1995). Therefore, if the objective is to monitor or conserve popula-
tion size, then it is important to not cause desertion.
9.13 How to count eggs in a nest

Valuable information can be obtained by counting embryos in a nest. Clutch size
is a fundamental fitness component for studies ranging from natural selection
and evolution to conservation biology. Obtaining an accurate count of embryos
will depend on the number of embryos in the nest, how the embryos are con-
figured within the nest, and the line of sight from observer to nest. When there
are few eggs arranged in a monolayer in easy sight, counts should be accurate.
Nest photographs can be used to count eggs when all embryos are visible. When
embryos are laid in a three-dimensional cluster, accurate counts are more diffi-
cult, especially for large clutches. Amphibian embryos are often attached to each
other in nests, so moving embryos must be done carefully to avoid damaging
adjacent embryos. Once an investigator has experience with counting embryos,
accurate estimates of the number of embryos without actually seeing each one
can be obtained by understanding how clusters of embryos are arranged. Such a
procedure should be verified by first estimating and then counting all embryos.
Measurement error should be estimated by replicating counts within a nest.
Nests in crevices or cavities can sometimes be surveyed with a strong light source
and tools, such as a dental mirror (Corser 2001).
9.14 Estimating hatching success

Hatching success is calculated as number of embryos that hatch divided by
the initial number of embryos, and requires that both measurements are made
accurately. The number of embryos needs to be estimated before predation and
oophagy compromise the count. The number of hatchings needs to be esti-
mated just before hatching occurs and larvae or young have dispersed. A table
of embryonic stages can be very helpful in determining when hatching is just
about to occur. Care must be taken at this stage because movement of late-stage
embryos can induce hatching. Estimates of hatching or embryonic success typ-
ically take several visits to nest sites.
9.15 Analysis of egg-mass count data

For species where egg-mass counts represent accurate estimates of annual breed-
ing effort by females at a site, then population trends could be modeled using a
variety of regression approaches (Meyer et al. 1998; Alford and Richards 1999;
Loman and Andersson 2007). Information theoretic model selection based
on the Akaike Information Criterion may help identifying the best models.
Population growth rates and density dependence can be tested using relatively
simple formulae: the population growth rate can be measured using N method
(Houlahan et al. 2000):
N log( N 1)t −1 log( N 1)t
where N represents the population size (number of the egg masses) at time t.
Density dependence can be tested by regressing N (growth rate) against the
number of egg masses (see also Hartel 2008 for R. dalmatina and Meyer et al.
1998 for R. temporaria).
If one is interested in annual changes in site occupancy (i.e. the proportion
of ponds with breeding populations), then recent developments by MacKenzie
et al. (2006) could be of interest. This is relevant for species with lower detection
probabilities, particularly for species or areas where it is difficult to determine
total population size based on total egg mass or nest counts.
9.16 Summary
Egg-mass counts and nest counts are both viable techniques that have been used
successfully by herpetologists to monitor population trends of amphibian spe-
cies. Egg mass and nest counts are particularly powerful because one can model
annual breeding effort or true population sizes, rather than crude indices of popu-
lation sizes that are provided by other techniques such as calling surveys. In add-
ition, for conservation biologists interested in assessing factors that might affect
populations, egg counts and nest counts are useful tools. In particular, biologists
have been using egg counts recently to assess the effects of habitat structure at
a variety of spatial scales on estimates of annual reproductive effort (e.g. Egan
2001; Egan and Paton 2004; Skidds et al. 2007). Biologists working in the tropics,
where monitoring information is critical given recent disease events there, might
be particularly interested in incorporating egg counts or nest counts in monitoring
schemes as a supplement to existing calling surveys and visual encounter surveys.
9.17 References
Adams, M. H. and Bury, R. B. (2002). The endemic headwater stream amphibians of
the American Northwest: associations with environmental gradients in a large forested
preserve. Global Ecology and Biogeography, 11, 169–78.
Adams, M. H., Bury, R. B., and Swarts, S. A. (1998). Amphibians of the Fort Lewis
Military Reservation, Washington: Sampling techniques and community patterns.
Northwestern Naturalist, 79, 12–18.
Alford, R. A. and Richards, S. J. (1999). Global amphibian declines: a problem in applied
Andreone, F., Cadle, J. E., Cox, N., Glaw, F., Nussbaum, R. A., Raxworthy, C. J., Stuart,
S. N., Vallan, D., and Vences, M. (2005). Species review of amphibian extinction
risks in Madagascar: conclusions from the global amphibian assessment. Conservation
Biology, 19, 1790–1802.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004). Estimating site occupancy
and species detection probability parameters for terrestrial salamanders. Ecological
Applications, 14, 692–702.
Banning, J. L., Weddle, A. L., Wahl III, G. W., Simon, M. A., Lauer, A., Walters, R. L.,
and Harris, R. N. (2008). Antifungal skin bacteria, embryonic survival, and communal
nesting in four-toed salamanders, Hemidactylium scutatum. Oecologia, 156, 423–9.
Barton, K. and Rafinski, J. (2006). Co-occurrence of Agile Frog (Rana dalmatina Fitz. in
Bonaparte) with Common Frog (Rana temporaria L.) in breeding sites in southeastern
Poland. Polish Journal of Ecology, 54, 151–7.
Bickford, D. (2004). Differential parental care behaviors of arboreal and terrestrial micro-
hylid frogs from Papua New Guinea. Behavioral Ecology and Sociobiology, 55, 402–9.
Breitenbach, G. L. (1982). The frequency of joint nesting and solitary brooding in the
salamander, Hemidactylium scutatum. Journal of Herpetology, 16, 341–6.
Brodman, R. (1995). Annual variation in breeding success of two syntopic species of
Ambystoma salamanders. Journal of Herpetology, 29, 111–13.
Buckley, J. and Beebee, T. J. C. (2004). Monitoring the conservation status of an endan-
gered amphibian: the natterjack toad Bufo calamita in Britain. Animal Conservation,
7, 221–8.
Bury, R. B. and Adams, M. J. (2000). Inventory and monitoring of amphibians in North
Cascades and Olympic National Parks, 1995–1998. Final Report. Forest and Rangeland
Ecosystem Science Center, Corvallis, OR.
Calhoun, A. J. K., Walls, T. E., Stockwell, S. S., and McCollough, M. C. (2003).
Evaluating vernal pools as a bases for conservation strategies: a Maine case study.
Wetlands, 23, 70–81.
Chalmers, R. J. and Loftin, C. S. (2006). Wetland and microhabitat use by nesting four-
toed salamanders in Maine. Journal of Herpetology, 40, 478–85.
Cooke, A. S. (1985). The deposition and fate of spawn clumps of the common frog
Rana temporaria at a site in Cambridgeshire, 1971–1983. Biological Conservation, 32,
165–87.
Corn, P. S. and Livo, L. J. (1989). Leopard frog and wood frog reproduction in Colorado
and Wyoming. Northwestern Naturalist, 70, 1–9.
Corser, J. D. (2001). Decline of disjunct green salamander (Aneides aeneus) populations
in the southern Appalachians. Biological Conservation, 97, 119–26.
Corser, J. D. and Dodd Jr, C. K. (2004). Fluctuations in a metapopulation of nesting four-
toed salamanders, Hemidactylium scutatum, in the Great Smoky Mountains National
Park, USA, 1999–2003. Natural Areas Journal, 24, 135–40.
Crochet, P.-A., Chaline, O., Cheylan, M., and Guillaume, C. P. (2004). No evidence
of general decline in an amphibian community of Southern France. Biological
Conservation, 119, 297–304.
Crouch, W. B. and Paton, P. W. C. (2000). Using egg mass counts to monitor wood frog
populations. Wildlife Society Bulletin, 28, 895–901.
Crouch, W. B. and Paton, P. W. C. (2002). Assessing the use of call surveys to monitor
breeding anurans in Rhode Island. Journal of Herpetology, 36, 185–92.
DeAlmeida Prado, C. P., Uetanabaro, M., and Lopes, F. S. (2000). Reproductive strat-
egies of Leptodactylus chaquensis and L. podicipinus in the Pantanal, Brazil. Journal of
Dodd Jr, C. K. (2003). Monitoring amphibians in Great Smoky Mountains National Park.
Circular Number 1258. US Geological Survey, Tallahassee, FL.
Doody, J. S. and Young, J. E. (1995). Temporal variation in reproduction and clutch mor-
tality of leopard frogs (Rana utricularia) in South Mississippi. Journal of Herpetology,
29, 614–16.
Driscoll, D. A. (1998). Counts of calling males as estimates of population size in
the endangered frogs Geocrina alba and G. vitellina. Journal of Herpetology, 32,
475–81.
Egan, R. S. (2001). Within-pond and Landscape-level Factors Influencing the Breeding
Effort of Rana sylvatica and Ambystoma maculatum. MSc thesis, University of Rhode
Island, Kingston, RI.
Egan, R. S. and Paton, P. W. C. (2004). Within-pond parameters affecting oviposition by
wood frogs and spotted salamanders. Wetlands, 24, 1–13.
Elmberg, J. (1990). Long-term survival, length of breeding season, and operational sex
ratio in a boreal population of common frogs, Rana temporaria L. Canadian Journal of
Zoology, 68, 121–7.
Felton, A., Alford, R. A., Felton, A. M., and Schwarzkopf, L. (2006). Multiple mate
choice criteria and the importance of age for male mating success in the microhylid
frog, Cophixalus ornatus. Behavioral Ecology and Sociobiology, 59, 786–95.
Ficetola G. F., Valota, M., and de Bernardi, F. (2006). Temporal variability of spawning
site selection in the frog Rana dalmatina: consequence for habitat management. Animal
Biodiversity and Conservation, 29, 157–63.
Freda, J., Sadinski, W. J., and Dunson, W. A. (1991). Long term monitoring of amphib-
ian populations with respect to the effects of acidic deposition. Water, Air, and Soil
Pollution, 55, 445–62.
Gamradt, S. C. and Kats, L. B. (1997). Impact of chaparral wildfire-induced sedimen-

tation on oviposition of stream-breeding California newts (Taricha torosa). Oecologia,
110, 546–9.
Gollmann, G., Gollmann, B., Baumgartner, C., and Waringer–Löschenkohl, A. (2002).
Spawning site shifts by Rana dalmatina and Rana temporaria in response to habitat
change. Biota, 3, 35–42.
Gottsberger, B. and Gruber, E. (2004). Temporal partitioning of reproductive activity in
a neotropical anuran community. Journal of Tropical Ecology, 20, 271–80.
Grant, E. H. C., Jung, R., Nichols, J., and Hines, J. (2005). Double-observer approach
to estimating egg mass abundance of pool-breeding amphibians Wetlands Ecology and
Management, 13, 305–20.
Griffith, R. A. and Raper, S. J. (1994). How many clumps are there in a mass of frog
spawn? British Herpetological Bulletin, 50, 14–17.
Harris, R. N. (2005). Hemidactylium scutatum. In M. Lannoo (ed.), Amphibian Declines:
the Conservation Status of United States Species, pp. 780–1. University of California
Press, Berkeley, CA.
Harris, R. N. (2008). Body condition and order of arrival affect cooperative nesting
behaviour in four-toed salamanders Hemidactylium scutatum. Animal Behaviour, 75,
229–33.
Harris, R. N. and Ludwig, P. M. (2004). Resource level and reproductive frequency in
female four-toed salamanders, Hemidactylium scutatum (Caudata: Plethodontidae).
Ecology, 85, 1585–90.
Harris, R. N., Hames, W. W., Knight, I. T., Carreno, C. A., and Vess, T. J. (1995).
An experimental analysis of joint nesting in the salamander Hemidactylium scutatum
(Caudata: Plethodontidae): the effects of population density. Animal Behaviour, 50,
1309–16.
Hartel, T. (2008). Weather conditions, breeding date and population fluctuation in Rana
dalmatina from Central Romania. Herpetological Journal, 18, 40–4.
(1994). Measuring and Monitoring Biological Diversity: Standard Methods for
Hobel, G. (1999). Facultative nest construction in the gladiator frog Hyla rosenbergi
(Anura: Hylidae). Copeia, 1999, 797–801.
Houlahan, J. E., Findlay, C. S., Schmidt, B. R., Meyer, A. H., and Kuzmin, S. L.
(2000). Quantitative evidence for global amphibian population declines. Nature,
412, 499–500.
Icochea, J., Quispitupac, E., Portilla, A., and Ponce, E. (2004). Framework for assessment
and monitoring of amphibians and reptiles in the Lower Urubamba Region, Peru.
Environmental Monitoring Assessment, 76, 55–67.
Jensen, J. B., Bayley, M. A., and Blankenship, E. L. (2003). The relationship between
breeding by the gopher frog Rana capito (Amphibia: Ranidae) and Rainfall. American
Midland Naturalist, 150, 185–90.
Kats, L. B. and Sih, A. (1992). Oviposition site selection and avoidance of fish by stream-
side salamanders (Ambystoma barbouri). Copeia, 1992, 468–73.
Kuramoto, M. (1978). Correlations of quantitative parameters of fecundity in amphib-
ians. Evolution, 32, 287–96.
Kusano, T., Sakai, A., and Hatanaka, S. (2005). Natural egg mortality and clutch size
of the Japanese treefrog Rhacophorus arboreus (Amphibia: Rachophoridae). Current
Lips, K. R., Reaser, J. K., Young, B. E., and Ibáñez, R. (2001). Amphibian monitoring in
Latin America: a protocol manual. Society for the Study of Amphibians and Reptiles,
Herpetological Circular No. 30, 1–115.
Loman, J. and Andersson, G. (2007). Monitoring brown frogs Rana arvalis and Rana
temporaria in 120 south Swedish ponds 1989–2005. Mixed trends in different habitats.
Biological Conservation, 35, 46–56.
MacKenzie, D. I., Nichols, J. D., Royle, J. A., Pollock, K. H., Bailey, L. L., and Hines, J. E.
(2006). Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species
Occurrence. Elsevier Press, Amsterdam.
Marsh, D. M. (2001). Fluctuations in amphibian populations: a meta-analysis. Biological
Martin, A. A. and Cooper, A. K. 1972. The ecology of terrestrial anuran eggs, genus
Crinia (Leptodactylidae). Copeia, 1972, 163–9.
Mazerolle, M. J., Bailey, L. L., Kendall, W. L., Royle, J. A., Converse, S. J., and Nichols, J. D.
(2007). Making great leaps forward: accounting for detectability in herpetological field
studies. Herpetologica, 41, 672–89.
Meyer, A. H., Schmidt, B. R., and Grossenbacher, K. (1998). Analysis of three amphib-
ian populations with quarter-century long time-series. Proceedings of the Royal Society of
London Series B Biological Sciences 265, 523–8.
Muths, E., Corn, P. S., Pessier, A. P., and Green, D. E. (2003). Evidence for disease-
related amphibian decline in Colorado. Biological Conservation, 110, 357–65.
Neckel-Oliveira, S. and Wachlevski, M. (2004). Predation on the arboreal eggs of three
species of Phyllomedusa in Central Amazônia. Journal of Herpetology, 38, 244–8.
Nichols J. D., Hines, J. E., Sauer, J. R., Fallon, F., Fallon, J., and Heglund, P. J. (2000).
A double-observer approach for estimating detection probability and abundance from
avian point counts. Auk, 117, 393–408.
Parmelee, J. R., Knutson, M. G., and Lyon, J. E. (2002). A Field Guide to Amphibian
Larvae and Eggs of Minnesota, Wisconsin, and Iowa. Information and Technology
Report USGS/BRD/ITR-2002-0004. US Geological Survey, Washington DC.
Paton, P. W. C. and Crouch, W. B. (2002). Using the phenology of pond-breeding
amphibians to develop conservation strategies. Conservation Biology, 16, 194–204.
Pechmann, J. H. K., Scott, D. E., Gibbons, J. W., and Semlitsch, R. D. (1989). Influence
of wetland hydroperiod on diversity and abundance of metamorphosing juvenile
amphibians. Wetland Ecology and Management, 1, 3–11.
Peterson, M. G. (2000). Nest, but not egg, fidelity in a terrestrial salamander. Ethology,
106, 781–94.
Petranka, J. W. (1998). Salamanders of the United States and Canada. Smithsonian
Petranka, J. W. and Holbrook, C. T. (2006). Wetland restoration for amphibians: should
local sites be designed to support metapopulations or patchy populations? Restoration
Ecology, 14, 404–11.
Petranka, J. W., Smith, C. K., and Scott, A. F. (2004). Identifying the minimal demographic
unit for monitoring pond-breeding amphibians. Ecological Applications, 14, 1065–78.
Regester, K. J. and Woosley, L. B. (2005). Marking salamander egg masses with Visible
Fluorescent Elastomer: retention time and effect on embryonic development. American
Midland Naturalist, 152, 52–60.
Richards, S. J. and Alford, R. A. (1992). Nest construction by an Australian rainforest
frog of the Litoria lesueuri complex (Anura: Hylidae). Copeia, 1992, 1120–3.
Richter, S. C. (2000). Larval caddisfly predation on the eggs and embryos of Rana capito
and Rana sphenocephala. Journal of Herpetology, 34, 590–3.
Richter, S. C., Young, J. E., Johnson, G. N., and Seigel, R. A. (2003). Stochastic variation
in reproductive success of a rare frog, Rana sevosa: implications for conservation and for
monitoring amphibian populations. Biological Conservation, 111, 171–7.
Roberts, W. E. (1994). Explosive breeding aggregations and parachuting in a neotropical
frog, Agalychis saltator (Hylidae). Journal of Herpetology, 28, 193–9.
Rödel, M. O. and Ernst, R. (2004). Measuring and monitoring amphibian diversity in
tropical forests. I. An evaluation of methods with recommendations for standardiza-
tion. Ecotropica, 10, 1–14.
Seale, D. (1982). Physical factors influencing oviposition by the wood frog, Rana sylvatica,
in Pennsylvania. Copeia, 1982, 627–35.
Sherman, C. K. and Morton, M. L. (1993). Population declines of Yosemite Toads in the
eastern Sierra Nevada of California. Journal of Herpetology, 27, 186–98.
Skidds, D. E., Golet, F. C., Paton, P. W. C., and Mitchell, J. C. (2007). Habitat correlates
of reproductive effort in wood frogs and spotted salamanders in an urbanizing water-
shed. Journal of Herpetology, 41, 439–50.
Sofianidou, T. S. and Kyriakopoulou-Sklavounou, P. (1983). Studies on the biology of
the frog, Rana dalmatina, Bonap. during the breeding season in Greece. Amphibia-
Reptilia, 4, 125–36.
Stewart, M. M. and Pough, F. H. (1983). Population density of tropical forest frogs: rela-
tion to retreat sites. Science, 221, 570–2.
Storfer, A. (2003). Amphibian declines: future directions. Diversity and Distributions, 9,
151–63.
Summers, K., McKeon, C. S., and Heying, H. (2006). The evolution of parental care and
egg size: a comparative analysis in frogs. Proceedings of the Royal Society of London Series
B Biological Sciences 273, 687–92.
Thurley, T. and Bell, B. D. (1994). Habitat distribution and predation on a western popu-
lation of terrestrial Leiopelma (Anura: Leiopelmatidae) in the northern King Country,
New Zealand. New Zealand Journal of Zoology, 21, 431–6.
Townsend, D. S. (1989). The consequences of microhabitat choice for male reproductive
success in a tropical frog (Eleutherodacylus coqui). Herpetologica, 45, 451–8.
Townsend, D. S. and Stewart, M. M. (1994). Reproductive ecology of the Puerto Rican
frog Eleutherodactylus coqui. Journal of Herpetology, 28, 34–40.
University of Canberra (2003). Survey Standards for Australian Frogs. Final Report.
Applied Ecology Research Group, Canberra.
Veith, M., Lötters, S., Andreone, F., and Rödel, M.-O. (2004). Measuring and monitor-
ing amphibian diversity in tropical forests. II. Estimating species richness from stand-
ardized transect censing. Ecotropica, 10, 85–99.
Vonesh, J. (2000). Dipteran predation on the arboreal eggs of four Hyperolius frog species
in western Uganda. Copeia, 2000, 560–6.
Wahl III, G. W., Harris, R. N., and Nelms, T. (2008). Nest site selection and embryonic
survival in four-toed salamanders, Hemidactylium scutatum (Caudata: Plethodontidae).
Waringer-Löschenkohl, A. (1991). Breeding ecology of Rana dalmatina in lower Austria:
a 7-year study. Alytes, 9, 121–34.
Weir, L. A., Royle, J. A., Nanjappa, P., and Jung, R. E. (2005). Modeling anuran detection
Wells, K. D. (2007). The Ecology and Behaviour of Amphibians. University of Chicago
Press, Chicago, IL.
Werner, E. E., Yurewicz, K. L., Skelly, D. K., and Relyea, R. A. (2007). Turnover in
an amphibian metacommunity: the role of local and regional factors. Oikos, 116,
1713–25.
Williams, L. R. (2005). Restoration of ponds in a landscape and changes in Common
Frog (Rana temporaria) populations, 1983–2005. Herpetological Bulletin, 94, 22–9.
Williams, B. K., Nichols, J. D., and Conroy, M. J. (2002). Analysis and Management of
Animal Populations: Modeling, Estimation and Decision Making. Academic Press, San
Diego, CA.
Windmiller, B. S. (1996). The Pond, the Forest, and the City: Spotted Salamander Ecology
and Conservation in a Human-dominated Landscape. PhD dissertation, Tufts University,
Medford, MA.
Zheng, Y. and Fu, J. (2006). Making a doughnut-shaped egg mass: oviposition behaviour
of Vibrissaphora boringiae (Anura: Megophryidae). Amphibia-Reptilia, 28, 309–11.
Appendix 9.1 Oviposition strategies of anuran families of the world, with comments on
whether egg counts could be used monitor their populations (modified from Wells 2007).
Family Distribution Deposition site Eggs could be counted?
Allophrynidae South America Ponds No?1

Arthroleptidae Africa Terrestrial No?
Ascaphidae Northwestern Strings attached to Maybe?2
USA rocks in streams
Astylosternidae Africa Rivers and streams No?
Bombinatoridae Europe and Ponds/streams Yes?
Asia
Brachycephalidae Northern Terrestrial No
South America
Bufonidae Worldwide, Long strings in still Small populations of
except water some species
Australia
Centrolenidae Central Arboreal over water Yes?
America
Dendrobatidae South and Terrestrial eggs carried No?
Central to water by adults
America
Discoglossidae Europe, Terrestrial (males No
Middle East, carry eggs), some
North Africa breed in ponds
Helephrynidae Africa Under rocks in No?
streams
Hemiphractinae South America Females carry eggs No
Hemisotidae Africa Terrestrial near ponds No
Hylidae Worldwide In ponds and on Some species3, 4,5
vegetation
Hyperoliidae Africa Terrestrial and aquatic No?
eggs
Leiopelmatidae New Zealand Terrestrial and water- Some species6
filled depressions
Leptodactyidae Southern USA, Ponds, foam nests in Some species?7
Central and water or near water
South America
Mantellidae Madagascar In ponds or streams No?8
Megophryidae Asia Pools, under rocks in No?9
streams
Appendix 9.1 Continued

Family Distribution Deposition site Eggs could be counted?
Microhylidae Worldwide Variety of terrestrial, Yes?10,11

arboreal, and aquatic
habitat
Myobatrachidae Australia and Aquatic foam nests or No?12
New Guinea burrows or terrestrial
or ponds
Pelodytidae Europe and Strings in ponds Yes13
Asia
Petropedetidae Africa Ponds and terrestrial No?14
Pipidae South Strictly aquatic; in No?15
American and some species females
Africa incubate eggs on back
Ranidae Worldwide Aquatic Yes (see Table 9.1)16
Rhacophoridae Asia, Africa Foam nests in trees, Yes? 17
on rocks, and surface
of surface
Rhinodermatidae South America Terrestrial, No
Rhinophrynidae Central Shallow ponds No?
America
Sooglossidae Seychelles Terrestrial No?
1
Gottsberger and Gruber (2004); 2Bury and Adams (2000); 3Adams et al. (1998); 4Roberts (1994);
5
Neckel-Oliveira and Wachlevski (2004); 6Thurley and Bell (1994); 7DeAlmeida Prado et al. (2000),
8
Andreone et al. (2005); 9Zheng and Fu (2006); 10Bickford (2006); 11Summers et al. (2006);
12
Driscoll (1998); 13Crochet et al. (2004); 14Veith et al. (2004); 15Icochea et al. (2004); 16Freda et al.
(1991); 17Kusano et al. (2005).
10
Dietary assessments of adult amphibians
Mirco Solé and Dennis Rödder
10.1 Introduction
Dietary information is essential to understand amphibian life history, popu-
lation fluctuations, and the impact of habitat modification, and to develop
conservation strategies (Anderson 1991). In ecology, trophic relationships
represent a functional connection between different taxa and are a key
subject in many aut- and synecological studies. Studies on diets are essen-
tial for assessments of energy flow and food webs in ecological communi-
ties. Variables involved in an animal’s diet are seldom obvious and include
behavioral, physiological, and morphological features of both predator and
prey (Simon and Toft 1991). Therefore, dietary information from locally
imperilled species is a necessary component in designing management and
conservation programs.
Traditionally, amphibians are described as generalist predators feeding
mainly on arthropods, but mollusks, annelids, and even small vertebrates
are also frequently consumed. Amphibians are able to distinguish between
different prey types allowing for different degrees of feeding specialization
(Freed 1982). Dietary specialization is often associated with morphological,
physiological, and behavioral characteristics that facilitate location, identifi-
cation, capture, ingestion, and digestion of prey items. Feeding dynamics can
vary considerably between species, but differences between the three major
amphibian lineages and between species feeding on land and in water become
obvious. Therefore, the feeding mechanisms and dietary specialties of the dif-
ferent groups are discussed separately (for amphibian larvae see Chapter 5).
10.1.1 Adult anurans

Anurans are carnivores and their main diet consists of arthropods. Parts of
plants or fruits are accidentally consumed with other food items by adult
amphibians, although exceptions have been documented, such as the hylid

frog Xenohyla truncata (Silva and Britto-Pereira 2006). In most anurans, prey
capture is enhanced by the use of projectile tongues fi xed at the anterior part
of the mouth. The posterior part is commonly free, and the whole tongue
can be thrust out in order to capture prey items. Among anurans, the family
of the predominately aquatic Pipidae lack tongues and uses suction feeding.
The species are characterized by broad elements of the hyobranchial skeleton,
especially the ceratohyals and ceratobranchials. Sokol (1969) observed that
the small pipid Hymenochirus boettgeri depressed its hyobranchium prior to
mandibular depression, so that maximum suction is achieved as soon as the
mouth is opened. The larger pipid genera Pipa in South America and Xenopus
in Africa augment buccal suction by stuffing their prey into their mouth with
their forelimbs.
Differences in feeding behavior among frogs with different diets have been
observed, encompassing a continuum between sit-and-wait and opportunistic
search strategies. The composition of the diet can reflect the feeding strat-
egy to some degree. Whereas sit-and-wait foragers commonly consume larger,
mobile prey and only a few food items per time unit, opportunistic searchers
consume greater numbers of smaller food items. These behavioral patterns
are also reflected in morphological properties of the species. Emerson (1985)
found a relationship between the skull shapes of frogs and their diet, whereby
frogs that eat relatively small, slow prey have relatively long jaws and an asym-
metrical feeding cycle; time to capture prey is less than the time to bring prey
into the mouth. On the other hand, snail-eating frogs have lower jaws with a
relatively long distance from the insertion point of the adductor muscle to the
jaw articulations. Two main diet patterns in anurans were identified by Toft
(1980a, 1981, 1985). They comprise “ant specialists” eating slow-moving,
strongly chitinized arthropods such as ants, termites, and mites, and “non-ant
specialists” eating generally more variable and less chitinous arthropods (e.g.
spiders and grasshoppers).
10.1.2 Adult caudata

Terrestrial salamanders typically feed by capturing their prey with a large, moist,
sticky tongue. Although predominately carnivorous, several species of the genus
Siren include larger plant parts in their diet (Duellman and Trueb 1994). Some
neotropical salamanders exhibit lower prey diversity than most temperate-zone
terrestrial and aquatic species, a difference that might be explained by the dis-
proportionately high contribution of ants in the diets of the neotropical species
(Anderson and Mathis 1999).
10 Dietary assessments of adult amphibians | 169
10.1.3 Adult gymnophiona

Despite being widely distributed in the tropics, only few dietary studies have
been performed on caecilians (O’ Reilly 2000). Most dietary studies of terres-
trial and aquatic caecilians suggest that they are generalized carnivores. Kupfer
et al. (2005) studied the diet of both larval and adult Ichthyophis cf. kohtaoensis
from Thailand. The larvae feed exclusively on aquatic invertebrates, preferring
benthic prey. The adults consume mostly soil invertebrates such as earthworms,
ants, and termites. Ontogenetic dietary shifts and the trophic position of caecil-
ians are more similar to salamanders than frogs with biphasic life cycles.
10.2 Methods to obtain prey items:

historical overview
Most accounts of amphibian feeding are anecdotal and involve only a few taxa.
As a result, much is unknown about prey selection and foraging strategies.
The diet of amphibians has commonly been studied by sacrificing the ani-
mals and dissecting their stomachs. Technically, this is the easiest method and,
when applied to series of preserved specimens housed in museum collections,
ethically justified. Prey items can be obtained in good condition, facilitating
identification and measurements. Today, the euthanasia of large samples of
study animals for dietary analyses is ethically questionable because alternative
methods are available. One of these alternative methods is stomach flushing.
It avoids killing large number of animals, and has been used in herpetological
studies since the 1970s. For example, Fraser (1976) used this method to study
salamander diets, and it was used in later studies of both frogs and reptiles
(Legler 1977; Legler and Sullivan 1979). Most researchers anesthetize the ani-
mals prior to stomach flushing, but this occasionally has resulted in mortality
(Joly 1987). Despite early attempts at alternative study methods, most dietary
studies continued to rely upon freshly killed specimens, probably because the
description of stomach-flushing techniques proved rather inadequate and diffi-
cult to replicate. Solé et al. (2005) published a protocol clearly explaining which
materials were needed, and which steps had to be followed to achieve successful
stomach flushing. Their protocol has since been used in dietary research on
different anurans (e.g. Miranda et al. 2006; Mahan and Johnson 2007; Solé
and Pelz 2007; Wang et al. 2008). Wu et al. (2007) compared data on stomach
dissection with those obtained by stomach flushing to assess dietary diversity,
and found no significant differences, tending to validate the flushing method.
They stated that “our study indicates that the result of stomach flush is accurate
in the diet analysis. Stomach flush should be used more widely and stomach
dissection should be thrown away ultimately.” Since prey size, composition,
and diversity are often correlated with the sex and size of the predator (Lima
and Magnusson 2000), it is important to note the snout–vent length and the
gender of each specimen.
10.2.1 Stomach flushing: materials

A complete flushing set consists of two containers, a spatula, forceps, two
syringes (20 ml for small and 60 ml for large frogs), an infusion tube (for
small frogs) or a rubber or PVC hose (commonly used for aquarium pumps),
small airtight vials, and 70% ethanol solution (Figure 10.1). We recommend
acquiring syringes with screw-on threading and infusion tubes with internal
threading to prevent the detachment of the tube from the syringe during the
flushing procedure. For use directly in the field, all materials can be stored in
the larger container. Spring water or previously filtered pond water can be used
as a flushing solution.
Fig. 10.1 Complete flushing set consisting of two containers, a spatula, forceps, two
syringes, an infusion tube (for small frogs) or a rubber or PVC hose, small airtight vials,
and 70% ethanol solution.
10.2.2 When should sampling be conducted?

Feeding-activity patterns of the target species should be considered to avoid a
bias resulting from digestion, the rate of which is directly related to environmen-
tal temperature. Amphibians should be sampled during their foraging activity
period or shortly thereafter, as digestion compromises accurate identification
of prey items. For example, most frogs are crepuscular or nocturnal (Duellman
and Trueb 1994), and searching for them is easier at night due to the reflective
aspect of the eyes of most frogs when illuminated with a head lamp (Rocha et al.
2000). We recommend sampling frogs about 2–4 h after nightfall. During this
time most frogs have already managed to find some food, and the chance to
recover unfragmented prey items is much higher than if sampling and flushing
are performed the next morning.
10.2.3 How to perform the flush

Before handling the frog, all material should be ready to be used. One con-
tainer should be filled with flushing solution and the other positioned to collect
the stomach contents. The syringe should be filled with flushing solution and
attached to the tube. The specimen should be held safely by fixing its forelimb
with one hand, while the syringe is held in the other. As both hands are in use,
the spatula used to open the frog’s mouth should be held between one’s teeth.
The spatula should be inserted gently at the hind part of the mouth and then
turned at a 45° angle. The tube is then inserted into the mouth and down the
esophagus to the stomach (Figure 10.2). The pyloric end of the stomach can eas-
ily be detected, but gentle care is highly recommended throughout the process.
Once the tube is in the correct position, the entire syringe contents are flushed
into the stomach. The right amount of force is dependent on the size of the
flushed frog. We recommend starting using little pressure and increase the pres-
sure steadily until stomach contents start pouring out of the mouth. The empty
syringe is then separated from the tube, which is kept inside the stomach, refilled
with flushing solution, and attached to the tube again in order to allow a second
flush. This procedure should be repeated as long as stomach contents are forced
out. When no more stomach contents appear, the frog should be flushed one last
time. The flushing solution containing the stomach contents is then decanted
into a sieve. Items are then picked up by using a forceps and stored in a vial in
70% ethanol or formalin. The latter has the advantage that the prey items are
not leached as much compared to when ethanol is used. On the other hand
storage in ethanol allows subsequent DNA analyses and formalin does not. For
specimens with very small prey, such as mites and Collembolans, gauze with a
mesh size of at least 6 μm should be used instead of the sieve. Stomach contents
Fig. 10.2 Stomach flushing applied to Trachycephalus mesophaeus.
can be directly washed with 70% ethanol from the gauze into a small vial. Frogs
should be released where they were captured as soon as possible after stomach
flushing to facilitate a return to normal foraging activity and behavior.
10.3 Data analysis

Once the prey items are obtained, the most comfortable way to identify them
is to spread each sample separately on a petri dish containing water and then
examine them under a stereo microscope. Prey identification is commonly car-
ried out to ordinal level, but prey categories must not all correspond to taxo-
nomic orders. Most authors, for example, separate the Order Hymenoptera into
the family Formicidae (ants) and non-Formicidae because in many amphibian
species ants make up a large part of the diet. Depending on the degree of diet-
ary specialization and the question at hand it might be useful to identify prey
items to the lowest taxonomic level as possible. If identification keys for the
local fauna exist, these can be used for taxonomic identification of prey. In areas
with a lack of such data, general ordinal keys can be used (e.g. Triplehorn and
Johnson 2005) or prey items can be sent to colleagues specialized in specific
groups. Some authors categorize the prey items into ecological guilds, such as
‘rather active’ or ‘sedentary’ taxa, or flying, ground-dwelling, or burrowing spe-
cies. These methods facilitate assessments of foraging strategies in the absence of
behavioral information, since the quantity and volume of prey items belonging
to different guilds allow researchers to draw conclusions about the feeding strat-
egy of the amphibian predator.
Even if researchers manage to capture amphibians shortly after they have
started foraging, and even if the flushing procedure is applied shortly thereafter,
researchers frequently find that parts of the stomach contents have already been
digested. Soft tissues generally are digested more rapidly than strongly chitinized
parts, thus making prey items break into fragments. Hirai and Matsui (2001)
noticed that certain specific parts of prey remain intact even after digestion has
started. Examples are the heads of ants or wasps, abdomen and elytra of beetles,
thoraxes of flies, moths, beetles, or Orthopterans (grasshoppers, crickets, katy-
dids), forewings of flies and bees, and the saltatorial legs of Orthopterans. Hirai
and Matsui (2001) calculated regression formulae between body parts and total
body length, which can be used as a rough estimate of prey size. To be complete,
however, information on the size of undamaged items is required for a thorough
assessment of prey use.
In most dietary studies, information about the number, frequency, and vol-
ume and/or mass of the prey items of each taxonomic or ecological group is
presented. These data allow direct comparison between different prey groups.
The mass of prey items can be assessed with an analytic balance. Here, storage
in formalin is superior to storage in ethanol since the later dehydrates the prey
items. Furthermore, since stomach contents are commonly wet, it is important
to remove excess moisture from the prey item before weighing to avoid bias.
The best way would be to store the prey items in a drying oven at approximately
50°C until a constant dry mass is achieved. Information about prey number,
frequency, and mass are relatively easy to obtain, as there are a number of differ-
ent techniques used to measure the volume of prey items. Volume and mass of
prey are directly correlated with the amount of energy they can provide; thus,
this information is most appropriate in combination with the identification of
important prey groups. Most researchers attempt to reconstitute volumes from
linear measures using formulae for geometric shapes; body appendices of prey
items should be ignored when measurements are taken. The most commonly
used formula is the one for ellipsoid bodies (Colli and Zamboni 1999):
2
4 L ⎛ W ⎞
V ⎜ ⎟
3 2⎝ 2 ⎠
where V is volume, L is length, and W is width of prey item. This formula is

also referred to as prolate spheroid (Matori and Aun 1994; Vitt et al. 1999).
Other formulas used to estimate volume are the formula for a parallelepiped
(V LWH) (Schoener 1967; Bergallo and Magnusson 1999) and for a cylinder
(Werner et al. 1995):
2
⎛W ⎞
V L ⎜ ⎟
⎝ 2⎠
The volume of prey items also can be measured directly by introducing them
into a measuring cylinder graduated in intervals of 0.1 ml. Water is added from a
pipette graduated in intervals of 0.01 ml until the item is completely covered and
the water has reached the next highest graduation in the cylinder. The volume
can then be taken as the water level in the cylinder minus the amount added
with the pipette (Magnusson et al. 2003; Cuello et al. 2006).
10.3.1 Measuring prey availability and electivity

Although Schoener (1974) recommended that the study of a species’ trophic
niche should include prey availability, only a few studies have done so (Toft
1980a, 1981; MacNally 1983; Juncá and Eterovick 2007). There are many
different methods used to measure prey availability, and techniques should be
selected in keeping with the different foraging microhabitats, such as leaf litter
and arboreal species. Traps are usually either mechanical or based on attractants
such as food or pheromones; the latter should be avoided because of sampling
biases.
To assess prey availability in a leaf-litter habitat, Maneyro and da Rosa (2004)
used pitfall traps consisting of 300 ml cans with a solution of liquid soap and
10% formalin. This formulation may prove useful in sampling ground-dwelling,
mobile prey items. Soap is added to prevent the escape of arthropods, and it
reduces the surface tension of the solution, facilitating the drowning of small
arthropods such as Collembolans. Furthermore, it makes the bucket walls of
the trap slippery. Instead of formalin (which may be legally prohibited for use
under natural settings and the use of which is often considered unethical in field
research), ethanol may be substituted. Pitfalls also may be fitted with modified
lids or funnels to prevent escape.
Pitfall traps may overestimate more mobile fauna (Ausden 1996). On the
other hand, it is reasonable to assume that such bias would not be misleading in
the assessment of prey availability since prey mobility may increase the detection
probability by amphibians which use vision to locate prey (Duellman and Trueb
1994). Juncá and Eterovick (2007) recommended using Berlese–Tullgren funnel
traps (Brower et al. 1998) in addition to pitfall traps and combining the results.
These traps sample sedentary taxa in addition to mobile fauna. Both methods
underestimate flying prey items.
In grassy or herbaceous habitats, the composition and abundance of prey
items can be estimated along transects using linear sweeps with insect nets.
Each sweep sample should be transferred into a Berlese–Tullgren trap for at
least 30 min to separate prey items from litter (leaves, seeds, sticks). Properly
conducted, this method can provide a comprehensive assessment of inverte-
brate availability. It should be noted that highly mobile flying insects such as
Odonates (dragonflies) might be underrepresented. Furthermore, the time and
temperature at which the sweeps are conducted can have a strong influence on
the result depending on the activity periods of the insects.
Arboreal habitat can be sampled most easily using fogging techniques.
Applying insecticides through fogging is widely used in pest control, and a wide
range of commercial fogging machines is available. Fumigated insects falling
from tree canopies are collected on plastic sheets laid out on the ground or by
plastic trays or aluminum funnels (for a short review, see Adis et al. 1998).
Insecticide fogging has become established as a standard technique since the
1980s for assessing arthropod diversity, especially in tropical forest canopies
(Adis et al. 1998). A variety of “knockdown” insecticides have been tested, but
natural pyrethrum is now established as a standard as it allows for collection of
live specimens and its biodegradability facilitates application in ecologically sen-
sitive habitats. Adis et al. (1998) suggested recommendations to standardize fog-
ging procedures. When applying fogging procedures, researchers must consider
that the entire vertical habitat that is sampled may overestimate the prey avail-
able to amphibians, which inhabit only parts of the vertical habitat structure.
Aquatic habitats can be sampled using a fine mashed net, such as the D-frame
kick net or surber stream-bottom sampler. The D-frame kick net is either cone-
or bag-shaped for the capture of organisms and commonly used by lying the
spine of the net firmly on the ground and hauling it a specific number of times
along a specific distance. This type of net is easy to transport and can be used in
a variety of habitat types. However, the D-frame kick net must have a defined or
delimited area that is sampled/kicked in order to compare results obtained from
different sites or samples. Cao et al. (2005) investigated the performance and
comparability between surber and D-frame kick-net samples and found that
subsamples with the same number of individuals were highly and consistently
comparable between sampling devices. Rocky stream habitats are best sampled
by two persons. One collector should place a fine mashed net so that its bottom
lies on the ground. Another collector should turn over as many rocks as possible
upstream from the net for 1 min. Wiping off clinging insects from rocks and
branches may enhance collection quantities and species diversity.
Once relative abundances of the prey items in the habitat and in the stomach
contents are identified, an electivity index can be calculated, as proposed by
Jacobs (1974):
Rk Pk
D
(Rk Pk ) (2Rk Pk )
whereby Rk is the proportion of prey category k in stomach contents, and Pk is the
proportion of prey category k in the environment. D varies from 1 (complete
selection of preference for prey category k) through 0 (prey category k is taken in
the same proportion as found in the environment) to 1 (prey category k is pre-
sent in the environment but absent in the diet). When interpreting the results,
biases caused by sample methods need to be considered.
10.3.2 Comparing trophic niche structure

A number of indices have been proposed for the identification of (trophic) diver-
sity, niche breadth, and niche overlaps between multiple species. Calculation
of indices can be facilitated using the software EcoSim (N. J. Gotelli and
G. L. Entsminger, version 7.0, available at http://together.net/~gentsmin/
ecosim.htm, 2001). Trophic diversity can be assessed using Hurtubia (1973) or
the Gladfelter–Johnson index (Gladfelter and Johnson 1983). A modification
of this latter index was recently proposed by Cardona (1991) to measure trophic
niche breadth using occurrence frequencies:
B′
∑ p k
100R
where pk is the occurrence frequency of kth prey, is the standard deviation of
the occurrence frequencies, and R is the number of prey species exploited by the
guild.
In order to compare the niche breath between multiple species or changes in
the niche breadth within different seasons, the Shannon–Weaver diversity index
H (Weaver and Shannon 1949; see Chapter 18) can be used:
H ′ ∑ pi In Pi
i
where pi is the relative abundance of each prey category, calculated as the pro-
portion of prey items of a given category to the total number of prey items.
The index of relative importance (IRI) is a measure that reduces bias in

descriptions of animal dietary data. Introduced by Pianka et al. (1971) and
Pianka (1973), it is mainly used currently to describe the diet of fish:
IRI t ( POt ) ( PI t PVt )
where POt is the percentage of occurrence (100 number of stomachs containing

t item/total number of stomachs), PIt is the percentage of individuals (100 total
number of individuals of t in all stomachs/total number of individuals of all taxa
in all stomachs), and PVt is the percentage of volume (100 total volume of indi-
viduals of t in all stomachs/total volume of all taxa in all stomachs). Hart et al.
(2002) suggested that IRI should be applicable to a wide range of taxa. To calcu-
late the trophic niche breadth, the index of Simpson (1949) can be used:
1
B
∑ pi2
where B is Simpson’s index, pi is the fraction of items in the food category i, and
the range is 1 to n.
To compare various diets with a different number of food categories, the
standardized Simpson’s index (BS) form as proposed by Hulbert (1978) can be
used:
B 1
BS
n 1
Here, n is the number of food categories and the index ranges from 0 to 1.
Overlap in trophic niches between two species can be assessed with an index
proposed by May and MacArthur (1972) as modified by Pianka (1974):
n
∑P P ij ik
O jk n
i
n
∑P ∑P
i
ij
2
i
2
ik
Here, Ojk is niche overlap and pij and pik represent the proportions of the ith
resource used by the jth and kth species.
10.3.3 Fatty acid and stable-isotope analyses

Analyses of isotopic composition and fatty acid profiles of tissues can provide
reliable information on diets and trophic status because they reflect assimilation
over long periods of time. Isotope signatures—that is, ratios of heavier (e.g.
15
N and 14C) and lighter (14N and 13C) isotopes—of consumers are generally
similar to those of their food and can thus provide insights into food-web struc-
ture. Fatty acids are essential energy storage lipids in consumers that cannot
be readily synthesized and must be obtained from their food. Differences in
fatty acid composition among available food types are reflected in fatty acid
compositions in consumers; therefore, profiles can be good indicators of diets.
For both techniques, tissue samples (or in the case of small individuals or food
items, the whole organism) are simply collected and preserved in a manner that
does not alter isotopic or fatty acid composition, depending on the analyses to
be performed, and then sent to an analytical laboratory for analyses (for a more
detailed description of these techniques see Chapter 5).
10.3.4 Further biochemical analyses

Some frog species in the neotropics and Madagascar secrete alkaloids for protec-
tion against predators. These so-called poison frogs do not produce the alkal-
oids, but instead obtain them from their insect-rich diet (Daly et al. 1994).
While the diet of neotropical poison frogs is well studied (Daly et al. 1994,
2002), data on the Malagasy frogs was only provided recently by Clark et al.
(2005). These authors extracted alkaloid samples from both Malagasy frogs and
their food. Alkaloid fractions can be extracted and isolated from skin samples
using methanol (one part skin/two parts methanol) (Daly et al. 1992). Possible
methods for biochemical analyses comprise gas chromatography in conjunc-
tion with mass spectrometry and infrared spectroscopy to identify alkaloids
(for detailed descriptions see Daly et al. 1992, 1993, 1994). They found that
Malagasy frogs, like the neotropical species, acquire their alkaloids from the
ants they prey upon. Thirteen of the 16 alkaloids detected in the Malagasy
frogs also exist in insects and frogs in the neotropics. As neither the ants nor the
frogs in these two regions are closely related, these authors suggested that the
evolution of acquisition mechanisms for alkaloids in these ant species was likely
responsible for the subsequent convergent evolution of the frogs that preyed on
them. The researchers also found nicotine, a plant alkaloid, in one Malagasy
frog species, suggesting a possible plant–ant–frog toxin food chain.
10.4 General considerations

Samples should be taken randomly throughout the habitat of the target spe-
cies in order to avoid bias, such as the predominance of ants which would be
expected if all samples were obtained next to an ant hill. Furthermore, sample
sizes are an important issue for dietary analyses. Considering that diet may vary
between the sexes, adequate samples from both genders should be obtained.
When working with stomach contents, researchers should be aware that they
will probably not get a picture of what prey items the focal species feeds upon,
but rather a picture of the items that are hard to digest. Frogs are capable of
digesting prey rapidly: Blanchard’s cricket frogs (Acris crepitans blanchardi) fed
dyed Drosophila (fruit flies) and killed 7 h later had no trace of prey remain-
ing in their stomachs (Johnson and Christiansen 1976). In another study with
A. crepitans immediately preserved frogs had a mean of seven prey items per
stomach, whereas frogs preserved 6 h after capture revealed 3.6 prey items per
stomach (Caldwell 1996). In a preliminary study with 20 Physalaemus lisei,
frogs were fed three different prey categories (Oligochaeta, Curculionid beetles,
and Aranea) at the same time, and four frogs were stomach-flushed every 4 h.
Oligochaeta (worms) could not be detected after 4 h, whereas the pronotum of
spiders was still undigested after 8 h and the beetle elytra were still present after
12 h (M. Solé, unpublished results). In conclusion, researchers will underesti-
mate the prey categories that basically consist of soft tissue (worms, mollusks,
and fly larvae) and overestimate those with strongly keratinized bodies (ants and
beetles). One way in which biologists can mediate this problem is to carefully
choose the sampling period according to the feeding period of each species.
10.4.1 Ontogenetic changes

Diet composition of amphibians may change during ontogeny, since larger indi-
viduals may be able to eat larger prey items, which significantly increases the var-
iety of available prey types. However, Lima (1998) found changes in types and
sizes of prey that were caused other than by a simple passive effect of selection for
larger prey with growth of the frogs. Foraging modes and microhabitat use may
also change during ontogeny, which can affect the composition of available prey
items (Lima and Magnusson 2000). To assess ontogenetic changes in the com-
position of the diet, different snout–vent size classes of the target species can be
compared. The best visualization provides a cluster analysis based on the relative
frequency of the different prey categories. Data must be normally distributed to
meet the assumptions of the statistical test (Zar 1999). If the data do not meet the
required statistical assumptions, they can be transformed (see Chapter 18).
10.4.2 Seasonal changes

Availability and diversity of prey items can change between different seasons
(Houston 1973). Although there are many records concerning feeding habits in
amphibians, few include seasonal patterns. In Amazonian Peru, Toft (1980b)
observed a decrease in arthropod biomass and in frog abundance in the dry

season. However, she argued that frog species and their arthropod prey may
respond differently to seasonal changes, since the principal feeding patterns
of the litter anurans studied did not appear to change seasonally. Yilmaz and
Kutrup (2006) analyzed seasonal changes in the diet of the European common
frog, Rana ridibunda, and found changes in prey diversity between July (62 prey
categories) and August (38). The highest mean prey number was observed in
July. The mean prey size also changed during the year: frogs were more likely
to consume large prey items in June than in the other months. These results
suggested that the monthly variation in diet composition of this species was
correlated with prey availability. Finally, large sample sizes are important when
studying seasonal dietary changes.
10.5 Conclusions
Dietary information is pivotal for successful development of conservation strat-
egies on species level and the understanding of ecosystem function. Unfortunately,
this kind of information is not available for the vast majority of taxa and is often
incomplete. Seasonal variations, ontogenetic shifts, and relationships between
site-specific prey availability, presence of potential competitors, and diet compos-
ition are commonly not addressed. Future studies should focus on these issues.
In the light of global amphibian declines an Amphibian Conservation Action
Plan was formulated by the International Union for Conservation of Nature to
prevent further biodiversity loss and captive breeding programmes are being
developed for the most threatened species (Gascon et al. 2007). Such approaches
depend crucially on autecological knowledge such as information on the diet of
the target species. Here, every kind of information can be valuable.
10.6 Acknowledgments
We are grateful to Ken Dodd, Marian Griffey, and Matt Whiles for helpful com-
ments on the manuscript and to Klaus Riede for fruitful discussions concerning
quantitative insect sampling methods and help with the literature. The work of
DR was funded by the Graduiertenförderung des Landes Nordrhein-Westfalen.
10.7 References
Adis, J., Basset, Y., Floren, A., Hammond, P. M., and Linsenmair, K. E. (1998). Canopy fog-
ging on an overstorey tree- recommendations for standardization. Ecotropica, 4, 93–7.
Anderson, M. T. and Mathis, A. (1999). Diets of two sympatric Neotropical salaman-

ders, Bolitoglossa mexicana and B. rufescens, with notes on reproduction for B. rufescens.
Anderson, S. H. (1991). Managing our Wildlife Resources. Merrill Publishing Co.,
Columbus, OH.
Ausden, M. (1996). Invertebrates. In W. J. Sutherland (ed.), Ecological Census Techniques,
pp. 139–77. Cambridge University Press, Cambridge.
Bergallo, H. G. and Magnusson, W. E. (1999). Effects of climate and food availability on
four rodent species in southeastern Brazil. Journal of Mammalogy, 80, 472–86.
Brower, J. E., Zar, J. H., and von Ende, C. N. (1998). Field and Laboratory Methods for
General Ecology. WCB McGraw-Hill, Boston, MA.
Caldwell, J. P. (1996). The evolution of myrmecophagy and its correlates in poison frogs
(family Dendrobatidae). Journal of Zoology, London, 240, 75–101.
Cao, Y., Hawkins, C. P., and Storey, A. W. (2005). A method for measuring compar-
ability of different sampling methods used in biological surveys: implications for data
integration and synthesis. Freshwater Biology, 50, 1105–15.
Cardona, L. (1991). Measurement of trophic niche breadth using occurrence frequencies.
Journal of Fish Biology, 39, 901–3.
Clark, V. C., Raxworthy, C. J., Rakotomalala, V., Sierwald, P., and Fisher, B. L. (2005).
Convergent evolution of chemical defense in poison frogs and arthropod prey between
Madagascar and the Neotropics. Proceedings of the National Academy of Sciences USA,
102, 11617–22.
Colli, G. R. and Zamboni, D. S. (1999). Ecology of the worm-lizard Amphisbaena alba in
the Cerrado of central Brazil. Copeia, 1999, 733–42.
Cuello, M. E., Bello, T. M., Kun, M., and Ubeda, C. A. (2006). Feeding habits and
their implications for the conservation of endangered semiaquatic frog Atelognathus
patagonicus (Anura, Neobatrachia) in a northwestern Patagonian pond. Phyllomedusa,
5, 67–76.
Daly, J. W., Secunda, S. I., Garraffo, H. M., Spande, T. F., Wisnieski, A., Nishihira, C.,
and Cover Jr, J. F. (1992). Variability in alkaloid profiles in Neotropical poison frogs
(Dendrobatidae). Toxicon, 30, 887–98.
Daly, J. W., Garraffo, H. M., and Spande, T. F. (1993). Amphibian alkaloids. In G. A.
Cordell (ed.), The Alkaloids, vol. 43, pp. 185–288. Academic Press, San Diego, CA.
Daly, J. W., Garraffo, H. M., Spande, T. F., Jaramillo, C., and Rand, A. S. (1994). Dietary
source for skin alkaloids of poison frogs (Dendrobatidae)? Journal of Chemical Ecology,
20, 943–55.
Daly, J. W., Kaneko, T., Wilham, J., Garraffo, H. M., Spande, T. F., Espinosa, A., and
Donnelly, M. (2002). Bioactive alkaloids of frog skin: combinatorial bioprospect-
ing reveals that pumiliotoxins have an arthropod source. Proceedings of the National
Duellman, W. E. and Trueb, L. (1994). Biology of Amphibians. John Hopkins University
Press, Baltimore, MD.
Emerson, S. B. (1985). Skull shape in frogs—correlations with diet. Herpetologica, 41, 177–88.
Fraser, D. F. (1976). Coexistence of salamanders in the genus Plethodon, a variation of the
Santa Rosalia theme. Ecology, 57, 238–51.
Freed, A. N. (1982). A treefrog’s menu: selection for an evening’s meal. Oecologia, 53,
20–6.
Gascon, C., Collins, J. P., Moore, R. D., Church, D. R., McKay, J. E., and Mendelson III,
J. R. (2007). Amphibian Conservation Action Plan - Proceedings: IUCN/SSC Amphibian
Conservation Summit 2005. IUCN, Gland.
Gladfelter, W. B. and Johnson, W. S. (1983). Feeding niche separation in a guild of trop-
ical reef fish (Holocentridae). Ecology, 64, 552–63.
Hart, R. K., Calver, M. C., and Dickman, C. R. (2002). The index of relative import-
ance: an alternative approach to reducing bias in descriptive studies of animal diets.
Wildlife Research, 29, 415–21.
Hirai, T. and Matsui, M. (2001). Attempts to estimate the original size of partly digested
prey recovered from stomachs of Japanese anurans. Herpetological Review, 32, 14–16.
Houston, W. W. K. (1973). The food of the common frog, Rana temporaria, on high
moorland in northern England. Journal of Zoology, 171, 153–65.
Hulbert, S. H. (1978). The measurement of niche overlap and some relatives. Ecology,
59, 67–77.
Hurtubia, J. (1973). Trophic diversity measurement in sympatric predatory species.
Journal of Ecology, 549, 885–90.
Jacobs, J. (1974). Quantitative measurement of food selection: a modification of the for-
age ration and Ivlev’s electivity index. Oecologia, 14, 413–17.
Johnson, B. K. and Christiansen, J. L. (1976). The food and food habits of Blanchard’s
cricket frog, Acris crepitans blanchardi (Amphibia, Anura, Hylidae), in Iowa. Journal of
Joly, P. (1987). Le regime alimentaire des amphibiens: methods d’etude. Alytes, 6, 11–17.
Juncá, F. A. and Eterovick, P. C. (2007). Feeding ecology of two sympatric species of
Aromobatidae, Allobates marchesianus and Anomaloglossus stepheni, in Central Amazon.
Kupfer, A., Nabhitabhata, J., and Himstedt, W. (2005). From water into soil: trophic
ecology of a caecilian amphibian (Genus Ichthyophis). Acta Oecologica, 28, 95–105.
Legler, J. M. (1977). Stomach flushing: a technique for chelonian dietary studies.
Legler, J. M. and Sullivan, L. J. (1979). The application of stomach-flushing to lizards and
anurans. Herpetologica, 35, 107–10.
Lima, A. P. (1998). The effects of size on the diets of six sympatric species of postmeta-
morphic litter anurans in central Amazonia. Journal of Herpetology, 32, 392–9.
Lima, A. P. and Magnusson, W. E. (2000). Does foraging activity change with ontogeny?
An assessment for six sympatric species of postmetamorphic litter anurans in Central
Amazonia. Journal of Herpetology, 34, 192–200.
MacNally, R. (1983). Trophic relationships of two sympatric species of Ranidella (Anura).
Magnusson, W. E., Lima, A. P., Silva, W. A., and Araújo, M. C. (2003). Use of geomet-
ric forms to estimate volume of invertebrates in ecological studies of dietary overlap.
Copeia, 2003, 13–19.
Mahan, R. D. and Johnson, J. R. (2007). Diet of the gray treefrog (Hyla versicolor) in rela-
tion to foraging site location. Journal of Herpetology, 41, 16–23.
Maneyro, R. and da Rosa, I. (2004). Temporal and spatial changes in the diet of Hyla
pulchella (Anura, Hylidae) in southern Uruguay. Phyllomedusa, 3, 101–13.
Matori, R. and Aun, L. (1994). Aspects of the ecology of a population of Tropidurus spinu-
losus. Amphibia-Reptilia, 15, 317–26.
May, R. M. and MacArthur, R. H. (1972). Niche overlap as a function of environmental
variability. Proceedings of the National Academy of Sciences USA, 69, 1109–13.
Miranda, T., Ebner, M., Solé, M., and Kwet, A. (2006). Spatial, seasonal and intrapopu-
lational variation in the diet of Pseudis cardosoi (Anura: Hylidae) from the Araucária
Plateau of Rio Grande do Sul, Brazil. South American Journal of Herpetology, 1, 121–30.
O’Reilly, J. C. (2000). Feeding in caecilians. In K. Schwenk (ed.), Feeding, Form, Function
and Evolution in Tetrapod Vertebrates. Academic Press, New York.
Pianka, E. R. (1973). The structure of lizard communities. Annual Review of Ecology and
Systematics, 4, 53–74.
Pianka, E. R. (1974). Niche overlap and diffuse competition. Proceedings of the National
Pianka, E. R., Oliphant, M. S., and Iverson, Z. L. (1971). Food habits of albacore bluefin,
tuna and bonito in Califorina waters. California Department of Fish and Game Bulletin,
152, 1–350.
Rocha, C. F. D., Van Sluys, M., Alves, M. A. S., Bergallo, H. G., and Vrcibradic, D. (2000).
Activity of leaf-litter frogs: when should frogs be sampled? Journal of Herpetology, 34,
285–7.
Schoener, T. W. (1967). The ecological significance of sexual dimorphism in size in the
lizard Anolis conspersus. Science, 155, 474–7.
Schoener, T. W. (1974). The compression hypothesis and temporal resource partitioning.
Proceedings of the National Academy of Sciences USA, 71, 4169–72.
Silva, H. R. and Britto-Pereira, M. C. (2006). How much fruit do fruit-eating frogs
eat? An investigation on the diet of Xenohyla truncata (Lissamphibia: Anura: Hylidae).
Journal of Zoology, 270, 692–8.
Simon, M. P. and Toft, C. A. (1991). Diet specialization in small vertebrates: mite-eating
in frogs. Oikos, 61, 263–78.
Simpson, E. H. (1949). Measurement of diversity. Nature, 163, 688.
Sokol, O. M. (1969). Feeding in the pipid frog Hymenochirus boettergi (Tornier).
Solé, M. and Pelz, B. (2007). Do male tree frogs feed during the breeding season? Stomach
flushing of five syntopic hylid species in Rio Grande do Sul, Brazil. Journal of Natural
History, 41, 2757–63.
Solé, M., Beckmann, O., Pelz, B., Kwet, A., and Engels, W. (2005). Stomach-flushing for
diet analysis in anurans: an improved protocol evaluated in a case study in Araucaria
forests, southern Brazil. Studies on Neotropical Fauna and Environment, 40, 23–8.
Toft, C. A. (1980a). Feeding ecology of thirteen syntopic species of anurans in a seasonal
tropical environment. Oecologia, 45, 131–41.
Toft, C. A. (1980b). Seasonal variation in populations of Panamanian litter frogs and
their prey: a comparison of wetter and drier sites. Oecologia, 47, 34–8.
Toft, C. A. (1981). Feeding ecology of Panamanian litter anurans: patterns in diet and
foraging mode. Journal of Herpetology, 15, 139–44.
Toft, C. A. (1985). Resource partitioning in amphibians and reptiles. Copeia, 1985,

1–21.
Triplehorn, C. A. and Johnson, N. F. (2005). Borror and Delong’s Introduction to the Study
of Insects, 7th edn. Thomson Brooks/Cole, Belmont, CA.
Vitt, L. J., Zani, P. A., and Espósito, M. C. (1999). Ecological differences in tropical
sympatric skinks (Mabouya macrorhyncha and Mabouya agilis) in southeastern Brazil.
Wang, Y. P., Zhang, P., and Li, Y. (2008). Diet composition of post-metamorphic bull-
frogs (Rana catesbeiana) in the Zhoushan archipelago, Zhejiang Province, China.
Frontiers of Biology in China, 3, 219–26.
Weaver, W. and Shannon, C. E. (1949). The Mathematical Theory of Communication.
University of Illinois, Urbana, IL.
Werner, E. E., Wellborn, G. A., and McPeek, M. A. (1995). Diet composition in post-
metamorphic bullfrogs and green frogs: implications for interspecific predation and
competition. Journal of Herpetology, 29, 600–7.
Wu, Z. J., Li, Y. M., and Wang, Y. P. (2007). A comparison of stomach flush and stomach
dissection in diet analysis of four frog species. Acta Zoologica Sinica, 53, 364–72.
Yilmaz, Z. C. and Kutrup, B. (2006). Seasonal changes in the diet of Rana ridibunda
Pallas, 1771 (Anura: Ranidae) from the Gorele River, Giresun, Turkey. In M. Vences,
J. Köhler, T. Ziegler, and W. Böhme (eds), Herpetologia Bonnensis II. Proceedings of the
13th Congress of the Societas Europaea Herpetologica, pp. 201–4. Bonn, Germany.
Zar, J. H. (1999). Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, NJ.
11
Movement patterns and radiotelemetry
Dale M. Madison, Valorie R. Titus, and Victor S. Lamoureux
11.1 Introduction
There is no greater need for details on the year-round movements and refuges of
amphibians in their natural habitats than today, mainly due to the pressure to
develop lands for human habitation, agriculture, and commerce, and to define
better wetland/buffer zones to protect amphibian communities. Unfortunately,
there is a dearth of information on habitat use for most amphibians due to their
relatively small size and secretive nature. Efforts to track individuals using radio-
isotopes began decades ago (Madison and Shoop 1970), and other techniques
have followed, including fluorescent powder, spool tracking, radiotelemetry,
and harmonic radar, but for most amphibians details on habitat use through
the annual cycle are still lacking. The recent surge in radiotelemetry use is pro-
viding new insights into habitat use and movements, but only a few studies have
addressed whether the data might reflect transmitter transport, attachment,
implantation, or the periodic disturbance of the investigator.
Favorable outcomes in the use of radiotelemetry require having a clear pur-
pose for the study, obtaining reliable equipment and knowing how to use it,
choosing the transmitter and external/internal tagging technique that best
fit the need and the animal chosen for study, having the knowledge and skills
necessary to use the attachment or implant techniques, anticipating the poten-
tial indirect and direct impacts of the radiotransmitter on the individual during
and after tracking, and planning in advance how to analyze the anticipated data
set. In this chapter, we discuss our experiences with amphibians to reduce dupli-
cation of effort and trauma to the animals studied, and to accelerate technical
advances. We do not report all radiotelemetry studies, nor do we repeat the con-
tent of earlier reviews (see Heyer et al. 1994), but focus instead on more recent
studies and models for future application of radiotelemetry procedure.
11.2 Equipment
The basic radiotelemetry equipment for hand-tracking amphibians includes
radiotransmitters, a portable radio receiver, an antenna, and a cable. There are
many international vendors of telemetry equipment but no single authority that
has direct experience with all products and applications. Our advice is to rely
on the recommendation of experienced users having the broadest biological and
technical experience, and then choose the most reliable equipment available.
11.2.1 Receivers and antennas

The complexity and price of a receiver will vary with the scale of the study and
the types of information being electronically recorded. There are basic receivers,
such as the R-1000 (Communications Specialists), which can be programmed to
999 different animal frequencies between 148 and 174 MHz. Wider frequency
ranges occur in more complex receivers, such as the ATS R5400S (Advanced
Telemetry Systems), which can also upload information from data loggers in
the transmitters. The cost of good-quality receivers is between US$700 and
$2500. The small size of most amphibians usually restricts the transmitter size
and application to the use of less expensive receivers like the R-1000.
The most commonly used antenna is a three- or four-element, hand-held Yagi
antenna. Larger, five- (or more) element Yagi antennas are typically mounted
on a vehicle for tracking wide-ranging animals, and give slightly better direc-
tion and distance information (Heyer et al. 1994). The main disadvantage of
the five-element antenna in hand-tracking amphibians is that its larger size,
necessitated by the fixed spacing between the elements for maximum frequency
sensitivity, makes it more difficult to use in thick undergrowth. There are also H
and loop antennas, but these somewhat reduce signal range/resolution. When
ordering an antenna, tell the manufacturer what frequency range will be used
so the antenna can be set up for optimal reception. Antennas typically range
from US$100 to $300. We recommend a fold-up, three-element antenna for
hand-held applications, and always carrying an extra antenna cable in case one
fails in the field.
11.2.2 Transmitters
The types and sizes of transmitters, transmitter antennas and batteries vary,
but the options grow fewer with smaller animals. In general, the weight of the
transmitter/battery/attachment device should not exceed 10% of body weight
on land (see Heyer et al. 1994), it should be neutrally buoyant in water, and
the detection range should be at least 30–50 m line-of-sight. The smaller the
11 Movement patterns and radiotelemetry | 187
transmitter/battery unit required to remain under the 10% weight rule, the
more one has to compromise signal distance, lifetime, or pulse rate. There
are three trade-offs in selecting a transmitter/battery combination at a given
weight: (1) the greater the pulse rate (from about 50 beats/min to continuous),
the more quickly and reliably you can determine signal direction, but the
shorter the life of the transmitter/battery unit, (2) the larger (more powerful)
the transmitter, the greater the detection distance, but the shorter the life of
the transmitter/battery unit, and (3) the longer the transmitter antenna, the
greater the detection distance, but the greater the impact on the animal. One
should begin by determining the weight of the completed unit (including the
waistband if used), and then choose the transmitter type/battery/attachment
option that best serves the purpose of the study. In addition, in choosing trans-
mitter frequencies one should avoid CB frequencies to ensure that there is no
interference at the field site, especially near major highways or commercial
properties. Finally, one may want to choose a vendor (e.g. Holohil Systems,
Carp, Ontario, Canada) that will refurbish transmitters with spent batteries
for a significant cost reduction.
11.2.2.1 External transmitters

In anurans, the transmitter is attached to a waist belt made out of several mater-
ials, few of which are satisfactory (Goldberg et al. 2002; Weick et al. 2005).
Waist-belt trauma is frequent when the belts are attached snug enough so as
not to slip off (Weick et al. 2005). External attachment has been achieved in
cryptobranchid salamanders by sutures to the skin (Okada et al. 2006), but we
agree with a similar study in hellbenders that doesn’t recommend this procedure
(Blais 1996).
Besides causing lesions and open wounds, especially in frogs with more deli-
cate skin than toads, waist belts may cause trauma and death when the attach-
ment bands or antennas become snagged in vegetation (e.g. McAllister et al.
2004). The whip antenna alone would likely affect movement in the thick grass
and grass runways used by green frogs during summer foraging (Lamoureux
et al. 2002). Slippage off the animal is also common, either following seasonal
weight loss or being pulled off after snagging on vegetation (Rathbun and
Murphey 1996; McAllister et al. 2004; Baldwin et al. 2006; Rittenhouse and
Semlitsch 2007). Although some of these problems may be minimal over short
time periods, many animals whose signals have failed or whose movements carry
them outside the range of the tracking equipment will end up wearing these
belts for the rest of their lives with increasing fitness reduction (e.g. weight loss,
lower growth) and risk of injury/death. There are no known records to our
knowledge of a transmitter-belt complex being worn successfully by an anuran

spanning two breeding seasons. Even in one comprehensive and exceptionally
well-documented study, including up to four re-beltings per frog, no mention of
skin trauma or potential negative effect of transmitter belts was reported, and yet
only five of 43 frogs had been tracked for more than 8 weeks (mean 25.6 days/
frog) between 16 April and 12 November (Baldwin et al. 2006). Just one frog had
its transmitter removed at the end of tracking, 18 had lost signals, 12 had been
depredated, nine had “slipped out” of their waist bands, and three were found
dead with no sign of predation. There is room here, and in essentially all other
transmitter-belt studies lasting for more than a few days, for extensive direct and
indirect negative effects of external belt attachment. It is prudent to consider any
amphibian telemetry study relying on waist belts to be “long-term” unless the
belts are removed after a week or so, or a release mechanism is built into the units.
Benefits of waist bands include freedom from surgery, and being able to release an
animal within moments of capture and transmitter attachment.
11.2.2.2 Internal transmitters

Whereas transmitter belts likely have value in certain circumstances (e.g.
Fukuyama et al. 1988), preferably following laboratory-validated safety tests, we
recommend implanted units for most amphibians, as have others (Madison 1997;
Spieler and Linsenmair 1998; Lamoureux and Madison 1999; Lamoureux et al.
2002; Lemckert and Brassil 2003; Johnson 2006; Rittenhouse and Semlitsch
2006; McDonough and Paton 2007; Peterman et al. 2008). The implanted
transmitter/battery unit has various antenna strands or coils embedded in the
potting material to make a smooth-surfaced implant package. Alternatively,
Jehle and Arntzen (2000) guided the 3–5 cm antenna out through the body
wall (to increase signal distance reception) and sutured it to the body surface,
an application we do not recommend. The potting material is an impervious
coating of epoxy, dental acrylic, Elvax paraffin, beeswax, heat-shrink plas-
tic, or non-reactive butyl rubber (Heyer et al. 1994). Care should be taken to
use potting substances that do not elicit an immune response or release toxins
upon implantation, which in the past were thought to stimulate cyst formation
around the transmitter. Elongated implants should not have a length more than
the diameter of the animal’s greatest body or head width, because such units
have caused an animal to get permanently stuck turning around in a narrow
tunnel or passage (Madison 1997).
One variation of the “internal placement” of a radiotransmitter is down the
esophagus as a gastric pill (Oldham and Swan 1992; Blais 1996; Schabetsberger
et al. 2004), which in hellbenders were passed naturally after between 16 and
25 days, allowed short-term movement records, were recovered without the

necessity of salamander recapture, and provided the guarantee of no long-term
burden (Blais 1996). With this application, sizes just large enough to force
regurgitation or small enough to be passed completely through the digestive
system need to be considered carefully (Schabetsberger et al. 2004).
11.3 Surgical techniques

Surgery on amphibians requires knowing how to enter the coelomic cavity and
properly close the wound. Individuals performing surgery should have training
in the company of someone with previous surgical experience on amphibians.
Amphibian skin, especially for most salamanders and smaller frogs, is much
more delicate than mammalian skin, and has much slower healing rates in cool/
cold environments. Transmitter collar and implant techniques have been devel-
oped for field studies of small mammals (see references in McShea and Madison
1992), but these techniques are inappropriate for, and even detrimental to,
amphibians. The latest surgical advances for amphibians are incorporated in
our recommendations below.
11.3.1 Surgery
Transmitters are typically implanted into the coelomic cavity. Some studies have
implanted transmitters under the skin (Blais 1996; Lemckert and Brassil 2003),
but sloughing of the transmitter through the skin (without edema or infection)
can occur when using this technique (Blais 1996). Lumps on the body also pro-
mote abrasion/injury and suture failure as the animal moves through restrictive
environments (Weick et al. 2005). We do not recommend subcutaneous trans-
mitter implants for amphibians.
The following surgical procedure with minor variations has been used suc-
cessfully (Colberg et al. 1997; Madison and Farrand 1998; Faccio 2003; Crook
and Whiteman 2006; Rittenhouse and Semlitsch 2006; Cecala et al. 2007). For
coelomic implants, all instruments and transmitters should be sterilized, such as
in 95% ethanol followed by distilled water rinse and drying. For anesthetiza-
tion, we recommend near total submersion in MS-222 (3-aminobenzoic acid
ester methanosulfate salt) diluted with distilled water to a 0.25% solution and
buffered to pH 7.0 with sodium bicarbonate (this concentration is specifically
for adult Ambystoma tigrinum). The concentration should be adjusted to animal
size, species, and habitat condition (see Lowe 2004; Crook and Whiteman 2006;
Peterman and Semlitsch 2006), so that it takes about 10 min or more for the loss of
the “righting” response or reaction to the poke of a probe. More rapid anesthesia,
because either of higher temperatures or concentrations, increases the risk of death

(Weick et al. 2005). Animals anesthetized too rapidly, or those over-anesthetized,
should be rinsed continuously under cold tap water for at least 60 s to promote
recovery. Animals should also be rinsed briefly under cool tap water immediately
after anesthesia but before surgery to minimize the mild anticoagulant effect of
MS-222 in the surgical field (Blais 1996). Benzocaine anesthetic has also been
used (Crook and Whiteman 2006; Cecala et al. 2007), but MS-222 is preferred
due to its wider margin of safety. For surgery, damp toweling placed under and
over the animal, except for the surgical field, will reduce drying. A longitudinal
incision should be made in the ventral posteriolateral abdominal wall with cutting
iris scissors just long enough to insert the transmitter without tearing the tissue
(Figure 11.1). Non-absorbable, braided, polyviolene sutures (5–0), applied by a
3/8 circle, reverse-cutting needle and secured by a standard surgeon’s knot are
recommended for wound closure. Other non-absorbable sutures are not recom-
mended due to the lack of flexibility in the nylon. Just one layer of sutures is needed
for most salamanders because the muscles are bonded to the skin. Otherwise, and
for anurans in general, two layers of sutures are required, one for the inner mus-
culature layer and the second for the skin (see Goldberg et al. 2002). A maximum
of 1–1.5 mm spacing between all sutures is recommended to completely close the
inner/outer incisions for prolonged healing and to put less stress on any single
suture. Investigators routinely use too few sutures and consequently observe more
dehiscence. Sutures drawn too tightly can also lead to tissue necrosis, severance,
and dehiscence. Finally, some studies have used dissolvable sutures (Peterman
et al. 2008), which we discourage, especially when healing is slowed, even for up
to 6 months, during cooler times of the year (Colberg et al. 1997).
Fig. 11.1 Implanting of an internal transmitter into an eastern tiger salamander

(Ambystoma tigrinum tigrinum). Photographs by Valorie R. Titus.
A final precaution regarding surgery involves gravid females. Amphibian

eggs are hydrophilic, and exposing coelomic eggs to any water during implant
surgery can cause coelomic egg-mass expansion and difficulty in suturing the
wound closed. Too few sutures, or a permanent opening into the body cavity to
accommodate an antenna, can also permit some water entry and possible injury
or death. We also do not recommend implant surgery on females that will soon
undergo amplexus. The physical stress on the sutures, even in anurans with gen-
erally tougher skins, may cause dehiscence.
11.3.2 Recovery and healing

The surgical field should not be treated before or after surgery with ethanol
or other strong antiseptics, which will kill the animal’s natural biota and pro-
mote bacterial/fungal infection soon after field release in a soiled environment.
Probiotics have not been developed for amphibian skin. In addition, since the
skin of amphibians is thin and permeable, ethanol or other antiseptics may
be absorbed directly into the bloodstream and cause physiological stress, espe-
cially in aquatic organisms (see Crook and Whiteman 2006). Upon completion
of surgery and rinsing in clean tap water, the animals should be placed under
damp toweling or gauze for recovery. Animals should revive within 1 h and
show full recovery in 3 h. Although complete healing may take several weeks
when temperatures fall below 15°C, release soon after recovery from the anes-
thetic does not appear to have negative effects on survival or movements and
is most likely to return the animal to its pre-capture activity (Madison 1997;
Spieler and Linsenmair 1998; Johnson 2006; Rittenhouse and Semlitsch 2006;
McDonough and Paton 2007; Peterman et al. 2008).
11.4 Tracking procedures

Radiotracking technique takes practice to glean the most information from
a signal while minimizing any negative observer effect on the animal and its
environment. Simply stated, a novice quickly becomes mesmerized by the signal
and unaware of possible habitat disturbance and the risk of trampling the ani-
mal in trying to get an exact position, whether tracking in the water or on land.
In this section, we discuss strategies and precautions relating to animal release,
search, and data collection.
11.4.1 Animal release

Animals should normally be released under cover at the capture site at the begin-
ning of the next active period, usually just after dusk the following evening. Just
prior to release, check for transmitter functioning one last time, and tune the
receiver to the strongest signal output of the transmitter. Tuning on an additional
channel is also suggested (Baldwin et al. 2006). Re- or multi-programming the
receiver to an exact frequency setting improves the chance of re-locating a lost
signal. We also suggest that the investigator stay nearby after release and obtain
positions every 30–60 min for several hours to determine the initial movement
headings, which are generally straight and can often be used to find a lost sig-
nal the next day. Searching for a lost signal without any idea which direction to
search can be time-consuming and frustrating.
11.4.2 Locating signal source

In order to obtain the best, most rapid directional information when searching
for, or approaching, a distant signal, hold the antenna parallel to the ground
and continually sweep it to the signal nulls right and left of the loudest signal
intensity, rather than trying to aim the shaft in the direction of the loudest
signal toward the animal. Signal information can be used to estimate the dis-
tance between the observer and the animal. When the initial signal is detected,
adjust the volume on the receiver to be barely audible and continually re-adjust
to this low volume as you approach the animal. At greater distances from the
animal, white noise will dominate the signal background. The closer you are
to the animal, the purer the signal will be. Once a relatively noise-free signal is
obtained, we recommend a local triangulation procedure (about 3 m from the
animal) to finally fix the animal’s position (Mineau and Madison 1977). This
procedure is also used to locate amphibians in a pond, stream, or up a tree (e.g.
Blais 1996; Madison 1998); it involves the observer moving on a tangent to the
signal and estimating two independent headings that define approximately a
90° angle of convergence to the signal source. Local triangulation minimizes
the chance of disturbing the animal and/or its microenvironment, which might
induce subsequent movement, and it reduces the chance of trampling the ani-
mal. To get an exact location, one should approach the last 3 m by holding the
antenna shaft perpendicular to the ground (and the elements perpendicular to
the signal source) and continually move the antenna tip away from your feet
and up through an arc of about 135°, keeping the loudest signal in front of you.
This procedure can alternate with a side-to-side scan to refine the directional
heading. Soon the animal will become visible, or a location on the ground/leaf
litter above the animal can be identified to within 10 cm (Madison and Farrand
1998; Faccio 2003). Avoid disturbing the animal unless it needs to be retrieved
for weighing, replacing a transmitter, or checking on its condition. We generally
do not search for an unseen animal until it occurs in exactly the same location
for multiple position checks, especially following rain when animals are most
likely to move. Removing an animal from a subterranean retreat should follow
a circular excavation procedure to avoid injuring the animal and to enhance the
description of the refuge (Madison and Farrand 1998).
Several factors can influence signal direction and detectability. Animals near
or under metal objects, boulders, bluffs, small ridges and depressions, utility
poles, or fences, as well as underground or underwater, will be more difficult to
locate because of signal deflection and dampening. Madison and Farrand (1998)
experienced about a 1 m distance reduction for each centimeter of transmitter
depth, but this dampening can vary with different equipment. The reduction
for water depth is much less than for soil. If an arboreal animal is being tracked,
or if the observer and/or animal is on higher ground, the normal signal range
will be increased and the animal will sound closer than it actually is. An obser-
ver looking for a lost signal should raise the antenna high overhead and, where
possible, move to higher ground to improve detection distance.
11.4.3 Data collection

Most investigators use hand-held equipment to obtain positions, which minim-
ally requires a GPS unit, flagging, a topographical map, and a waterproof field
data book. Typically, 10–20 animals can be tracked concurrently, but this varies
with topographic features, understory thickness, how dispersed the animals are,
and how much data are taken during a position check. GPS data are less accurate
on steep slopes and in ravines. Flagging should be used to mark each fix or key
locations with the animal ID, date, and time of observation written in ballpoint
pen on the flagging. A wide variety of environmental and refuge data can also be
recorded for terrestrial or aquatic locations using data loggers, weather stations,
and local field and laboratory techniques (e.g. Heyer et al. 1994; Madison 1997;
Madison and Farrand 1998; Faccio 2003; Watson et al. 2003; Baldwin et al.
2006; Rittenhouse and Semlitsch 2007).
11.5 Analysis of movement data

Telemetry generates extensive good or poor data, depending on whether the study
objectives, sampling methods and statistical procedures were planned in advance.
There are several resources on analyzing directionality, space use and movement
data, depending on the study objectives (cited in White and Garrott 1990; Kie
et al. 1996; Madison 1997; Millspaugh and Marzluff 2001; Blomquist and Hunter
2007), including applications of advanced Geographic Information System (GIS)
software, such as ArcGIS (Environmental Systems Research Institute, Redlands,
Wetland
margin
Meadow
#10 #3
= Summer areas
Shrubs
= Over-wintering
sites
Pond
Woods = Foraging
forays
#9
= Migratory
movements
Swamp
80 m
(a)
(c) (b)
Fig. 11.2 Tracks for three radio-implanted northern green frogs (Rana clamitans),
Broome County, NY, USA, showing vegetation types, habitat photographs, and
major movement/refuge details. Wetland margins were structurally complex,
consisting of Leersia, Polygonum, Sparganium, Scirpus, and Carex. The swamp had an
CA, USA). The latter permits detailed maps of space use to be overlaid with vege-
tation and soil maps for comprehensive studies of habitat use.
A common mistake in quantifying space use is rushing to label and quan-
tify the “home range” for an amphibian, which should include all behaviors
relating to reproduction and survival (Burt 1943), or calculating the average or
maximum distance moved from a breeding pond without qualifying that the
estimates are based on a small subset of individuals over a few weeks or months.
Lacking for the most part is information on year-round movements in struc-
turally different habitats. Results are seldom scaled or weighted according to
the duration of a telemetry record, for example averaging a one-week distance
record with a 4-month distance record can underestimate the potential for a
species to move long distances (Madison and Farrand 1998). The potential for
individuals to occupy widely spaced and diverse habitats for different purposes
through the annual cycle must be kept in mind, and a good but preliminary
descriptive model for these movements is that for green frogs (Figure 11.2), as
described in Lamoureux and Madison (1999) and Lamoureux et al. (2002).
More quantitative indices of multi-seasonal space use also occur (e.g. Spieler and
Linsenmair 1998; Watson et al. 2003; Baldwin et al. 2006).
The most common home-range representation is the minimum convex
polygon, and while this method may give fairly accurate short-term “daily”
Fig. 11.2 Continued

understory of ferns, various thick grasses, Scirpus, Juncus, and Carex, and a canopy
consisting largely of Acer rubrum. The deciduous forest consisted of Fagus, Betula,
Acer, Tsuga, and Quercus. Subject numbers at track origins are consistent with
Lamoureux and Madison (1999) and Lamoureux et al. (2002). Frog #3 (a 31
g female released 12 September 1995) occupied a wetland margin, made four
back-and-forth foraging forays to self-made forms or rodent runways in dense
grasses during October (photograph a), migrated 240 m on 3 November to an
over-wintering site in a large beaver dam, and returned to its release (breeding)
location in April 2006. Frog #9 (a 45 g male released 10 August 1996) made
one out-and-back foray into the swamp, migrated across the swamp and 240 m
up a forested hillside on 8 September 1996, made additional movements with a
succession of leaf litter refuges (e.g. photograph b) 200 m further up the hillside
to its 19 October 1996 over-wintering location consisting of a small spring with
approximately 10 cm-diameter tunnels under a large root (photograph c), and
finally was recaptured there May 1997 prior to the spring migration. Frog #10 (a
63 g male released on 26 August 1996) moved 250 m on 30 September from its
wetland/woodland-edge habitat to the swamp, moved 40 m further into the swamp
on 3 October 1996, and then was lost until relocated on 11 October 1996 360 m
into forested habitat buried in a small, seasonal stream under flat stones and small
pebbles, where it over-wintered. Photographs by Victor S. Lamoureux.
or “activity” range estimates for a mobile individual (Madison 1985; McShea

and Madison 1992), many migratory amphibians make only a few large move-
ments with most activity being in small subterranean food or shelter patches
strung out somewhat longitudinally through upland habitat, such as in
A. tigrinum (Figure 11.3). Similar observations have been recorded by others
VEGETATION W E
Pinus rigida-Quercus (alba, Velutina)-Vaccinium palladium
Quercus coccinea-Vaccinium palladium/Gaylussacia baccata S
Pinus rigida /Quercus (coccinea, alba)-Vaccinium palladium
Grass 6
Road 5
0 20 40 80 m
Water
(a)
1 (c)
(b)
3
2
Fig. 11.3 Year-long telemetry track of an implanted 33 g male Ambystoma tigrinum,

Suffolk County, NY, USA, showing vegetation types, habitat photographs, and major
movement/refuge details as follows: Star (30 March 2005): initial position following
breeding in the vegetation at the edge of the pond. Point 1 (4 April 2005): initial
movement 137 m from pond edge into small self-made borrow 8 cm from surface.
Point 2 (21 June 2005): movement 21 m further from pond into a small self-made
borrow 10 cm from surface. Point 3 (2 July 2005): movement 23 m back toward
pond into a small mammal borrow 8 cm from surface. Point 4 (1 October 2005):
movement 78 m toward pond into a small mammal borrow 12 cm from surface.
Point 5 (4 December 2005): movement 68 m back into the breeding pond the
following season. Point 6 (15 March 2006): movement 33 m out of pond; later only
the transmitter was found on the surface (possible predation). All movements were
associated with rain events. Photographs include (a) grassland and road traversed at
least twice during migrations, (b) salamander near burrow at point 3, and (c) typical
woodland/understory at points 1, 2, 3, and 6. Photographs by Valorie R. Titus.
(Fukuyama et al.1988; Madison 1997; Madison and Farrand 1998; Lemckert

and Brassil 2000; Faccio 2003; Lemckert and Brassil 2003; Rittenhouse and
Semlitsch 2006, 2007). If an estimate of the home range or the buffer zone
for a species is required, we recommend caution in whatever model is used,
and a clear statement concerning the limitations of such values so they are not
misinterpreted by others to represent the year-round resource needs or habitat
buffer for a sustainable amphibian population (see McDonough and Paton
2007).
11.6 Validation of telemetry procedures

Radiotelemetry and drift fences with pitfall traps, although standard tech-
niques for monitoring amphibian movements, have not been fully examined
for shortcomings (see Heyer et al. 1994; Madison and Farrand 1998). The
following data suggest that telemetry can be used to monitor amphibians with
reduced impact on survival and reproduction, at least in short-term studies.
11.6.1 Internal condition and mass

In external attachment studies, no effect on body mass was shown for two
free-ranging toad species, since carrying transmitters of different weight or
over different time periods had no differential effect on body mass (Indermaur
et al. 2008). In another study, 16 frogs (Rana aurora) gained an average of
5.7 g after 2 months of carrying an external transmitter on a beaded chain
(Rathbun and Murphey 1996), but how much the frogs might have weighed
without carrying transmitters wasn’t known. In implant studies during the
post-breeding season, frogs (Rana clamitans) with transmitter implants
gained the same weight as frogs without implants (Lamoureux et al. 2002),
and weight and growth gains also occurred before and during the breeding
season in Hoplobatrachus occipitalis (Spieler and Linsenmair 1998). However,
Hyla versicolor showed a 9% average weight loss (1.1 g) following 3 weeks of
carrying an implant, and other individuals with long-term implants coated
with dental acrylic showed fibrous connective tissue cysts or strands around
the implants, but neither of these effects was considered detrimental (Johnson
2006). Similar connective tissue was associated with toad implants and was
thought to anchor the implants away from body organs (Gray et al. 2005).
Failure to form such cysts may result in cloacal expulsion, as observed in
leopard frogs (Weick et al. 2005). In a laboratory study of male spotted sala-
manders (Ambystoma maculatum), implanted animals showed significantly
depressed feeding compared to non-implanted salamanders for 2 weeks after
surgery, but no significant decrease in feeding behavior or body weight rela-

tive to the controls for the remaining 8 months of the study (Madison 1997).
Collectively, these data show no major effects on body weight or internal
condition.
11.6.2 Movements
Little is known about the full extent of habitat use in free-ranging amphib-
ians other than that obtained using radiotelemetry, and so it is difficult to test
whether transmitters might be affecting movements and habitat use. In field
enclosures, Rana sylvatica and Rana pipiens with and without belted transmit-
ters with 15 cm antennas showed some species-specific differences in behavior
in response to simulated predation threat, prompting some caution in the use of
the attached units studied (Blomquist and Hunter 2007). Studies of a rainforest
frog (Litoria lesueuri) in laboratory enclosures showed no difference in the dis-
tance or frequency of movement before and after transmitter attachment, except
for a slight reduction the first day after attachment (Rowley and Alford 2007).
In nature, casual observations on non-implanted A. maculatum salamanders
during tracking of implanted individuals revealed comparable movements,
occurrence, and mortality (Madison 1997). Cross-study comparisons show
that wood frogs monitored by drift fences (Vasconcelos and Calhoun 2004)
emigrated comparable distances to implanted individuals (Baldwin et al. 2006;
Rittenhouse and Semlitsch 2007). In addition, implanted tiger salamanders
showed courtship movements in a breeding pond similar to those observed for
non-implanted salamanders in captivity (Madison 1998). The above studies,
although quite incomplete, suggest no major effects of implanted or attached
transmitters on short-term movements.
11.6.3 Reproduction
Reproduction is seldom verified for amphibians carrying transmitters, although
there are exceptions. Six gravid female frogs (Buergeria buergeri) were given
waist-band transmitters (with a 3 cm whip antenna) and monitored along a
stream for up to 8 days during successful mating and oviposition (Fukuyama
et al. 1988). In implant studies, Johnson (2006) recorded a female Hyla versicolor
ovipositing a year after carrying an implant and undergoing two implant surger-
ies. Tiger salamanders showed normal breeding activity soon after implantation
(Madison 1998), and an implanted male spotted salamander returned 190 m to
the same location at a breeding pond after carrying an implant in upland habitat
for a year (Madison 1997). Further studies on possible long-term reproductive
effects are obviously needed.
11.6.4 Injury and survivorship

For external attachment, the recurrent problem has been skin injury or getting
entangled in vegetation or caught in tight spaces, perhaps more so for frogs than
toads (e.g. Rathbun and Murphey 1996; Weick et al. 2005; Indermaur et al.
2008), which increases with track duration (Rittenhouse and Semlitsch 2007).
For implant procedures, the main concern has been the somewhat high rate of
recovery of bare transmitters, suggesting a high population turnover rate due to
predators (herons, shrews, snakes) and/or the cloacal expulsion of the implants.
Predation is strongly supported (Madison 1997; Madison and Farrand 1998;
McDonough and Paton 2007), but recent evidence also supports expulsion
(Weick et al. 2005).
Little is known about the effects of transmitter implants or belts on the long-
term survival of amphibians. And even during short-term tracking where preda-
tion is the cause of death, the attached/implanted transmitters may predispose
animals to predation, as cautioned by Blomquist and Hunter (2007). Clearly,
more data are needed to evaluate effects on long-term survivorship.
11.7 Conclusions
In order to fully understand amphibian movement patterns and habitat use,
radiotelemetry is an indispensable tool, despite some shortcomings in meth-
odology. Waist bands for external transmitters with short whip antennas (less
than ≈3 cm), or antennas built into the waist band rather than trailing behind
the animal, are tentatively recommended for short-term tracking of anurans
under conditions where the animals can be inspected periodically and the waist
bands removed at the end of study (e.g. Fukuyama et al. 1988). However, long-
term tracking using waist-band transmitters is not recommended because of the
increased likelihood of skin trauma and getting snagged or trapped by objects in
post-breeding and over-wintering refuges. External transmitters and antennas
sutured to amphibians are not recommended.
Transmitters with antennas coiled in the potting material are recommended
for short-term studies as gastric pills (e.g. Schabetsberger et al. 2004), or for
extended studies as coelomic implants in both anurans (e.g. Lamoureux and
Madison 1999) and salamanders (Madison 1997; Madison and Farrand 1998;
Faccio 2003; Rittenhouse and Semlitsch 2006; McDonough and Paton 2007).
Especially needed, however, is information on potential weight effects on fitness
from the long-term transport of transmitter implants.
The future challenge in the use of radiotelemetry in amphibians is threefold:
to minimize duplication in method development and unnecessary trauma
to amphibians by fostering apprenticeship prior to use by an investigator, to

educate journal editors and academic units of the appropriate procedures,
and to encourage publication of data on injury, mortality, or other problems
resulting from responsible radiotelemetry procedure, a notable example being
Weick et al. (2005).
11.8 References
Baldwin, R. F., Calhoun, A. J. K., and deMaynadier, P. G. (2006). Conservation planning
for amphibian species with complex habitat requirements: A case study using move-
ments and habitat selection of the Wood Frog Rana sylvatica. Journal of Herpetology,
40, 442–53.
Blais, D. P. (1996). Movement, Home Range, and Other Aspects of the Biology of the Eastern
Hellbender (Cryptobranchus alleganiensis alleganiensus): a Radiotelemetric Study.
Master’s thesis, State University of New York at Binghamton, Binghamton, NY.
Blomquist, S. M. and Hunter Jr, M. L. (2007). Externally attached radio-transmitters
have limited effects on the antipredator behavior and vagility of Rana pipiens and Rana
sylvatica. Journal of Herpetology, 41, 430–8.
Burt, W. H. (1943). Territoriality and home range concepts as applied to mammals.
Journal of Mammalogy, 24, 346–52.
Cecala, K. K., Price, S. J., and Dorcas, M. E. (2007). A comparison of the effectiveness
of recommended doses of MS-222 (tricaine methanesulfonate) and Orajel (R) (benzo-
caine) for amphibian anesthesia. Herpetological Review, 38, 63–6.
Colberg, M. E., Denardo, D. F., Rojek, N. A., and Miller, J. W. (1997). Surgical proced-
ure for radio transmitter implantation into aquatic, larval salamanders. Herpetological
Review, 28, 77–8.
Crook, A. C. and Whiteman, H. H. (2006). An evaluation of MS-222 and benzocaine as
anesthetics for metamorphic and paedomorphic tiger salamanders (Ambystoma tigri-
num nebulosum). American Midland Naturalist, 155, 417–21.
Faccio, S. D. (2003). Postbreeding emigration and habitat use by Jefferson and spotted
salamanders in Vermont. Journal of Herpetology, 37, 479–89.
Fukuyama, K., Kusano, T., and Nakane, M. (1988). A radio-tracking study of the behav-
iour of females of the frog Buergeria buergeri (Rhacophoridae, Amphibia) in a breeding
stream in Japan. Japanese Journal of Herpetology, 12, 102–7.
Goldberg, C. S., Goode, M. J., Schwalbe, C. R., and Jarchow, J. L. (2002). External
and implanted methods of radiotransmitter attachment to a terrestrial anuran
(Eleutherodactylus augusti). Herpetological Review, 33, 191–4.
Gray, M. J., Miller, D. L., and Smith, L. M. (2005). Coelomic response and signal range
of implant transmitters in Bufo cognatus. Herpetological Review, 36, 285–8.
(eds) (1994). Measuring and Monitoring Diversity: Standard Methods for Amphibians,
Smithsonian Institution Press, Washington DC.
Indermaur, L.B., Schmidt, B.R., and Tockner, K. (2008). Effect of transmitter mass and
tracking duration on body mass of two anuran species. Amphibia-Reptilia, 29, 263–9.
Jehle, R. and Arntzen, J. W. (2000). Post-breeding migrations of newts (Triturus crista-

tus and T. marmoratus) with contrasting ecological requirements. Journal of Zoology,
London, 251, 297–396.
Johnson, J. R. (2006). Success of intracoelomic radiotransmitter implantation in the tree-
frog (Hyla versicolor). Lab Animal, 35, 29–33.
Kie, J. G., Baldwin, J. A., and Evans, C. J. (1996). CALHOME: a program for estimating
animal home ranges. Wildlife Society Bulletin, 24, 342–4.
Lamoureux, V. S. and Madison, D. M. (1999). Overwintering habitats of radio-implanted
green frogs, Rana clamitans. Journal of Herpetology, 33, 430–5.
Lamoureux, V. S., Maerz, J. C., and Madison, D. M. (2002). Premigratory autumn for-
aging forays in the green frog, Rana clamitans. Journal of Herpetology, 36, 245–54.
Lemckert, F. and Brassil, T. (2000). Movements and habitat use of the endangered giant
barred river frog (Mixophyes iteratus) and the implications for its conservation in timber
production forests. Biological Conservation, 96, 177–84.
Lemckert, F. and Brassil, T. (2003). Movements and habitat use by the giant burrowing
frog, Heleioporus australiacus. Amphibia-Reptilia, 24, 207–11.
Lowe, J. (2004). Rates of tricaine methanesulfonate (MS-222) anesthetization in rela-
tion to pH and concentration in five terrestrial salamanders. Herpetological Review, 35,
352–4.
Madison, D. M. (1985). Activity Rhythms and Spacing. In R.H. Tamarin (ed.), Biology of
New World Microtus, pp. 373–419. American Society of Mammalogists, Lawrence, KA.
Madison, D. M. (1997). The emigration of radio-implanted spotted salamanders,
Ambystoma maculatum. Journal of Herpetology, 31, 542–51.
Madison, D. M. (1998). Habitat-contingent reproductive behaviour in radio-implanted
salamanders: a model and test. Animal Behaviour, 55, 1203–10.
Madison, D. M. and Shoop, C. R. (1970). Homing behavior orientation and home range
of salamanders tagged with tantalum-182. Science, 168, 1484–7.
Madison, D. M. and Farrand, L. (1998). Habitat use during breeding and emigration in
radio-implanted tiger salamanders, Ambystoma tigrinum. Copeia, 1998, 402–10.
McAllister, K. R., Watson, J. W., Risenhoover, K., and McBride, T. (2004). Marking
and radiotelemetry of Oregon spotted frogs (Rana pretiosa). Northwestern Naturalist,
85, 20–5.
McDonough, C. and Paton, P. W. C. (2007). Salamander dispersal across a forested land-
scape fragmented by a golf course. Journal of Wildlife Management, 71, 1163–9.
McShea, W.J. and Madison, D.M. (1992). Alternative approaches to the study of
small mammal dispersal: insights from radiotelemetry. In W. Z. Lidicker Jr and
N. C. Stenseth (eds), Dispersal: Small Mammal as a Model, pp. 319–32. Chapman and
Hall, New York.
Millspaugh, J. J. and Marzluff, J. M. (2001). Radio Tracking and Animal Populations,
Mineau, P. and Madison, D. M. (1977). Radio-tracking of Peromyscus leucopus. Canadian
Okada, S., Utsunomiya, T., Okada, T., and Felix, Z. I. (2006). Radio transmitter attach-
ment by suturing for the Japanese giant salamander (Andrias japonicus). Herpetological
Review, 37, 431–4.
Oldham, R. S. and Swan, M. J. S. (1992). Effects of ingested radio transmitters on Bufo

bufo and Rana temporaria. Herpetological Journal, 2, 82–5.
Peterman, W. E. and Semlitsch, R. D. (2006). Effects of tricaine methanosulfonate
(MS-222) concentration on anesthetization and recovery in four plethodontid sala-
manders. Herpetological Review, 37, 303–4.
Peterman, W. E., Crawford, J. A., and Semlitsch, R. D. (2008). Productivity and sig-
nificance of headwater streams: population structure and biomass of the black-bellied
salamander (Desmognathus quadramaculatus). Freshwater Biology, 53, 347–57.
Rathbun, G. B. and Murphey, T. G. (1996). Evaluation of a radio-belt for ranid frogs.
Rittenhouse, T. A. G. and Semlitsch, R. D. (2006). Grasslands as movement barriers for
a forest-associated salamander: migration behavior of adult and juvenile salamanders
at a distinct habitat edge. Biological Conservation, 131, 14–22.
Rittenhouse, T. A. G. and Semlitsch, R. D. (2007). Postbreeding habitat use of wood
frogs in a Missouri oak-hickory forest. Journal of Herpetology, 41, 645–53.
Rowley, J. J. L. and Alford, R. A. (2007). Techniques for tracking amphibians: the effects
of tag attachment, and harmonic direction finding versus radio telemetry. Amphibia-
Reptilia, 28, 367–76.
Schabetsberger, R., Jehle, R., Maletzky, A., Pesta, J., and Sztatecsny, M. (2004).
Delineation of terrestrial reserves for amphibians: Post-breeding migrations of the
Italian crested newt (Triturus c. carnifex) at high altitudes. Biological Conservation, 117,
95–104.
Spieler, M. and Linsenmair, K. E. (1998). Migration patterns and diurnal use of shelter
in a ranid frog of a West African savannah: a telemetric study. Amphibia-Reptilia, 19,
43–64.
Vasconcelos, D. and Calhoun, A. J. K. (2004). Movement patterns of adult and juvenile
Rana sylvatica (LeConte) and Ambystoma maculatum (Shaw) in three restored seasonal
pools in Maine. Journal of Herpetology, 38, 551–61.
Watson, J. W., McAllister, K. R., and Pierce, D. J. (2003). Home ranges, movements, and
habitat selection of Oregon Spotted Frogs (Rana pretiosa). Journal of Herpetology, 37,
292–300.
Weick, S. E., Knutson, M. G., Knights, B. C., and Pember, B. C. (2005). A comparison
of internal and external radio transmitters with northern leopard frogs (Rana pipiens).
White, G. C. and Garrott, R. A. (1990). Analysis of Wildlife Radio-Tracking Data,
12
Field enclosures and terrestrial cages
Elizabeth B. Harper, Joseph H.K. Pechmann, and
James W. Petranka
12.1 Introduction: amphibians in the

terrestrial environment
Amphibians exhibit many life-history patterns and occur in biomes ranging
from deserts and grasslands to boreal and tropical forests (Chapter 1). Most
species have complex life cycles with a free-living aquatic larval stage and a non-
gilled post-metamorphic stage. Of the more than 6000 extant species that have
been described, almost all use terrestrial habitats during some or all of their life
cycle. Common terrestrial specializations include arboreal, fossorial, surface-
dwelling, and riparian species (Chapter 1).
The aquatic larval stage, if present, is typically much shorter than the post-
metamorphic stage. The juveniles and adults of most species live in semi-aquatic
or terrestrial habitats. Despite these facts, researchers have conducted surprisingly
few studies on the terrestrial ecology of amphibians, especially pond-breeding
amphibians (Pechmann 1995). Even some of the most basic questions—such as
the extent to which population regulation occurs during the aquatic and terres-
trial stages of the life cycle—are poorly resolved because of a shortage of infor-
mation on density dependence during the terrestrial stages.
Conservation biologists have identified factors that are contributing to the
global declines of amphibians (e.g. environmental pollutants, forest fragmen-
tation, urbanization, and road construction), but the relative extent to which
these are affecting the aquatic and terrestrial stages is often uncertain (Biek et al.
2002). Widespread gaps in knowledge about the terrestrial ecology of amphib-
ians continue to hamper our ability to develop effective conservation strategies
for the hundreds of species that are in long-term decline. Experiments using field
enclosures can help resolve critical issues concerning the terrestrial ecology of
amphibians and the ability of animals to tolerate environmental stressors.
Given that the use of enclosures is often expensive and time-consuming, why
should a researcher use them? Enclosures allow researchers to expose animals to
experimental treatments that test explicit hypotheses and to recapture animals
to measure response variables. More importantly, field experiments maximize
realism relative to laboratory or mesocosm studies and provide the strong-
est inferences as to whether the factors under investigation operate in nature
(Hairston 1989; Morin 1998; Boone and James 2005).
When designing enclosure experiments, researchers must balance trade-offs
in experimental design because it is impossible to simultaneously maximize
realism, precision, and generality (Hairston 1989). Precision is maximized by
creating replicates of experimental treatments that are as similar as possible to
one another. Generality, which is the ability to extrapolate to a broad area or
to a number of conditions, is maximized by placing replicate enclosures in as
many locations or conditions as possible. In general, experimental designs that
maximize generality compromise precision (and visa versa), and designs that
maximize realism also compromise precision.
Optimally balancing these trade-offs depends on the questions that are being
addressed. For example, a researcher who wishes to address how density affects
the growth of juvenile amphibians may elect to maximize precision and reduce
experimental error by placing enclosures at one site and in close proximity to one
another (Altwegg 2003). In contrast, a researcher who wishes to determine the
rate at which air-borne agricultural pesticides are accumulating in amphibians
in a nearby mountain range may want to establish test sites at numerous loca-
tions to maximize generality.
In the sections that follow we provide examples of how terrestrial enclosures
have been used to address questions about the ecology and conservation biology
of amphibians. We also provide information on cage designs, add cautionary
notes about experimental artifacts, and discuss trade-offs that should be consid-
ered when designing experiments.
12.2 What are the purposes of terrestrial enclosures?

Terrestrial enclosures have been used by amphibian ecologists for four general
purposes, as follows.
1) To confine amphibians so that they can be subjected to an experimen-
tal treatment. The treatment can be the characteristics of the area itself
(e.g. forested or non-forested; Chazal and Niewiarowski 1998; Todd and
Rothermel 2006), or one applied by the researcher (e.g. population density;
12 Field enclosures and terrestrial cages | 205
Pechmann 1995; Harper and Semlitsch 2007). Enclosures allow these

experiments to be conducted under realistic conditions in the field on
organisms that could otherwise migrate from their assigned treatment.
2) To facilitate recapture of amphibians subjected to experimental treat-
ments before they were placed in the enclosures, for example investigating
the post-metamorphic carryover effects of larval exposure to pesticides
(Boone 2005). In this case, enclosures permit evaluation of treatment
effects under natural climatic conditions and diets, and perhaps under
natural competition and predation regimes, depending on the type and
size of the enclosures.
3) To facilitate recapture of dispersing or migrating amphibians to measure
habitat choice and movement distances (Rothermel and Semlitsch 2002;
Croshaw and Scott 2006). Enclosures in this case consist of walled corri-
dors. Although these topics can also be addressed by tracking unenclosed
individuals (Chapter 11), enclosures allow large groups of animals to be
followed and recaptured. Enclosures can also confine choices and move-
ments to particular habitats. For example, Croshaw and Scott (2006)
examined the elevations marbled salamanders chose for nesting in enclos-
ures in which the amount of cover was equalized across elevations.
4) To estimate terrestrial densities of amphibians (Regosin et al. 2003a,
2005). Enclosing an area containing naturally occurring amphibians
limits their movement and can improve the accuracy of density estimates
obtained by removal sampling or mark–recapture.
12.3 Defining the research question

The first step in designing an effective experiment using terrestrial enclosures is
to clearly define the research question. This is essential in deciding: (1) which
species to use, (2) the location, size, and number of enclosures, (3) methods
of pen construction, (4) the number of amphibians per enclosure, (5) the fre-
quency and methods of recapturing individuals, (6) the duration of the experi-
ment, (7) the response metrics, and (8) how the data will be analyzed. The
majority of research questions fall into two general categories: questions with
habitat types as the experimental treatment and questions with treatments that
can be randomly assigned to enclosures or to individuals within them.
12.3.1 Questions related to habitat types

If the research question examines the effects of specific habitat types, then the
treatment consists of the conditions inside each enclosure and the surrounding
conditions. Examples of habitat-focused research questions include studies of

amphibian responses to clear-cutting (Chazal and Niewiarowski 1998; Todd
and Rothermel 2006), vegetation density (Denton and Beebee 1994), forest
edges (Marsh and Beckman 2004; Rothermel and Semlitsch 2006), and exotic
plants (Maerz et al. 2005). Studies of this sort often do not permit researchers
to randomly assign treatments to experimental units, and particular attention
should be given to issues concerning spatial autocorrelation and pseudorepli-
cation (Hurlbert 1984). To maximize generality, enclosures should be widely
spaced. The trade-off is increased variability within treatments, meaning that
increased replication may be necessary to achieve adequate statistical power.
Fig. 12.1 Small cages can be used to confine amphibians to specific microhabitats
during short-term experiments.
Measuring habitat variables for use as covariates in later analyses can be useful
in understanding the mechanisms underlying this within-treatment variabil-
ity (e.g. aspect and surface temperature; Harper 2007). Densities of predators,
competitors, and prey may also be important explanatory variables.
Research questions focused on habitat type have important implications for
enclosure design and construction. Because the treatment includes the surround-
ing habitat as well as the conditions within the enclosure, realism may be improved
by using mesh rather than solid materials for the enclosure walls. Mesh allows
movement of air, water, nutrients, and small organisms in and out of enclosures
and creates conditions within the enclosure that are more typical of the habitat type
as a whole. Large enclosures ( 200 m2) are more likely to incorporate most micro-
habitats available within a larger habitat type and to provide the range of choices
that would be available to free-roaming individuals, thereby increasing realism.
Large enclosures are also likely to support a more representative community includ-
ing competitors, predators, and small mammals that dig burrows used by some
amphibians (Regosin et al. 2003b). Conversely, if the research question focuses
on microhabitats, small enclosures (Figure 12.1) can purposefully limit choice to
determine the effects of specific microhabitats and their associated organisms on
amphibian performance (Maerz et al. 2005; Rothermel and Luhring 2005).
12.3.2 Questions with treatments that can be

assigned within enclosures
A broad range of research questions can be addressed using treatments that
are randomly assigned to enclosures. Examples include manipulations of dens-
ity (Cohen and Alford 1993; Pechmann 1995; Altwegg 2003; Harper and
Semlitsch 2007), conditions during the larval stage (Boone 2005), and inter-
specific interactions (Southerland 1986; Denton and Beebee 1994; Price and
Shields 2002). Because these questions do not focus on the surrounding habitat
type, enclosures can be constructed within close proximity of one another, per-
haps even sharing walls, while still achieving interspersion of treatments. If walls
are shared, pens should be constructed of solid material so that the individual
enclosures remain independent and are not influenced by treatments applied
to neighboring pens (Chalcraft et al. 2005). Building pens in close proxim-
ity minimizes variability within treatments and increases precision; however, it
compromises generality. A good compromise is to construct pens in blocks that
are spaced widely apart, which allows extension of the results beyond a single
site (Hurlbert 1984). One major advantage of questions that allow treatments
to be assigned randomly within enclosures is that enclosure arrays built for one
experiment can later be used for a range of other experiments (Figure 12.2).
Fig. 12.2 Terrestrial enclosure (1 m 2 m) at the University of Missouri’s Research

Park in Columbia, MO, USA. These enclosures were first built to test the terrestrial
fitness of hybrids in the Rana pipiens complex and have since been used in experiments
focusing on ecotoxicology and density dependence. A hole dug in the center of each
pen and covered with a board provides a moist refuge.
12.4 Constructing enclosures

Decisions involved in enclosure construction include the location, number,
dimensions, and construction materials. Options are typically constrained by
funds, labor, and the logistical constraints unique to each study site.
12.4.1 Location of enclosures

Enclosures built in close proximity to one another maximize precision and
increase statistical power, but compromise generality. Generality is increased
by widely spacing blocks of enclosures hundreds of meters or several kilometers
apart; however, it is important to consider the time and resources required to
build and monitor pens that are widely spaced. Enclosures can be positioned
randomly, haphazardly, or uniformly, across the entire study site or in a stratified
fashion within it. Random or stratified random-site selection is the most statis-
tically defensible, especially if habitat is the treatment. Completely random or
uniform site selection may not be the best option if there are areas within the
habitat that amphibians avoid or areas where it is impossible to build suitable

pens. If the study animals prefer specific microhabitats, then stratified designs
that place all or a disproportionately large number of pens in preferred habitats
may be desirable. Areas that expose amphibians to high temperatures due to lack
of canopy cover or aspect can result in high mortality (Harper 2007), as can lack
of moisture, burrows, or cover objects (Rothermel and Luhring 2005). In some
studies these factors may be part of the treatments, while in others they may be
sources of experimental error that reduce statistical power. Haphazard place-
ment of enclosures may be ideal for ensuring that pens are located in areas where
installation is feasible and habitats seem suitable, but it often compromises gen-
erality and may introduce bias.
12.4.2 Size and number of enclosures

Limited resources explain the inverse relationship between the number and size
of enclosures used in experiments and the prevalence of small (
10 m2) enclo-
sures (Table 12.1). Striking the right balance between the number and size of
enclosures depends largely on the research question and the duration of the
experiment. In general, it is best to first determine the size of enclosure that best
suits the research question and then ascertain the number of enclosures that can
feasibly be built, stocked, and monitored given the circumstances. Then decide
whether this number of enclosures is adequate to answer the research question
without a high probability of type II error (i.e. failing to reject a null hypothesis
when the alternate is true), preferably by doing a statistical power analysis. Keep
in mind that some enclosures may be lost to falling trees or other misfortunes,
especially during long-term experiments.
Small pens compromise realism because they rarely encompass the home
range and associated microhabitats that are used by amphibians. They may
also exclude or restrict the movements of competitors, predators, and prey more
than large pens. This can result in large variation in the organisms included
in replicate pens. Cage effects, including unnatural behaviors resulting from
confinement, are expected to be more prevalent in small enclosures because
the available behavioral options, including habitat selection and predator avoid-
ance, are limited.
Some amphibian species migrate over 1 km among over-wintering habitats,
breeding habitats, and summer foraging areas, and the area of habitat used can
vary dramatically depending on life-history stage and time of year (Beebee
1996). Experiments using these species can still achieve a reasonable level of
realism if the duration of the experiment is limited; for example, to the juven-
ile stage or to the summer foraging period. Realism can also be improved by
Table 12.1 Summary of methods used in terrestrial enclosure experiments.
Reference Species Enclosure Number of Research Stocking Duration of Response metrics
size (m2) enclosures focus density experiment
(per/m2)
Pearson (1955) Scaphiopus holbrookii 74 5 Behavior 0.14–0.54 3 years Activity rates

Jaeger (1971) Plethodon richmondi 0.25 and 6 Competitive 20–40 2 months Survival
shenandoah and 0.5 exclusion
Plethodon cinereus
Southerland Desmognathus 0.25 and 9 Competitive 4–16 2 summers Change in mass and
(1986) quadramaculatus, 0.5 exclusion survival
Desmognathus
monticola,
Desmognathus
ochrophaeus
Southerland Desmognathus 5 4 Competitive 4 28 days Growth and habitat
(1986) quadramaculatus, exclusion selection
Desmognathus
monticola,
Desmognathus
ochrophaeus
Cohen and Bufo marinus 3 15 Density 3.3–16.7 3 weeks Growth and survival
Alford (1993) dependence
Denton and Bufo bufo, 4.5 12 Niche separation 0.89 2 months Growth, survival, and
Beebee (1994) Bufo calamita activity rates
Pechmann (1994) Ambystoma 100 16 Density 0.28–0.84 6 years Survival, size, and age at
talpoideum dependence first reproduction
Pechmann (1994) Gastrophryne 100 12 Density 0.6–2.4 6 years Survival, size, and age at
carolinensis dependence first reproduction
Pechmann (1995) Ambystoma opacum 225 18 Density 0.31–0.62 5 years Survival, size, and age at
and Ambystoma dependence, first reproduction
talpoideum competition
Chazal and Ambystoma 100 8 Forestry 0.8 5–6 months Growth, fecundity, age
Niewiarowski (1998) talpoideum practices at maturity, lipid storage
Hopkins et al. Bufo terrestris Not 16 Contaminant Not 7–12 weeks Body concentrations of
(1998) reported exposure reported trace elements
Beck and Bufo terrestris 0.5 2 Size-dependent 52 2 months Growth and survival
Congdon (1999) growth and
survival
Laposata and Ambystoma 0.1134 20 Contaminant 8.8 1 month Growth, body
Dunson (1999) jeffersonianum exposure concentrations
of water and elements
Parris (2001) Rana blairi, Rana 2 48 Hybrid fitness 3.5 1 year Growth and survival
sphenocephala and
hybrids
Beard et al. Eleutherodactylus 1 20 Nutrient cycling 7 4 months Concentrations of
(2002) coqui carbon and
nitrogen in the body
urine and feces
Table 12.1 Continued
(per/m2)
Price and Shields Plethodon cinereus and 0.26 52 Interspecific 3.8–7.71 8 weeks Growth
(2002) Plethodon glutinosus interactions
Rothermel and Ambystoma maculatum, 125 8 Juvenile Variable (total 2 months Rate of movement and
Semlitsch (2002) Ambystoma texanum, emigration n 179) distance moved
Bufo americanus
Rothermel and Ambystoma maculatum, 0.025 8 Dehydration 401 22–25 h Change in mass
Semlitsch (2002) Ambystoma texanum,
Bufo americanus
Altwegg (2003) Rana lessonae 9 12 Density 2.8–8.3 1 year Growth and survival
dependence
Altwegg and Reyer Rana lessonae and 9 12 Carryover effects 2.8–8.3 1 year Growth and survival
(2003) Rana esculenta
Regosin et al. Rana sylvatica 272 17 Over-wintering Not stocked 9 months Sex ratio, density,
(2003a) ecology distance from pond
Marsh and Beckman Plethodon cinereus 0.81 24 Road effects 2.5 2 months Detectibility and
(2004) surface activity
Moseley et al. Ambystoma 20.88 and 9 and 18 Microhabitat use Not reported 8 months Movement and habitat
(2004) talpoideum 9 selection
Regosin et al. Ambystoma 3.8 10 Burrow-occupancy 0.26–0.531 8 days Probability of burrow
(2004) maculatum patterns occupancy
Williams and Plethodon cinereus 1.18 124 Detection 0.85–1.71 2 months Proportion detected
Berkson (2004) probabilities
Yurewicz and Plethodon cinereus 0.81 28 Cost of 1.21 2 years Egg survival, female
Wilbur (2004) reproduction growth, production of
ova
Boone (2005) Rana sphenocephala, 2 48 Contaminant 3–5 7–12 months Survival and growth
Rana blairi, Rana exposure
clamitans,
Bufo woodhousii
Maerz et al. Rana clamitans 0.06 38 Foraging success 171 38 h Change in mass
(2005)
Rothermel and Ambystoma 0.025 48 Forestry practices 401 72 h Water loss and
Luhring (2005) talpoideum survival
Croshaw and Scott Ambystoma opacum 112–223 4 Nest-site selection 10.76 1 month Number of nests per
(2006) area
Greenlees et al. Bufo marinus 2.88 60 Effects on native Standard Not reported Invertebrate
(2006) invertebrates biomass abundance
Rothermel and Ambystoma maculatum 18 12 Forest 1.3 2 years Survival to maturity
Semlitsch (2006) and Ambystoma opacum fragmentation
Todd and Bufo terrestris 16 8 Forestry practices 1.8 2 months Growth and survival
Rothermel (2006)
Harper (2007) Rana sylvatica and 9 64 Forestry practices 2 3 months Survival
Bufo americanus
Table 12.1 Continued
(per/m2)
Harper and Rana sylvatica, 2 48 Density 1–10 1 year Survival, growth,

Semlitsch (2007) Bufo americanus dependence reproductive
development
Todd et al. (2007) Ambystoma opacum 0.025 96 Fire ant predation 401 48 h Survival
and Ambystoma
talpoideum
Blomquist (2008) Rana pipiens and 14.4 28 Forestry practices 1.73 and 1.39 3 months Survival and growth
Rana sylvatica
Patrick et al. (2008) Rana sylvatica 100 and 8 1 and 24 Density dependence 2–7 2–3 weeks Survival and habitat
and movement selection
1
Densities represent one animal per enclosure.
choosing species that have relatively small home ranges, as is often the case for
direct-developing species (Price and Shields 2002).
Large enclosures maximize realism, allowing increased movement and access
to a greater range of microhabitats and interacting organisms. One important
trade-off is that the probability of recapturing individuals decreases as enclosure
size increases. On the other hand, large enclosures can hold more individuals
and provide more precise estimates of mean responses.
Large enclosures can also allow densities to be at or near “natural levels”.
Unfortunately, estimates of natural densities for terrestrial life history stages
are not available for most amphibian species, and density can have significant
effects on growth and survival (Figure 12.3). It may be worthwhile to estimate
the natural densities of study species if estimates are unavailable. Terrestrial
densities can be extremely high following metamorphosis, but quickly decrease
thereafter due to mortality and migration (Cohen and Alford 1993). Therefore,
short-term experiments using recently metamorphosed individuals can use high
Fig. 12.3 The density of animals in enclosures can have profound effects on
individual growth rates, probabilities of survival, and age at sexual maturity. These
American toads were the same size at metamorphosis and were raised in similar
enclosures, but at densities of 1/m2 (left) and 10/m2 (right). This photo was taken 3
months after the animals were introduced to enclosures.
densities without sacrificing realism (Beck and Congdon 1999), but experiments
that are longer term or use adults may require larger enclosures to approximate
natural densities. Because amphibians occur at higher densities in high-quality
habitat (Patrick et al. 2008), ensuring high habitat quality within enclosures
(e.g. sufficient moisture, cover, and prey) can improve realism when enclosures
are small and densities are relatively high.
12.4.3 Building enclosures that minimize

escapes and trespasses
Most amphibians, especially fossorial species, are notorious escape artists and
enclosure designs should ensure that study animals do not escape. Very secure
pens are often less permeable to biotic and abiotic components of the ecosystem,
a feature which may or may not be desirable (see section 12.3). Walls made of
mesh materials such as hardware cloth allow greater exchange with the surround-
ing environment but are more easily climbed by some species than smooth, solid
materials such as aluminum flashing. Even taxa that do not usually climb, such
as Ambystoma and Bufo, will climb walls to get out of enclosures (E.B. Harper
and J.H.K. Pechmann, personal observations).
Enclosures for fossorial species may require a mesh or solid bottom (Marsh
and Beckman 2004; Williams and Berkson 2004; Yurewicz and Wilbur 2004).
In the case of caecilians, it is necessary to use solid materials because caecilians
can easily penetrate through flexible plastic (J. Measey, personal communica-
tion). Pre-made plastic containers, ranging in size from shoe boxes to large rain
barrels or cattle tanks, can serve as enclosures for fossorial species (Price and
Shields 2002; Williams and Berkson 2004). Caecilians have successfully been
enclosed in plastic rain barrels punctured with small drainage holes and filled
with sifted soil. To prevent caecilians from escaping, the container should be
filled to a level that leaves an unfilled top portion that is greater in length than
the longest individual (J. Measey, personal communication).
For less effective burrowers, for example most ranids, enclosures do not
require a bottom, but usually have sides that are buried in well-packed soil 5–50 cm
below ground depending on the species and size of enclosure. Small enclos-
ures, for example those constructed from buckets or coffee cans (Rothermel
and Semlitsch 2002; Maerz et al. 2005; Rothermel and Luhring 2005), can be
installed with a shovel or post digger. Large enclosures are typically built by first
digging a trench, then driving wooden, metal, or PVC posts into the ground at
each corner and at points in between. Walls are then attached to the posts, and
can be built with sheet metal, aluminum flashing, galvanized hardware cloth,
or silt fencing. These materials vary considerably in price and durability, but
even the least expensive silt fencing will survive at least a full year under most
conditions. After the walls are secured, soil is backfilled into the trench and
compressed so that a portion of the wall is underground. If pitfall traps are used
to recapture animals, they can be installed before the trench is filled. During
construction, soil, leaf litter, and vegetation are often disturbed. To maximize
realism, the time between the building of the enclosures and the introduction of
animals should be sufficient to allow accumulation of leaf litter and invertebrate
prey, and the regrowth of vegetation.
Animals are most likely to attempt to escape from the corners of the enclosures,
by climbing up the posts, or through the overlapping ends of the material used for
the walls. Rounded corners can be more difficult to climb, and placing the support
posts 10cm or more below the top of the wall can also reduce escapes. Strips of add-
itional material across the top edge of the walls can serve as baffles, both to keep
enclosed animals in and to prevent animals from natural populations from enter-
ing (Figure 12.4). Quick detection and repair of damaged or worn enclosures is
also necessary to prevent escapes. Enclosures with lids may be necessary for arbor-
eal species (Beard et al. 2002), and can be useful for other species to exclude preda-
tors, provide shade, and reduce escapes (Harper and Semlitsch 2007). However,
Fig. 12.4 Array of terrestrial enclosures (10 m 10 m each) used in studies of

density dependence in Ambystoma talpoideum and Gastrophryne carolinensis at the
Savannah River Ecology Laboratory in South Carolina, USA. Note baffles at the top of
the walls to deter escapes and immigration.
lids typically influence light, temperature, rainfall, litterfall, and movement of

prey, which can hinder realism and may confound the treatment.
12.4.4 “Standardizing” conditions among enclosures

One advantage of terrestrial enclosures is that they can expose animals to the
range of conditions found within a study area, such as the densities of prey,
competitors, and predators, as well as variation in leaf-litter depth, vegetation,
and solar radiation. Unfortunately, this variability decreases statistical power
by contributing to experimental error. One option to reduce this “noise” is to
standardize conditions within enclosures. For example, quantities of leaf litter
and woody debris can be equalized (Harper 2007). Devices that provide wet
refuges, such as a pit with moist leaf litter, a bowl of water, or a covering object,
can reduce the risk of desiccation (Greenlees et al. 2006), and reduce variation
in survival within treatments, especially if amphibians must be released into
enclosures during dry periods when they would not normally be active. Predators
including snakes may have to be removed, especially from small pens, because of
the difficulty of equalizing predation across enclosures without excessive mor-
tality. Competition can also be standardized by removing competitors or by
introducing an equal number to all enclosures. Standardization comes at the
cost of reduced realism, and it is important to consider how this may comprom-
ise the interpretation of the results. This is especially important when biotic and
abiotic factors comprise the treatments.
12.5 Study species

The choice of study species is an important consideration when designing
enclosure experiments. It must be possible to confine the species under realis-
tic conditions and densities and to have access to enough animals to design an
experiment with high statistical power.
12.5.1 Choice of species

Terrestrial enclosures are most appropriate for species that are primarily terrestrial
after metamorphosis. For species that are both aquatic and terrestrial (e.g. Rana
clamitans and Desmognathus monticola), short-term terrestrial enclosure experi-
ments that coincide with normal habitat use can be informative (Maerz et al.
2005). Enclosures that encompass both aquatic and terrestrial habitats can also
be used (Southerland 1986). For arboreal species, enclosures must be completely
covered to prevent escape, which decreases realism. Effective jumpers, such as
Acris, are difficult to keep in enclosures without very high walls or covers.
12.5.2 Source and age of animals

Animals can be wild-caught or raised from eggs in captivity (e.g. in artificial
ponds). Both methods have advantages and disadvantages in terms of logis-
tics, experimental design, and conservation. Using wild-caught animals is often
necessary for studies of adult amphibians when raising individuals to adulthood
is unfeasible due to constraints of time, space, and/or expense. Research on
the ecology of adult amphibians is important because amphibian populations
are especially sensitive to changes in adult vital rates (Biek et al. 2002); for this
same reason, however, the effects on wild populations should be carefully con-
sidered before adults are removed. Raising individuals through metamorphosis
in outdoor tanks or experimental ponds is often the best method for studies of
juvenile pond-breeders for several reasons. First, large numbers of individuals
can be produced with minimal effects on wild populations. Survival to meta-
morphosis in predator-free aquatic enclosures and tanks often exceeds 50–80%
compared with fewer than 5–10% in many wild populations (Petranka and
Sih 1986). Secondly, the conditions experienced by larvae can be controlled,
thereby maximizing precision. Experimental ponds can also be used in con-
junction with terrestrial enclosures to study carryover effects by manipulating
conditions during the larval period and observing their effects on terrestrial
life-history stages (Pechmann 1994; Altwegg and Reyer 2003; Boone 2005).
When collecting eggs for future experimental populations, one should consider
the genetic make-up in the context of the experiment. In some instances it may
be prudent to select eggs from multiple populations or from multiple clutches
within a population so the results are more representative of the species or popu-
lation of interest.
12.6 Census techniques

For most enclosure studies, gathering data requires recapturing amphibians
during and/or at the end of the study. Gathering data at regular intervals rather
than waiting until the end of the experiment is often a wise strategy because
mortality can be higher than expected. Multiple census intervals are also useful
in estimating capture probabilities, which are often necessary to calculate esti-
mates of survival (Chapter 24).
12.6.1 Individual marks

Individually marking animals before they are released into enclosures allows
monitoring of individual growth rates, estimation of capture probabilities,
and analysis of factors such as initial mass or time to metamorphosis (Beck
and Congdon 1999). There are many techniques that can be used to identify
individuals, including toe-clipping, visible implant elastomer, passive integra-
tive transponder tags (PIT tags), and photographing unique dorsal patterns
(Chapter 8). Group marks may be necessary when it is impractical to mark all
individuals. In general, the marking technique chosen should strive to minimize
the increased probability of mortality and stress to the animals while maximiz-
ing the probability of successful identification.
12.6.2 Methods of capture

The most common methods of recapturing amphibians in enclosures are hand
capture, pitfall traps, and coverboards (Chapter 13). A combination of methods
can be used to maximize capture probability. Hand captures work well for ani-
mals that are active on the surface or that jump when the leaf litter is disturbed.
However, this method can result in biased data if animals are easier to see and
catch in some treatments than others. Pitfall traps (Chapter 13) may provide a
more standardized method for recapturing amphibians and are especially useful
for fossorial species like ambystomatid salamanders that are not easily captured
by hand (Pechmann 1995). Traps can be covered or filled with leaf litter when
not in use and opened on rainy nights when the animals are most likely to be
active. A wet sponge in the bottom of the traps can reduce desiccation risks,
although only sponges that are free of chemicals should be used. Cover objects
such as rotting logs or boards can also be placed in enclosures as a method
for recapturing animals (Chapter 13) as long as their addition does not con-
found the effects of the treatment. Similarly, burrows can be created within
enclosures to provide retreats where animals can predictably be found (Regosin
et al. 2004). For caecilians and other truly fossorial amphibians that are rarely
active on the surface, it may be necessary to do a single census at the end of the
experiment by removing and thoroughly sifting through the soil and leaf lit-
ter. Radiotransmitters or harmonic radar tags (Moseley et al. 2004) can make
study animals easier to relocate, but the equipment may affect behavior and
expense can be an issue. Animals marked with PIT tags can be detected up to
13 cm underground (Blomquist et al. 2008). Pond-breeding amphibians can be
censused along the enclosure walls during breeding migrations by pitfall traps,
funnel traps, or hand capture as they attempt to migrate out of the enclosure to
find a breeding pond (Pechmann 1995; Regosin et al. 2003b).
12.6.3 Frequency of censuses

Because enclosure experiments usually require an initially large input of time,
labor, and resources, it is worthwhile in longer-term experiments to collect data
early and often to ensure that useful data are gathered even if natural disas-
ters or unexpectedly high mortality occur before the end of the experiment.
Frequent censuses also improve estimates of capture probability and survival
(Chapter 25). However, it is important to consider that handling animals too
frequently can affect survival and induce unnecessary stress. The optimal fre-
quency of censuses for estimating capture probability and survival also depends
on the statistical methods that will be used. Some methods require equal time
between sampling intervals or a minimum number of sample periods. Other
statistical methods rely on multiple consecutive samples spaced at regular inter-
vals (e.g. 3 days of sampling every 2 weeks). These methods are discussed in
greater detail in Chapter 25. It is important to become familiar with these stat-
istical methods prior to gathering the data to ensure that the frequency and
methods used in the censuses maximize statistical power.
12.7 Response metrics

Experiments conducted in enclosures can use a wide range of response vari-
ables to describe the effects of experimental treatments on amphibians. Vital
rates, including growth and survival, are among the most commonly used met-
rics (Table 12.1), but more subtle effects can be quantified with physiological
measurements such as lipid content (Chazal and Niewiarowski 1998) or stress-
hormone concentrations (Cooperman et al. 2004). Behavioral observations,
including habitat selection, can also yield useful response metrics (Patrick et al.
2008). Data that quantify conditions within enclosures can provide covariates
for analyses or can be used to provide metrics describing the role of amphibians
in the ecosystem, for example in nutrient cycling (Beard et al. 2002) or reducing
prey densities (Greenlees et al. 2006). In most enclosure experiments the enclos-
ure is the experimental unit, so for any response metric that is based on the indi-
vidual (e.g. mass), means per enclosure (or the equivalent) should be used in the
analysis (Pechmann 1995). Proportions can also be used (e.g. percentage survival
or the percentage that selected a given habitat type; Patrick et al. 2008).
12.7.1 Vital rates: survival, growth, age at

reproductive maturity, fecundity
Survival can be surprisingly difficult to measure in all but very small enclos-
ures. The minimum number known alive (MNKA) can be used as a proxy for
survival (Rothermel and Semlitsch 2006), but does not allow the estimation
of error resulting from undetected individuals. Because capture probabilities
are rarely 100%, capture–mark–recapture models can be useful for estimating
survival (Altwegg 2003; Chapter 24). However, when capture probabilities are
low, there is a large amount of error associated with these estimates. Methods
for incorporating errors from mark–recapture estimates into other statistical
analyses of multiple experimental populations, such as analysis of variance, have
received limited attention. Survival estimates can be improved by using an end-
point associated with breeding migrations, such as first reproduction, because
migrations increase the likelihood of capture (Pechmann 1995; Regosin et al.
2003b).
Growth is usually measured as change in mass or length over time. Some
researchers prefer measuring snout–vent length, often with additional measure-
ments including head width or tibia length (Altwegg and Reyer 2003), because
mass is prone to short-term fluctuations.
Reproductive maturity can be determined by dissection, candling, or exter-
nal signs of maturity such as nuptial pads or a swollen cloaca (Pechmann 1995;
Rothermel and Semlitsch 2006; Harper and Semlitsch 2007). Age at repro-
ductive maturity or the proportion of individuals that have reached maturity at
the conclusion of the experiment are demographically important response met-
rics. Although fecundity is a component of individual fitness and population
growth, clutch size has not typically been used as a response metric for enclosure
experiments (but see Yurewicz and Wilbur 2004).
12.7.2 Physiological responses

Response metrics that quantify physiological responses are especially useful
when the effects of experimental treatments are likely to be sublethal. Many of
the methods used to quantify physiological responses require that animals be
killed, so it is necessary to gather these data at the end of the experiment, or by
removing a subset of animals at points during the experiment. Lipids stored in
the body or in eggs can be measured and used as indicators of body condition
and parental investment (Chazal and Niewiarowski 1998). Whole-body con-
centrations of elements such as arsenic, selenium, and vanadium can indicate
bioaccumulation of contaminants (Hopkins et al. 1998; Laposata and Dunson
1999). Heightened concentrations of corticosterone circulating in the blood
indicate a stress response, and can be determined using non-lethal methods
(Cooperman et al. 2004).
12.7.3 Behavioral responses

Measures of behavioral responses in enclosure experiments have typically focused
on movement and habitat selection. Movement has been quantified as activity
rate (Pearson 1955; Rothermel and Semlitsch 2002; Marsh and Beckman 2004;
Moseley et al. 2004) and distance moved (Rothermel and Semlitsch 2002).
Studies of habitat selection typically quantify the proportion of individuals that
select a particular habitat type (Southerland 1986; Moseley et al. 2004; Patrick
et al. 2008). Regosin et al. (2004) used probability of burrow occupancy as a
response metric.
12.8 Thinking outside the box

Amphibian studies using terrestrial enclosures and cages have so far focused on
a relatively narrow range of species and experimental designs. However, there
is great scope for creativity in the use of these techniques. Some of the most
exciting and least studied aspects of amphibian ecology can be studied using
terrestrial enclosures. For example, a study by Beard et al. (2002) used cages
as “exclosures” to document the effects of both the presence and absence of a
terrestrial frog on nutrient cycling. Creative techniques for answering complex
amphibian research questions have also involved using enclosures in conjunc-
tion with other experimental techniques (Marsh and Beckman 2004; Todd and
Rothermel 2006) or conducting multiple enclosure experiments at a range of
spatial and temporal scales (Patrick et al. 2008).
Although most amphibian enclosure experiments have been conducted in
temperate regions, there is enormous potential for the use of enclosures in trop-
ical ecosystems. Many tropical amphibians may actually be far better suited
for enclosure experiments than their temperate counterparts. Leaf-litter species
such as the neotropical Eleutherodactylus, Arthroleptis of Africa, or many of the
Asian microhylids, occur at remarkably high densities and are unlikely to have
large home ranges. Well-designed enclosures can also be used to gain a better
understanding of the ecology of caecilians, for which few data are currently
available.
12.9 References
Altwegg, R. (2003). Multistage density dependence in an amphibian. Oecologia, 136,
46–50.
Altwegg, R. and Reyer, H.-U. (2003). Patterns of natural selection on size at metamor-
phosis in water frogs. Evolution, 57, 872–82.
Beard, K. H., Vogt, K. A., and Kulmatiski, A. (2002). Top-down effects of a terrestrial
frog on forest nutrient dynamics. Oecologia, 133, 583–93.
Beck, C. W. and Congdon, J. D. (1999). Effects of individual variation in age and size
at metamorphosis on growth and survivorship of southern toad (Bufo terrestris) meta-
morphs. Canadian Journal of Zoology, 77, 944–51.
Beebee, T. J. C. (1996). Ecology and Conservation of Amphibians. Chapman & Hall,

London.
Biek, R., Funk, W. C., Maxell, B. A., and Mills, L. S. (2002). What is missing in amphib-
ian decline research: insights from ecological sensitivity analysis. Conservation Biology,
16, 728–34.
Blomquist, S. M. (2008). Relative Fitness and Behavioral Compensation of Amphibians in a
Managed Forest. PhD dissertation, University of Maine, Orono, ME.
Blomquist, S. M., Zydlewski, J. D., and Hunter, M. L. (2008). Efficacy of PIT tags for
tracking the terrestrial anurans Rana pipiens and Rana sylvatica. Herpetological Review,
39, 174–9.
Boone, M. D. (2005). Juvenile frogs compensate for small metamorph size with terrestrial
growth: overcoming the effects of larval density and insecticide exposure. Journal of
Boone, M. D. and James, S. M. (2005). Aquatic and terrestrial mesocosms in amphibian
ecotoxicology. Applied Herpetology, 2, 231–57.
Chalcraft, D. R., Binckley, C. A., and Resetarits Jr, W. J. (2005). Experimental venue and
estimation of interaction strength: comment. Ecology, 86, 1061–7.
Chazal, A. C. and Niewiarowski, P. H. (1998). Responses of mole salamanders to clear-
cutting: using field experiments in forest management. Ecological Applications, 8,
1133–43.
Cohen, M. P. and Alford, R. A. (1993). Growth, survival and activity patterns of recently
metamorphosed Bufo marinus. Wildlife Research, 20, 1–13.
Cooperman, M. D., Reed, J. M., and Romero, L. M. (2004). The effects of terrestrial and
breeding densities on corticosterone and testosterone levels in spotted salamanders,
Ambystoma maculatum. Canadian Journal of Zoology, 82, 1795–1803.
Croshaw, D. A. and Scott, D. E. (2006). Marbled salamanders (Ambystoma opacum)
choose low elevation nest sites when cover availability is controlled. Amphibia-Reptilia,
27, 359–64.
Denton, J. S. and Beebee, T. J.C. (1994). The basis of niche separation during terrestrial
life between two species of toad (Bufo bufo and Bufo calamita): competition or special-
isation? Oecologia, 97, 390–8.
Greenlees, M. J., Brown, G. P., Webb, J. K., Phillips, B. L., and Shine, R. (2006). Effects
of an invasive anuran [the cane toad (Bufo marinus)] on the invertebrate fauna of a trop-
ical Australian floodplain. Animal Conservation, 9, 431–8.
Hairston Sr, N. G. (1989). Ecological Experiments: Purpose, Design, and Execution.
Cambridge University Press, Cambridge.
Harper, E. B.H. (2007). The Population Dynamics and Conservation of Pond-breeding
Amphibians. PhD dissertation, University of Missouri, Columbia, MO.
Harper, E. B. and Semlitsch, R. D., (2007). Density dependence in the terrestrial life his-
tory stage of two anurans. Oecologia, 153, 879–89.
Hopkins, W. A., Mendonca, M. T., Rowe, C. L., and Congdon, J. D., (1998). Elevated
trace element concentrations in southern toads, Bufo terrestris, exposed to coal combus-
tion waste. Archives of Environmental Contamination and Toxicology, 35, 325–9.
Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiments.
Jaeger, R.G. (1971). Competitive exclusion as a factor influencing the distributions of two
species of terrestrial salamanders. Ecology, 52, 632–7.
Laposata, M. M. and Dunson, W. A. (1999). Enclosure design for ecotoxicological stud-
ies of terrestrial salamanders. Herpetological Review, 30, 28–30.
Maerz, J. C., Blossey, B., and Nuzzo, V. (2005). Green frogs show reduced foraging
success in habitats invaded by Japanese knotweed. Biodiversity and Conservation, 14,
2901–11.
Marsh, D. M. and Beckman, N. G. (2004). Effects of forest roads on the abundance and
activity of terrestrial salamanders. Ecological Applications, 14, 1882–91.
Morin, P. J. (1998). Realism, precision, and generality in experimental tests of ecological
theory. In W. J. Resetarits and J. Bernardo (eds), Issues and Perspectives in Experimental
Ecology, pp. 50–70. Oxford University Press, Oxford.
Moseley, K. R., Castleberry, S. B., and Ford, W. M. (2004). Coarse woody debris and pine
litter manipulation effects on movement and microhabitat use of Ambystoma talpoi-
deum in a Pinus taeda stand. Forest Ecology and Management, 191, 387–96.
Parris, M. J. (2001). Hybridization in leopard frogs (Rana pipiens complex): terrestrial
performance of newly metamorphosed hybrid and parental genotypes in field enclos-
ures. Canadian Journal of Zoology, 79, 1552–8.
Patrick, D. A., Harper, E. B., Hunter Jr, M. L., and Calhoun, A. J. K. (2008). Terrestrial
habitat selection and strong density-dependent mortality in recently metamorphosed
amphibians. Ecology, 89, 2563–74.
Pearson, P. G. (1955). Population ecology of the spadefoot toad, Scaphiopus h. holbrooki
(Harlan). Ecological Monographs, 25, 233–67.
Pechmann, J. H. K. (1994). Population Regulation in Complex Life Cycles: Aquatic and
Terrestrial Density-dependence in Pond-breeding Amphibians. PhD dissertation, Duke
University, Durham, NC.
Pechmann, J. H. K. (1995). Use of large field enclosures to study the terrestrial ecology of
pond-breeding amphibians. Herpetologica, 51, 434–50.
Petranka, J. W. and Sih, A. (1986). Environmental instability, competition and density-
dependent growth and survivorship of a stream-dwelling salamander. Ecology, 67,
729–36.
Price, J. E. and Shields, J. A.S. (2002). Size-dependent interactions between two terrestrial
amphibians, Plethodon cinereus and Plethodon glutinosus. Herpetologica, 58, 141–55.
Regosin, J. V., Windmiller, B. S., and Reed, J. M. (2003a). Terrestrial habitat use and win-
ter densities of the Wood Frog (Rana sylvatica). Journal of Herpetology, 37, 390–4.
Regosin, J. V., Windmiller, B. S., and Reed, J. M. (2003b). Influence of abundance of
small-mammal burrows and conspecifics on the density and distribution of spot-
ted sala manders (Ambystoma maculatum) in terrestrial habitats. Canadian Journal of
Zoology, 81, 596–605.
Regosin, J. V., Windmiller, B. S., and Reed, J. M. (2004). Effects of conspecifics on the
burrow occupancy behaviour of spotted salamanders. Copeia, 2004, 152–8.
Regosin, J. V., Windmiller, B. S., Homan, R. N., and Reed, J. M. (2005). Variation
in terrestrial habitat use by four poolbreeding amphibian species. Journal of Wildlife
Rothermel, B. B. and Semlitsch, R. D. (2002). An experimental investigation of land-

scape resistance of forest versus old-field habitats to emigrating juvenile amphibians.
Rothermel, B. B. and Luhring, T. M. (2005). Burrow availability and desiccation risk
of mole salamanders (Ambystoma talpoideum) in harvested versus unharvested forest
stands. Journal of Herpetology, 39, 619–26.
Rothermel, B. B. and Semlitsch, R. D. (2006). Consequences of forest fragmentation for
juvenile survival in spotted (Ambystoma maculatum) and marbled (Ambystoma opacum)
salamanders. Canadian Journal of Zoology, 84, 797–807.
Southerland, M. T. (1986). Coexistence of three congeneric salamanders: the importance
of habitat and body size. Ecology, 67, 721–8.
Todd, B. and Rothermel, B. B. (2006). Assessing quality of clearcut habitats for amphib-
ians: effects on abundances versus vital rates in the southern toad (Bufo terrestris).
Todd, B. D., Rothermel, B. B., Reed, R. N., Luhring, T. M., Schlatter, K., Trenkamp, L.,
and Gibbons, J. W. (2007). Habitat alteration increases invasive fire and abundance to
the detriment of amphibians and reptiles. Biological Invasions, 10, 539–46.
Williams, A. K. and Berkson, J. (2004). Reducing false absences in survey data: detection
probabilities of red-backed salamanders. Journal of Wildlife Management, 68, 418–28.
Yurewicz, K. L. and Wilbur, H. M. (2004). Resource availability and costs of reproduc-
tion in the salamander Plethodon cinereus. Copeia, 2004, 28–36.
Part 4
Amphibian populations
13
Drift fences, coverboards, and other traps
John D. Willson and J. Whitfield Gibbons
13.1 Introduction
Many of the simplest yet most highly productive sampling methods in her-
petological field research use some type of trap or attraction device to increase
capture rates or target secretive species. These techniques fall into two general
categories: those that actually trap animals, accumulating captures on their own
over time (passive traps) and those that attract animals but require an observer
to actively capture them at the moment of the census (active traps). The most
popular examples of these two trap categories are drift fences (generally with
pitfall and/or funnel traps) and coverboards, respectively. Both of these methods
are inherently simple concepts, and their description and explanation need not
be made complex or complicated. Both techniques are usually best modified by
the investigator who can use common sense to focus on the needs of a particular
project that involves capturing animals in a field situation. However, we provide
a general discussion of some of the fundamental issues that investigators who
use these techniques must face, with particular emphasis on how choice of cap-
ture method and sampling design influence interpretation of capture data.
13.2 Drift fences, funnel traps, and other

passive capture methods
13.2.1 What are passive traps?
Passive capture methods are designed to restrain animals that enter the trap of
their own accord, accumulating captures that are then assessed upon regular cen-
suses by the observer. Passive traps are among the more intensive methods for sam-
pling amphibians in terms of time and effort, but often yield higher capture rates
and more standardized samples than opportunistic or visual searches. Perhaps
most importantly, passive traps are the most effective ways to sample many rare or
secretive amphibian species, many of which are of conservation concern.
A huge variety of passive amphibian traps have been developed for nearly any
imaginable habitat or situation; however, nearly all are variants on two basic trap
types, the funnel trap and the pitfall trap. Funnel traps consist of a tapering funnel-
shaped entrance that guides animals into a larger holding chamber (Figure 13.1).
Once within the chamber, animals are unable to find their way back out the small
entrance hole, becoming trapped. Pitfall traps work on a similar principle, con-
sisting of some type of container that is sunk into the ground, with the rim level
with the surface (Figure 13.2). Animals that fall into pitfalls are unable to climb
out, becoming trapped. First described for use with herpetofauna by Gibbons and
Bennett (1974) and Gibbons and Semlitsch (1981), drift fences are vertical bar-
riers that intercept the intended trajectory of amphibians moving from one loca-
tion to another. The fence typically guides animals toward a pitfall bucket, funnel
trap, or other capture device (Figure 13.2). In most terrestrial and some aquatic
situations, drift fences dramatically increase amphibian capture rates (Friend et al.
1989) and, under some circumstances, have been responsible for more amphibian
captures per day, month, year, or decade than any other method used in field stud-
ies of amphibians (Pechmann et al. 1991; Gibbons et al. 2006). Although drift
Fig. 13.1 Several varieties of funnel trap are commonly used to sample amphibians
in terrestrial and aquatic habitats. Front: soft-drink bottle funnel trap; back (left
to right): plywood and hardware cloth box trap, steel “Gee” minnow trap, plastic
minnow trap, and collapsible nylon trap. Photograph by John Willson, Savannah River
Ecology Laboratory.
13 Drift fences, coverboards, and other traps | 231
fences have been used most extensively in the southeastern USA, they have proven
effective for many amphibian species worldwide (e.g. Gittins 1983; Friend 1984;
Bury and Corn 1987; Friend et al. 1989; Jehle et al. 1995; Weddeling et al. 2004).
However, drift fences were relatively ineffective for sampling anurans in forests of
Queensland, Australia (Parris et al. 1999) and recorded lower numbers of anuran
species than automated recording devices or nocturnal line-transect surveys in
Taiwan (Hsu et al. 1985).
13.2.2 How are passive traps constructed,

aligned, and monitored?
Funnel traps can be nearly any size and have been constructed of a variety of
materials including twine netting, hardware cloth, window screen, nylon mesh,
plywood, PVC pipe, and plastic soft-drinks bottles (Figure 13.1; Griffiths 1985;
Shaffer et al. 1994; Adams et al. 1997; Buech and Egeland 2002; Willson et al.
2005). The wide range of funnel trap variants makes them effective for most
species in nearly any habitat, aquatic or terrestrial. Likewise, pitfalls can be any
size from small coffee cans to multi-gallon drums, but must be sufficiently deep
to prevent escape of the target species. Pitfall traps are typically metal or plastic.
Several studies have been conducted weighing the strengths and weaknesses of
different trap types for various amphibian species (e.g. Vogt and Hine 1982;
Friend 1984; Friend et al. 1989; Mitchell et al. 1993; Greenberg et al. 1994;
Enge 2001; Ryan et al. 2002; Stevens and Paszkowski 2005; Todd et al. 2007),
and we refer readers to these sources rather than discussing the merits of each
trap type in detail here.
Metal stakes
Plastic cable ties

Aluminum-
flashing
drift fence
Plastic bucket pitfall trap
Drainage holes
Fig. 13.2 Schematic of a terrestrial drift fence with large pitfall traps. Figure reprinted
from Gibbons and Semlitsch (1981).
The evolution of the drift-fence technique in herpetology has resulted in

the use of a variety of construction materials for the fence, including chicken
wire, hardware cloth, aluminum flashing, stiff plastic, and erosion/silt fencing
(Gibbons and Bennett 1974; Dodd and Scott 1994; Enge 1997a). Likewise,
habitat constraints and the traits of particular target species have given rise to
a variety of suggestions for the alignment of fencing within the habitat and an
endless array of configurations of fences and traps is possible (see Corn 1994;
Dodd and Scott 1994; Rice et al. 2006). No single construction material or
trap type is universally “the best” because of several factors whose importance
will vary depending upon the particular project. Among the issues that must
be considered are goals of the study and how the data will be analyzed, the
cost and availability of materials, anticipated longevity of the project and main-
tenance effort required, the size and behavior of target species, potential safe-
guards against predators, and the terrain and topography of the habitat itself.
Mechanisms to minimize trespass by climbing species, drowning of captured
individuals in buckets or dehydration in buckets or funnel traps, and predation
by a variety of species continue to be developed to address specific situations.
Rather than dictate what the most effective materials and trap configuration
should be, we recommend that the investigator tailor drift-fence applications on
the basis of budgets, time availability, and the general goals and specific object-
ives of the project. Following are examples of how differing goals can temper the
style of fencing and type of traps.
First, perhaps the situation for which drift fences are most commonly
employed is to intercept wetland-breeding amphibians as they undertake sea-
sonal breeding migrations between terrestrial refugia and aquatic breeding
sites (Figure 13.3a; e.g. Gittins 1983; Pechmann et al. 1991; Dodd and Scott
1994; Arntzen et al. 1995; Weddeling et al. 2004). In these applications the
wetland could be completely encircled by a drift fence, allowing enumeration
of the entire annual breeding population at that site. Alternatively, if the wet-
land is large or if resources are limited, smaller partial sections of drift fence
could be placed at intervals around all or part of the aquatic/terrestrial inter-
face (Figure 13.3a). Some studies have used concentric circular drift fences to
monitor terrestrial dispersal of pond-breeding salamanders away from wet-
lands (Johnson 2003). Trap types used to monitor pond-breeding amphib-
ians could vary depending on the target species. For studies focusing solely
on ambystomatid salamanders, for example, small, coffee-can-sized pitfalls
would be sufficient to capture individuals of all species, including the largest
(Ambystoma tigrinum; Gibbons et al. 2006). Large plastic buckets, however,
(a)
Small
wetland or
Large wetland
(b) Habitat 1
Habitat 2
Drift-fence array Replicated arrays
Fig. 13.3 Examples of drift-fence and passive-trap configurations used to address

different research questions. Solid lines represent sections of drift fence, filled
circles represent pitfalls, and open rectangles represent funnel traps. (a) Drift-
fence configurations used to sample pond-breeding amphibians migrating to and
from wetland breeding sites; (b) a drift-fence array pattern designed to compare
amphibian abundance across two habitat types.
would be necessary to capture high numbers of most species, especially lar-

ger anurans (e.g. ranid frogs and toads) that could jump out of a smaller can
(Mitchell et al. 1993). Larger buckets also increase the time that climbing hylids
and microhylids are likely to remain within the trap. In some situations, funnel
traps are the best single trap type for capturing many amphibian species (Enge
2001; Todd et al. 2007); however, funnel traps are generally more costly to
construct, more time-consuming to maintain, and more prone to desiccating
captured amphibians than are pitfalls. In nearly all situations, a combination of
trap types (e.g. alternating large pitfalls and funnel traps) will produce the best
possible assessment of the entire amphibian community (Vogt and Hine 1982;
Greenberg et al. 1994; Todd et al. 2007).
Second, another common application of drift fences in terrestrial habitats is to
compare abundances of amphibians in different areas (Figure 13.3b; e.g. habitat
types, experimental treatments). In such cases, fences are generally located far
from any obvious breeding wetland, hibernaculum, or other habitat focal point.
For such applications, drift fences are often constructed in cross- or X-shaped
arrays with a central trap and traps placed along each section of the fence (see
Corn 1994). Because the goal of this type of study is generally to test hypoth-
eses about amphibian abundance between areas, each array makes a convenient
sampling unit for statistical comparisons. In this case, care should be taken to
ensure that arrays are comparable (same length of fencing, number of traps, etc.)
and that arrays are located randomly or systematically across the treatments or
habitats.
Finally, passive traps can be used for quantitative sampling of aquatic
amphibians that are not easily captured by other methods. For this applica-
tion, a variety of aquatic funnel traps can be effective, with the size of the
trap being dictated by the size of the target species. For small species or life
stages, small funnel traps made from plastic soft-drinks bottles would suffice
(e.g. Griffiths 1985; Willson and Dorcas 2003), while targeting larger spe-
cies such at the giant salamanders, Siren and Amphiuma, would require lar-
ger traps such as commercially available minnow or crawfi sh traps (Johnson
and Barichivich 2004; Willson et al. 2005). In most cases, aquatic traps
should be set in water shallow enough to allow captured animals access to
air and care should be taken to monitor fluctuations in water level that could
submerge traps. Specific microhabitats where traps are set will vary by spe-
cies, but heavily vegetated shallow areas are often preferable. Capture rates
may also be increased by setting aquatic traps along natural barriers such as
submerged logs or the shoreline or by the use of aquatic drift fences to direct
amphibians into traps (Enge 1997b; Willson and Dorcas 2004; Palis et al.
2007). Finally, as in the previous example, traps may be set in any number
of spatial confi gurations. Often a simple linear transect along a shoreline
is sufficient. However, if the goal of the study is to compare captures stat-
istically across different treatments (habitats, wetlands, etc.) standardized
arrays of traps can be set that will serve as the sampling units in statistical
comparisons.
Passive traps restrain captured animals, so frequent (generally at least daily)
monitoring of traps is necessary to avoid mortality of captured animals. During
hot or dry periods it is often advisable to provide access to moisture (e.g. a water
bowl or damp sponge) within traps to avoid desiccation of captured animals.
Finally, natural amphibian predators such as mid-sized mammals, birds, and
large snakes often learn to target drift fences, and predator-control measures
(raised covers for pitfalls, live-trapping and removal, or wide-width steel mesh
trap covers, etc.) may be necessary to avoid undue predation.
13.2.3 What can passive traps tell you?

What can they not tell you?
Passive traps, especially when used in conjunction with drift fences, have proved
highly effective for determining the distribution and abundance of amphib-
ians both spatially and temporally. Many examples exist in which drift-fence
captures revealed the presence of species that were rare or not even known to
be present in an area, as well as providing a comparative assessment of annual
and seasonal activity among species. Drift fences have also been used to capture
large numbers of specific life stages of study species for laboratory experiments.
The application of drift fences to a conservation effort was aptly demonstrated
by Aresco (2005), who used silt fencing to create a barrier to prevent amphibians
and reptiles from crossing a busy highway. Despite considerable hand-waving
about the statistical approaches that could and should be applied to drift-fence
data, as long as an investigator is aware of potential biases in the effectiveness of
the fence and what is being revealed, the technique remains one that unques-
tionably can reveal natural history information about amphibians that may be
unobtainable or unlikely to be discovered in any other way.
Generally, amphibians are only captured in passive traps when they are
actively moving through the area where traps are deployed. Thus, the number
of amphibians captured (often expressed as a rate, such as captures per trap per
night) is ultimately a function of three major factors: (1) the density of animals
within the area sampled, (2) the activity (movement) levels of those animals,
and (3) the probability that an individual animal encountering a trap will be
captured and not escape. Although trap capture rates can be extremely inform-
ative, consideration of all three factors is critical to interpreting capture data.
In situations where the population density of amphibians can be assumed to be
relatively constant (e.g. within a single population over a fairly short time), most
of the variation in capture rates can be assumed to be due to shifts in activity.
Thus, drift fences have been instrumental in allowing investigators to identify
seasonality and orientation of amphibian breeding migrations and environmen-
tal correlates of migratory activity (e.g. Semlitsch 1985; Semlitsch and Pechmann
1985; Phillips and Sexton 1989; Todd and Winne 2006). Likewise, when trap
captures are pooled over relatively long intervals (e.g. seasons or years), thus
minimizing short-term variation in activity due to environmental conditions,
long-term shifts in abundance or activity can be assessed (e.g. Jehle et al. 1995).
Ideally, studies wishing to use trap capture rates as abundance indices should
test the assumption of equal catchability through mark–recapture, occupancy
modeling, or similar methods (Chapter 24; Mazerolle et al. 2007). Additionally,
marking captured animals allows the researcher to distinguish between novel
individuals and recaptures, improving census counts by eliminating multiple

counts of the same individual animal (Weddeling et al. 2004).
Although it is often tempting to attempt interspecific comparisons of abun-
dance using capture-rate data, such comparisons are nearly always tenuous,
given that species often differ in seasonal timing of activity levels and vary in
their susceptibility to being captured by a particular trap type. For example,
some species, such as many hylid and microhylid frogs, are adept at climbing
out of buckets and over fences, whereas even small pitfalls are highly effective for
capturing many salamanders and terrestrial anurans (Friend et al. 1989; Dodd
1991; Enge 2001; Todd et al. 2007). Thus, it would clearly be inappropriate
to conclude that because more salamanders were captured in drift fences, they
were actually more abundant than treefrogs within the habitat. Some conserva-
tive interspecific comparisons of abundance are possible, however, given careful
consideration of potential biases. For example, if a wetland were to be completely
surrounded by a drift fence with large pitfalls for an entire breeding season, it
would probably be safe to compare total annual breeding population sizes of
ambystomatid salamander species using that wetland. However, Arntzen et al.
(1995) demonstrated that even for relatively small terrestrial species (Bufo bufo
and Triturus cristatus), drift-fence efficiencies (proportion of breeding popula-
tion captured) were often low and varied between species.
13.3 Coverboards and other traps that

require active capture
13.3.1 What are coverboards and other active traps?
Unlike passive traps such as funnel traps or pitfalls, some so-called traps do
not actually restrain or capture animals, but instead concentrate free-ranging
amphibians to facilitate their capture by an active observer (usually by hand).
For example, a herpetologist may lay down boards or other artificial cover
objects that attract amphibians and allow them to be captured without disturb-
ing natural cover such as logs, rocks, or vegetation. Such active traps generally
operate on the principle of creating optimal microhabitats for the target species,
attracting animals that can then be collected more easily than would otherwise
be possible.
Because many amphibians are partially or exclusively fossorial for much of
their lives and most prefer moist habitats, coverboards and other artificial cover
objects are the most widely used active traps for amphibians. Coverboards sim-
ply consist of sections of cover material, most commonly wood or metal, which
are placed on the ground in habitats preferred by target species. Coverboards

act in the same way as natural-cover objects, trapping moisture and providing
refugia for a variety of amphibian species. Amphibians are captured when an
observer gently lifts the cover object and collects any animals observed hiding
beneath. Because coverboards generally create moist subterranean microhabi-
tats, this technique is most frequently used for amphibian species that prefer
such habitats and include most terrestrial salamander species and some of the
more fossorial anurans such as toads (e.g. Bufo and Scaphiopus) and microhylid
frogs (e.g. Gastrophryne).
Other active capture methods have been developed, many for specific species
or situations. One example is the use of PVC pipes for collecting hylid treefrogs
(Moulton et al. 1996; Boughton et al. 2000). Pipes are placed vertically within
the habitat, creating a moist arboreal microclimate favored by treefrogs for diur-
nal refugia. Pittman et al. (2008) used an extensive grid of PVC refugia within a
wetland and surrounding upland habitats to document seasonal activity patterns,
habitat use, and site fidelity in a North Carolina population of gray treefrogs
(Hyla chrysoscelis). Additional examples of active traps include leaf-litter bags to
aid in capture of aquatic salamanders and their larvae (Pauley and Little 1998;
Waldron et al. 2003) and artificial pools to assess breeding activity of various
anuran species (Resetarits and Wilbur 1991; Gascon 1994).
13.3.2 How are active traps constructed,

aligned, and monitored?
Coverboards may be constructed from nearly any material and the most suit-
able material likely varies depending on the target species, research budget,
habitat, and other characteristics of the study site (e.g. proximity to roads,
terrain). For most amphibians, wooden boards are probably the best all-round
option as they create moist conditions preferred by many species. Indeed, wood
coverboards have been used in studies of a variety of woodland and stream-
dwelling salamander species (Figure 13.4; Degraaf and Yamasaki 1992; Fellers
and Drost 1994; Houze and Chandler 2002; Moore 2005; Luhring and Young
2006). Moore (2005) suggested using boards cut in situ from native trees for
studying redback salamanders (Plethodon cinereus), providing a cost-effective
coverboard option that can be used in habitats that are difficult to traverse.
Although coverboards consisting of roofing tin or other metals are frequently
used in reptile studies, these materials heat quickly and create conditions too
hot and dry for amphibians in most situations. Indeed, Grant et al. (1992)
found that amphibian captures in South Carolina were highest under plywood
coverboards, while reptiles preferred tin. As most amphibian species are small,
Fig. 13.4 Example of a coverboard used to monitor woodland salamanders. Note

that the low barrier around the board is part of an experiment and would normally
not be used with coverboards used for general monitoring purposes. Photograph by
Thomas Luhring, Savannah River Ecology Laboratory.
boards generally need not be large, but at least one study reported a positive
correlation between the size of coverboards and the number of salamanders
captured per board (Moore 2005). Moreover, larger or thicker boards may be
preferable in warm or dry habitats as they generally hold moisture better than
smaller boards. One study conducted in southern Georgia, USA, noted lower
amphibian captures under coverboards than natural-cover objects, presumably
resulting from warmer and more variable temperatures under boards (Houze
and Chandler 2002). As some time is often necessary for suitable microhabi-
tats (e.g. rotten leaf litter, burrows) to develop under refugia, the investiga-
tor should consider allowing boards to “weather” for several weeks or months
before amphibian censuses are initiated.
Construction of other active traps varies, but the general goal is to create
microhabitat conditions that attract target amphibian species. For example,
sections of PVC pipe may be inserted vertically into the ground or affixed to
tree trunks to provide arboreal refugia for hylid treefrogs (Moulton et al. 1996;
Boughton et al. 2000). Johnson (2005) described several modifications to the
so-called hylid tube technique, maximizing standardization of microhabitat
within tubes and increasing ease of census and frog capture.
As uses of coverboards and other active traps vary, so will the designs for their
placement. In general active traps should be placed in habitats that are favorable
for amphibians, including well-shaded areas with abundant moisture. For spe-
cies that breed in aquatic habitats, breeding sites may be the most appropriate
places to maximize captures. For example, an obvious location to deploy PVC
tubes for hylid frogs would be around the periphery of wetland breeding sites.
The spatial distribution of active traps also depends on the goals of the study.
For simple amphibian inventories (documenting species presence) or collecting
individuals for use in the laboratory, placing devices haphazardly in the most
optimal habitats is the most cost-effective method. For studies where statistical
comparisons are to be made, replicated sampling units must be designated. In
general, because captures per individual trap are low, arrays of several boards or
other traps are usually designated as the sampling unit. An array can consist of
any arbitrary number of traps, generally arranged in a systematic pattern (e.g.
a grid or transect). Replicate arrays are then placed systematically or randomly
within treatments. Ideally, a power analysis can be used to determine the num-
ber of arrays (sample size) that is needed to obtain sufficient statistical power for
the analysis.
For example, a researcher might wish to compare salamander abundance
across three forest types within a relatively small geographical area. Having
determined that salamanders are generally fairly common (say, an average of
one salamander found per five coverboards checked), an array of 10 cover-
boards, spaced 5 m apart in a linear transect, might be determined as the sam-
pling unit. Geographic Information Systems (GIS) technology could then be
used to generate randomized locations for five arrays to be placed within each
of the three habitat types. Thus, the total number of boards to be used would
be 150 (10 boards/array 5 arrays per habitat treatment 3 habitats) and the
sample size would be five per treatment.
Unlike passive traps, active traps can be monitored on nearly any tem-
poral schedule, and the timing of censuses will reflect the question of inter-
est. Generally, as animals may remain within the same refuge for extended
periods, allowing some time (e.g. a few days or a week) between censuses may
minimize repeated counts of the same individual animal. Alternatively, ani-
mals may avoid boards that are disturbed too frequently; one study noted that
salamander captures under coverboards that were checked daily were reduced
compared to boards that were censused on longer time intervals (1 or 3 weeks;
Marsh and Goicochea 2003). Similarly, when designing a monitoring scheme,
care should be taken to make samples as repeatable as possible. This typic-
ally means that environmental conditions should be as comparable as possible
between samples, and many researchers set up environmental criteria for deter-
mining census times based on the biology of the study animal. For example,
coverboard arrays might be checked once-weekly at 7–9 am on days without
precipitation or only on nights with a temperature greater than 15°C and at
least 1 cm of rain.
13.3.3 What can active traps tell you?

What can they not tell you?
Coverboards and other active traps have several advantages as sampling methods
for amphibians. First, in many cases, these are among the best methods for col-
leting large numbers of target species and can be the only ways to collect highly
fossorial species or those that do not form breeding aggregations. Moreover,
unlike passive traps, which require high-intensity monitoring on a daily basis
to avoid mortality of captured animals, active traps can be monitored on a low-
intensity or periodic basis because animals are not restrained and are not prone
to accidental mortality or unnaturally high levels of predation. Also, because
active traps generally concentrate animals into a highly searchable area (e.g.
under a board, or in a PVC tube), these methods minimize the effects of obser-
ver bias, which can be substantial in other active capture methods such as visual
searches. Finally, because artificial refugia are standardized for size and material,
they create more repeatable microhabitats than natural-cover objects, yielding
more standardized measures of abundance, and can be censused with minimal
disturbance to the habitat (Heyer et al. 1994).
Because active traps yield high capture rates of target species, they can be use-
ful for assessing patterns of abundance over time or space (e.g. population trends
over time or variation in abundance across habitats). Generally, studies designed
to investigate these types of questions compare capture rates among arrays of
traps placed in different locations (e.g. two habitat types) or capture rates within
an array or set of arrays over time. In these types of studies, the dependent vari-
able is generally an index of relative abundance, typically the number of animals
captured over some unit of time and effort (e.g. captures per array per census).
However, as with any abundance index, data collected using these methods may
be biased in a variety of ways, all of which must be considered when interpreting
capture data.
The key assumption when comparing indices of relative abundance is that
the capture probability of an individual animal is the same across arrays or time
intervals. Thus, differences in capture rate among arrays reflect true differences
in population density between sampling units. Although this assumption may
be met in some situations, it is relatively easy to imagine situations where this
premise is violated. Perhaps the most obvious case where the equal catchability
assumption is violated is when comparing capture rates among species, as noted
for the drift-fence method. It is often the case, regardless of the sampling method,
that some species are highly catchable, while others are seldom encountered.
Thus, it is generally unreasonable to assume that simply because one species is
captured more frequently than another, that it is truly more abundant. Likewise,
differences in catchability across time are critical to consider in temporal com-
parisons and the potential for activity patterns to influence catchability must
be considered when interpreting capture rates. For example, in warm climates,
salamanders often retreat deep underground during the summer and are seldom
captured using any method. In this case, low capture rates of salamanders in the
summer are best explained by seasonal differences in catchability rather than by
changes in actual abundance within the landscape. A growing body of literature
uses mark–recapture or occupancy modeling to incorporate detection probabil-
ity (catchability) in interpretations of amphibian abundance data (Chapter 24;
Mazerolle et al. 2007). Ideally, any researcher wishing to use abundance indices
as indicators of population size or density should consider using these methods to
test the equal catchability assumption.
A final consideration when using active traps is the potential for use of these
methods to actually improve the overall quality of the habitat for target spe-
cies. For example, if one of the factors limiting population size in an amphibian
species is availability of suitable refugia, adding boards, hylid tubes, or other
artificial refugia to the habitat may permit an increase in population density.
Although such a situation would probably be favorable in studies designed to
conserve at-risk amphibian species, it is worth remembering that population
densities estimated in areas with abundant artificial refugia may not necessarily
be representative of those in unaltered habitats. To our knowledge the poten-
tial for artificial refugia to improve habitat quality for amphibians has not been
addressed experimentally and warrants future investigation.
13.4 References
Adams, M. J., Richter, K. O., and Leonard, W. P. (1997). Surveying and monitoring
amphibians using aquatic funnel traps. In D. H. Olson, W. P. Leonard, and R. B. Bury
(eds), Sampling Amphibians in Lentic Habitats, Northwest Fauna, no. 4, pp. 47–54.
Aresco, M. J. (2005). Mitigation measures to reduce highway mortality of turtles and other
herpetofauna at a north Florida lake. Journal of Wildlife Management, 69, 549–60.
Arntzen, J. W., Oldham, R. S., and Latham, D. M. (1995). Cost effective drift fences for
toads and newts. Amphibia-Reptilia, 16, 137–45.
Boughton, R. G., Staiger, J., and Franz, R. (2000). Use of PVC pipe refugia as a sampling
technique for hylid treefrogs. American Midland Naturalist, 144, 168–77.
Buech, R. R. and Egeland, L. M. (2002). Efficacy of three funnel traps for capturing
amphibian larvae in seasonal forest ponds. Herpetological Review, 33, 182–5.
Bury, R. B. and Corn, P. S. (1987). Evaluation of pitfall trapping in northwestern forests:
trap arrays with drift fences. Journal of Wildlife Management, 51, 112–19.
Corn, P. S. (1994). Straight-line drift fences and pitfall traps. In W. R. Heyer,
M. A. Donnelly, R. W. McDiarmid, and L. C. Hayek (eds), Measuring and Monitoring
Biological Diversity. Standard Methods for Amphibians, pp. 109–17. Smithsonian
Degraaf, R. M. and Yamasaki, M. (1992). A nondestructive technique to monitor the
relative abundance of terrestrial salamanders. Wildlife Society Bulletin, 20, 260–4.
Dodd, Jr, C. K. (1991). Drift fence-associated sampling bias of amphibians at a Florida
sandhills temporary pond. Journal of Herpetology, 25, 296–301.
Dodd, Jr, C. K. and Scott, D. E. (1994). Drift fences encircling breeding sites. In
W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, and L. C. Hayek (eds), Measuring
and Monitoring Biological Diversity. Standard Methods for Amphibians, pp. 125–30.
Enge, K. M. (1997a). Use of silt fencing and funnel traps for drift fences. Herpetological
Review, 28, 30–1.
Enge, K. M. (1997b). A Standardized Protocol for Drift-fence Surveys. Technical report
no. 14. Florida Game and Fresh Water Fish Commission, Tallahassee, FL.
Enge, K. M. (2001). The pitfalls of pitfall traps. Journal of Herpetology, 35, 467–78.
Fellers, G. M. and Drost, C. A. (1994). Sampling with artificial cover. In W. R. Heyer,
M. A. Donnelly, R. W. McDiarmid, and L. C. Hayek (eds), Measuring and Monitoring
Biological Diversity. Standard Methods for Amphibians, pp. 146–50. Smithsonian
Friend, G. R. (1984). Relative efficiency of two pitfall-drift fence systems for sampling
small vertebrates. Australian Zoologist, 21, 423–34.
Friend, G. R., Smith, G. T., Mitchell, D. S., and Dickman, C. R. (1989). Influence of pit-
fall and drift fence design on capture rates of small vertebrates in semi-arid habitats of
western Australia. Australian Wildlife Research, 16, 1–10.
Gascon, C. (1994). Sampling with artificial pools. In W. R. Heyer, M. A. Donnelly,
R. W. McDiarmid, and L. C. Hayek (eds), Measuring and Monitoring Biological
Diversity. Standard Methods for Amphibians, pp. 144–6. Smithsonian Institution Press,
Washington DC.
Gibbons, J. W. and Bennett, D. H. (1974). Determination of anuran terrestrial activity
patterns by a drift fence method. Copeia 1974, 236–43.
Gibbons, J. W. and Semlitsch, R. D. (1981). Terrestrial drift fences with pitfall traps: an
effective technique for quantitative sampling of animal populations. Brimleyana, 7,
1–16.
Gibbons, J. W., Winne, C. T., Scott, D. E., Willson, J. D., Glaudas, X., Andrews, K. M.,
Todd, B. D., Fedewa, L. A., Wilkinson, L., Tsaliagos, R. N. et al. (2006). Remarkable
amphibian biomass and abundance in an isolated wetland: Implications for wetland
conservation. Conservation Biology, 20, 1457–65.
Gittins, S. P. (1983). The breeding migration of the common toad (Bufo bufo) to a pond
in mid-Wales. Journal of Zoology, London, 199, 555–62.
Grant, B. W., Tucker, A. D., Lovich, J. E., Mills, A. M., Dixon, P. M., and Gibbons, J. W.
(1992). The use of coverboards in estimating patterns of reptile and amphibian bio-
diversity. In D. R. McCullough, and R. H. Barrett (eds), Wildlife 2001, pp. 379–403.
Elsevier Science, London.
Greenberg, C. H., Neary, D. G., and Harris, L. D. (1994). A comparison of herpetofaunal
sampling effectiveness of pitfall, single-ended, and double-ended funnel traps used
with drift fences. Journal of Herpetology, 28, 319–24.
Griffiths, R. A. (1985). A simple funnel trap for studying newt populations and an evalu-
ation of trap behavior in smooth and palmate newts, Triturus vulgaris and Triturus
helveticus. Herpetological Journal, 1, 5–10.
Heyer, W. R., Donnelly, M. A., McDiarmid, R. W., and Hayek, L. C. (eds) (1994).
Measuring and Monitoring Biological Diversity. Standard Methods for Amphibians,
Houze, C. M. and Chandler, C. R. (2002). Evaluation of coverboards for sampling terres-
trial salamanders in south Georgia. Journal of Herpetology, 36, 75–81.
Hsu, M. Y., Kam, Y. C., and Fellers, G. M. (2005). Effectiveness of amphibian monitor-
ing techniques in a Taiwanese subtropical forest. Herpetological Journal, 15, 73–9.
Jehle, R., Hodl, W., and Thonke, A. (1995). Structure and dynamics of central European
amphibian populations: a comparison between Triturus dobrogicus (Amphibia,
Urodela) and Pleobates fuscus (Amphibia, Anura). Australian Journal of Ecology, 20,
362–6.
Johnson, J. R. (2005). A novel arboreal pipe-trap designed to capture the gray treefrog
(Hyla versicolor). Herpetological Review, 36, 274–7.
Johnson, S. A. (2003). Orientation and migration distances of a pond-breeding salaman-
der (Notophthalmus perstriatus, Salamandridae). Alytes, 21, 3–22.
Johnson, S. A. and Barichivich, W. J. (2004). A simple technique for trapping Siren lac-
ertina, Amphiuma means, and other aquatic vertebrates. Journal of Freshwater Ecology,
19, 263–9.
Luhring, T. M. and Young, C. A. (2006). Innovative techniques for sampling stream-
inhabiting salamanders. Herpetological Review, 37, 181–3.
Marsh, D. M. and Goicochea, M. A. (2003). Monitoring terrestrial salamanders: biases
caused by intense sampling and choice of cover objects. Journal of Herpetology, 37,
460–6.
Mazerolle, M. J., Bailey, L. L., Kendall, W. L., Royle, J. A., Converse, S. J., and Nichols,
J. D. (2007). Making great leaps in herpetology: accounting for detectability in field
studies. Journal of Herpetology, 41, 672–89.
Mitchell, J. C., Erdle, S. Y., and Pagels, J. F. (1993). Evaluation of capture techniques for
amphibian, reptile, and small mammal communities in saturated forested wetlands.
Wetlands, 13, 130–6.
Moore, J. D. (2005). Use of native wood as a new coverboard type for monitoring red-
backed salamanders. Herpetological Review, 36, 268–71.
Moulton, C. A., Fleming, W. J., and Nerney, B. R. (1996). The use of PVC pipes to cap-
ture hylid frogs. Herpetological Review, 27, 186–7.
Palis, J. G., Adams, S. M., and Peterson, M. J. (2007). Evaluation of two types of
commercially-made aquatic funnel traps for capturing ranid frogs. Herpetological
Review, 38, 166–7.
Parris, K. M., Norton, T. W., and Cunningham, R. B. (1999). A comparison of tech-
niques for sampling amphibians in the forests of south-east Queensland, Australia.
Pauley, T. K. and Little, M. (1998). A new technique to monitor larval and juvenile sala-
manders in stream habitats. Banisteria, 12, 32–6.
Pechmann, J. H. K., Scott, D. E., Semlitsch, R. D., Caldwell, J. P., Vitt, L. J., and
Gibbons, J. W. (1991). Declining amphibian populations: the problem of separating
human impacts from natural fluctuations. Science, 253, 892–5.
Phillips, C. A. and Sexton, O. J. (1989). Orientation and sexual differences during
breeding migrations of the spotted salamander, Ambystoma maculatum. Copeia, 1989,
17–22.
Pittman, S. E., Jendrek, A. L., Price, S. J., and Dorcas, M. E. (2008). Habitat selection
and site fidelity of Cope’s gray treefrog (Hyla chrysoscelis) at the aquatic-terrestrial eco-
tone. Journal of Herpetology, 42, 378–85.
Resetarits, W. J. and Wilbur, H. M. (1991). Calling site choice by Hyla chrysoscelis: effects
of predators, competitors, and oviposition sites. Ecology, 72, 778–86.
Rice, A. N., Rice, K. G., Waddle, J. H., and Mazzotti, F. J. (2006). A portable non-
invasive trapping array for sampling amphibians and reptiles. Herpetological Review,
37, 429–30.
Ryan, T. J., Philippi, T., Leiden, Y. A., Dorcas, M. E., Wigley, T. B., and Gibbons, J. W.
(2002). Monitoring herpetofauna in a managed forest landscape: effects of habitat
types and census techniques. Forest Ecology and Management, 167, 83–90.
Semlitsch, R. D. (1985). Analysis of climatic factors influencing migrations of the sala-
mander Ambystoma talpoideum. Copeia, 1985, 477–89.
Semlitsch, R. D. and Pechmann, J. H.K. (1985). Diel patterns of migratory activity for
several species of pond-breeding salamanders. Copeia, 1985, 86–91.
Shaffer, H. B., Alford, R. A., Woodward, B. D., Richards, S. J., Altig, R. G., and Gascon, C.
(1994). Quantitative sampling of amphibian larvae. In W. R. Heyer, M. A. Donnelly,
R. W. McDiarmid, and L. C. Hayek (eds), Measuring and Monitoring Biological
Diversity. Standard Methods for Amphibians, pp. 130–41. Smithsonian Institution
Press, Washington DC.
Stevens, C. E. and Paszkowski, C. A. (2005). A comparison of two pitfall trap designs in
sampling boreal anurans. Herpetological Review, 36, 147–9.
Todd, B. D. and Winne, C. T. (2006). Ontogenetic and interspecific variation in timing
of movement and responses to climatic factors during migrations by pond-breeding
amphibians. Canadian Journal of Zoology, 84, 715–22.
Todd, B. D., Winne, C. T., Willson, J. D., and Gibbons, J. W. (2007). Getting the drift:
examining the effects of timing, trap type, and taxon on herpetofaunal drift fence sur-
veys. American Midland Naturalist, 158, 292–305.
Vogt, R. C. and Hine, R. L. (1982). Evaluation of techniques for assessment of amphibian
and reptile populations in Wisconsin. In N. J. Scott, Jr (ed.). Herpetological Communities,
pp. 201–17. Wildlife Research Report 13. United States Fish and Wildlife Service,
Washington DC.
Waldron, J. L., Dodd, Jr, C. K., and Corser, J. D. (2003). Leaf litterbags: factors affecting
capture of stream-dwelling salamanders. Applied Herpetology, 1, 23–6.
Weddeling, K., Hachtel, M., Sander, U., and Tarkhnishvili, D. (2004). Bias in estimation
of newt population size: a field study at five ponds using drift fences, pitfalls, and funnel
traps. Herpetological Journal, 14, 1–7.
Willson, J. D. and Dorcas, M. E. (2003). Quantitative sampling of stream salamanders:
comparison of dipnetting and funnel trapping techniques. Herpetological Review, 34,
128–30.
Willson, J. D. and Dorcas, M. E. (2004). A comparison of aquatic drift fences with
traditional funnel trapping as a quantitative method for sampling amphibians.
Willson, J. D., Winne, C. T., and Fedewa, L. A. (2005). Unveiling escape and capture
rates of aquatic snakes and salamanders (Siren spp. and Amphiuma means) in commer-
cial funnel traps. Journal of Freshwater Ecology, 20, 397–403.
14
Area-based surveys
David M. Marsh and Lillian M.B. Haywood
14.1 Introduction: what are area-based surveys?

Area-based surveys are frequently used to estimate the relative abundance, dens-
ity, or diversity of amphibians. While the methodological details of area-based
surveys can vary considerably, the basic principles behind these surveys are usu-
ally the same. Small, defined units, referred to as plots, quadrats, or transects,
are selected randomly from larger areas of interest. These small units are then
sampled for amphibians, and the data collected are used to make inferences
about the relative abundance or diversity of amphibians in the larger areas.
Area-based surveys can be employed in a variety of situations. For example,
one might want to compare amphibians in primary forest to amphibians in sec-
ondary forest or pasture. Or, one might want to analyze changes in amphibian
communities over time (i.e. monitoring) or across an environmental gradient
like elevation or distance from a disturbance. Area-based surveys are also used in
experimental studies, for instance when habitats are modified and amphibians
are compared between treatment and control areas. Finally, area-based studies
are commonly used to study the basic habitat relationships and community
ecology of amphibians (e.g. Toft 1980; Hairston 1987).
14.1.2 Why use area-based surveys?

Area-based surveys are useful primarily because they are rigorous and repeat-
able and because they can be replicated to put confidence intervals on estimates
of abundance or diversity. Although haphazard surveys (i.e. walking around a
large area and looking for amphibians) can yield count data and species lists,
data from haphazard surveys are limited in their utility. One problem with hap-
hazard surveys is that there is no way to repeat them consistently. If, years later,
one wants to re-survey the same site or compare the site to some other area, there
is simply no way to collect comparable data. Second, haphazard surveys result

in single estimates with no information on variation. As a result, any measure
of abundance or diversity will lack confidence intervals and hypothesis testing
will often be impossible.
Time-constrained surveys, in which a set amount of time is spent searching
for amphibians, are one step more rigorous than haphazard surveys. However,
when carried out in very large areas, time-constrained surveys lead to many of
the same problems concerning rigor and repeatability as do haphazard surveys.
Controlling the time period involved in a survey is important, but this can be
done within the context of area-based surveys.
In this chapter, we first outline the different kinds of area-based surveys and
illustrate their use in the amphibian research literature. We then discuss when
each of the different kinds of area-based surveys is preferable. Finally, we con-
sider the major methodological issues involved in designing area-based surveys.
In the interest of brevity, we focus on uses of area-based surveys in amphibian
conservation and management rather than basic population or community ecol-
ogy. Most of these uses involve the collection of count data (which are assumed
to reflect relative abundance; see Chapters 24 and 25) or else the estimation of
true density. Throughout this chapter, we use the term abundance to indicate
relative abundance or counts and density to signify an estimate of true density.
Our recommendations concerning area-based sampling are based in part on a
survey of approximately 100 studies that used area-based sampling to collect
data on amphibian abundance or diversity (see Table 14.1). They are also based
Table 14.1 Summary of a literature review of 89 studies that used area-based sampling.
Studies are divided by whether the primary sampling units were plots/quadrats, transects,
transects nested within plots, or quadrats nested within plots or transects. Summary data
show the percentage of studies of each type that were: (1) terrestrial, aquatic, or both;
(2) tropical or temperate; (3) single-species or multi-species studies, (4) targeting frogs,
salamanders, caecilians, or multiple groups; and (5) conducted during daytime, at night,
or both.
Approach N Terrestrial/ Tropical/ Single/ Frogs/salamanders/ Day/night/
aquatic/both temperate multiple caecilians/multiple both
species
Plots/quadrats 33 53/37/10 34/66 34/66 39/26/8/26 55/28/17

Transects 29 34/45/21 34/66 52/48 55/14/0/31 35/56/9
Nested 15 67/7/26 27/73 33/67 40/47/0/13 38/50/12
transects
Nested 12 92/0/8 42/58 50/50 42/8/50/0 29/0/71
quadrats
14 Area-based surveys | 249
on our own experience using plots and transects to sample terrestrial salaman-
ders, stream salamanders, and neotropical frogs.
14.2 Kinds of area-based survey

There are two basic types of area-based survey: plot or quadrat surveys and
transect surveys. The terms plot and quadrat are often used interchangeably
and refer to rectangular or square sampling units. Plots and quadrats can span
a wide range of sizes, from 0.5 m2 stream-bed quadrats to forest plots of several
hectares. Our literature review found that when plots were used as a primary
sampling unit, their median dimensions were 25 m (range 4–400 m) by 20 m
(range 2–240 m). The great variation in plot size in the literature highlights the
need to adapt plot designs specifically to the species and habitat of interest.
Transects are long, narrow plots, designed to be searched by one investigator
in a single pass. Our review found that the median length of amphibian transects
was 100 m (range 7–2000 m) with a median width of 2 m (range 1–8 m). The
most common transect lengths were 50 m and 100 m and the most common
widths were 1 and 2 m. Transects are often sampled visually, but frogs can also be
sampled by listening for calls (see Chapter 16).
In addition to these basic designs, researchers often use nested designs in
which one type of sampling unit is contained within another (e.g. transects or
quadrats nested within large plots). In these designs, researchers randomly select
plot locations within the areas of interest, and then randomly choose quadrat or
transect locations within the plots. These nested quadrats or transects are used
to make inferences about the abundance or diversity of amphibians within the
plots, and the plots are used to make inferences about amphibians in the larger
areas.
14.3 Specific examples of area-based surveys

Beyond the general classification of plots, quadrats, and transects, there are many
specific uses for area-based surveys. In this section, we give several examples
chosen to reflect the diversity of area-based surveys and some of their more com-
mon uses.
14.3.1 Leaf-litter plots

Anurans are abundant in the leaf-litter of tropical forests and can contribute
substantially to amphibian diversity in these habitats. Taxa frequently encoun-
tered in the leaf litter include terrestrial genera such as Eleutherodactylus in the
Neotropics, Arthroleptis in Africa, and Philautus in South Asia. The leaf-litter

amphibian fauna may also include Bufonids, Ranids, Leptodactylids, and other
amphibians that breed in aquatic habitat but spend most of their adult lives on
the forest floor.
Small plots have been frequently used to sample leaf-litter amphibians in
tropical forests (Inger 1980; Vonesh 2001; Rocha et al. 2001). To carry out
leaf-litter surveys, researchers mark off the boundaries of a plot and carefully
sift through the leaf litter for amphibians until they have covered the entire plot.
Searching leaf litter in this manner is time-consuming, so plots are typically
small: usually 1–64 m2. Rocha et al. (2001) argued from a direct comparison of
methods in Atlantic rainforest that small plots (i.e. 2 m2) surveyed very carefully
were likely to be superior to larger plots (i.e. 64 m2) surveyed more superficially.
Leaf-litter quadrats are typically searched during the day for reasons of conveni-
ence, though Rocha et al. (2000) found that both counts and species richness
were actually higher at night.
Because leaf-litter surveys disturb the habitat, researchers usually search each
plot only once. However, litter plots can be highly replicated within any area of
interest. Vonesh (2001), for example, searched a total of 150 25 m2 quadrats in
Kibale National Park, Uganda, 50 each in undisturbed forest, selectively logged
forest, and pine plantation. Conducting a few trial surveys will give a good idea
of how large a plot can be effectively surveyed, how many plots can be searched,
and how abundant amphibians are likely to be in any given sample. Habitat data
can also be an important component of leaf-litter surveys: variables measured
commonly include leaf-litter mass or depth (often correlated with amphibian
abundance: Allmon 1991; Vonesh 2001), herb cover, canopy cover, and soil
moisture, temperature, and pH.
14.3.2 Natural-cover surveys for terrestrial salamanders

Terrestrial salamanders are regularly found underneath natural-cover objects
(e.g. rocks and logs) on the forest floor. Natural-cover surveys for terrestrial sala-
manders can be carried out using plots or transects, though these are typically
much larger than leaf-litter plots, encompassing 100–2000 m2. In natural-cover
surveys, one turns over all accessible rocks and logs and searches for salaman-
ders underneath them. Although natural-cover surveys sample only a fraction
of salamander population, the resulting counts have been shown to be highly
correlated with density estimates from more extensive mark–recapture (Smith
and Petranka 2000). However, salamanders move back and forth between cover
objects, the leaf litter, and underground retreats, so proper timing of these sur-
veys is critical. Generally, the best time for natural-cover surveys is when the
soil underneath cover objects is wetter than the surrounding leaf litter. Cover-
object surveys, like leaf-litter surveys, can disturb the microhabitats of interest,
and Smith and Petranka (2000) recommend re-sampling plots once annually.
Habitat data collected during cover surveys usually include herb cover, canopy
cover, soil moisture, and soil temperature, as well as data on the number, type,
and decay classes of cover objects searched. Salamander counts can be divided
by the number of cover objects on a plot to correct for differences in search
effort (e.g. Marsh and Beckman 2004); however, whether or not this yields
more accurate estimates of relative abundance has not been tested.
The ability to resample salamanders more frequently without degrading cover
objects is one big advantage for artificial cover objects (ACOs; see Chapter 13)
over natural-cover surveys. However, counts from grids of ACOs may be only
weakly correlated with counts from natural-cover surveys (Hyde and Simons
2001), and ACOs may under-represent smaller size classes of salamanders
(Marsh and Goicochea 2003). Thus, natural-cover surveys are recommended
over ACOs for relative abundance and diversity estimation until ACOs have
been more thoroughly validated.
14.3.3 Nocturnal transects

Nocturnal transects have regularly been used to survey forest frogs in tropical
habitats. At night, frogs may crawl up onto vegetation where they can be seen at
or below eye level. Terrestrial-breeding frogs are commonly found with noctur-
nal transects. Additionally, researchers have reported low counts but a remark-
ably high richness of aquatic-breeding species (e.g. Pearman 1997).
Researchers typically set up transects by flagging vegetation in a straight line
of 30–100 m. These transects are walked slowly, searching with a flashlight or
headlamp for frogs. The width of transects is usually no more than 1–2 m to
each side, as beyond this distance too many frogs are likely to be missed (Funk
et al. 2003). Some frogs along transects may be found by localizing their calls, so
in a sense these transects may be considered a form of aural survey (Chapter 16).
However, female frogs of most species do not call, and many additional males
will be seen that are not heard calling. Thus, counts from nocturnal transects
will usually be higher than counts from aural surveys.
Nocturnal transects often involve only moderate disturbance to the vegeta-
tion, and in these cases surveys can be repeated multiple times and used in con-
junction with mark–recapture density estimation (Funk et al. 2003). However,
in habitats with dense understory vegetation, habitat degradation could lead
to reduced counts in subsequent surveys. The optimal conditions for carry-
ing out nocturnal transects can also vary widely among species and habitats.
Temperature and moisture are often positively associated with frog calling activ-
ity (Townsend and Stewart 1994; Duellman 1995), although very heavy rains
can actually reduce activity. Practicing transects in a variety of weather con-
ditions can allow one to determine the best times for carrying out nocturnal
transects. Habitat data measured in conjunction with nocturnal transects often
include herb cover, canopy cover, tree stems, and tree diameter at breast height.
14.3.4 Quadrats for stream amphibians

Amphibians with adult or larval stages that live in streams may be sampled under-
neath rocks in the stream bed or at the edge of the stream. Several researchers
have used quadrat surveys to sample amphibians, particularly larvae and small
adults, in rocky stream beds (Barr and Babbitt 2002). Quadrats of 0.25–4 m2 are
randomly selected within a channel unit (pool, run, or riffle) and then all rocks
are carefully removed to uncover the amphibians present. Quadrat surveys of
a stream bed are in several respects analogous to leaf-litter surveys for tropical
frogs: they are time-consuming so quadrats are typically small, they are moder-
ately destructive so each quadrat is only searched once, and many quadrats can
be surveyed in the course of a study. Habitat data collected during these kinds of
surveys usually includes information on the size and number of rocks turned, as
well as stream characteristics such as water temperature, pH, and flow rate.
14.3.5 Soil quadrats for caecilians

Several recent studies suggest that excavating soil quadrats may be a useful
method for sampling caecilians. Measey et al. (2003) counted caecilians in the
Western Ghats of India with 1 m2 quadrats that were nested within 100 m2 plots.
They used a bladed hoe to dig out the soil to a depth of 0.3 m and sift through
it for caecilians. This technique was surprisingly effective: they detected from
zero to 12 caecilians per quadrat, with a mean density of 0.51–0.63/m2. The one
downside of the technique was that some of the caecilians (≈13%) showed signs
of injury from the sampling procedure, although using a forked digging tool in
place of a bladed hoe could possibly reduce injury rates (G.J. Measey, personal
communication). Measey (2006) further suggested that because most caecil-
ians were found near the surface, sampling could be made more efficient by first
searching a larger plot (e.g. 25 m2) to only a shallow depth and then searching
one or more quadrats within each plot to a greater depth. It bears mention-
ing that not all caecilians are ecologically equivalent and some species may be
more likely than others to be detected in soil quadrats. Gower et al. (2004),
for example, found that timed digs in Tanzania resulted in high counts of one
caecilian species, whereas a second species was detected almost exclusively on
the surface during rainy nights.
14.4 Modifications
The previous examples highlight some of the basic uses of plot and transect
surveys. However, several modifications to these approaches have also been pro-
posed. We cover two in detail: distance sampling (a modification of transect
surveys) and adaptive cluster sampling (a modification of quadrat surveys).
14.4.1 Distance sampling

In bird surveys, distance sampling is widely used to estimate both density and
detection rates (Gregory et al. 2004). The general idea of distance sampling is to
make counts along a transect, but also keep track of the distance between the cen-
ter of the transect and each animal found. Any decrease in counts with distance
from the center of the transect is ascribed to changes in detectability. By mod-
eling the way that detectability decreases with distance, one can estimate dens-
ity from transect counts (for example with the program DISTANCE; Thomas
et al. 2006). Unfortunately, distance sampling has been judged only moderately
effective for amphibians (Fogarty and Vilella 2001; Funk et al. 2003), except
when estimating burrow abundance of the fossorial salamander Phaeognathus
hubrichti (Dodd 1990). This may be in part because many amphibians typically
go undetected even at the center of the transect, leading to imprecise density
estimates. Still, distance sampling does allow one to test whether detectability
differs among habitats when mark–recapture is not feasible.
14.4.2 Adaptive cluster sampling

Amphibians are often highly aggregated in space with many individuals in some
areas but few to none in other areas. This can be frustrating because it means
that many randomly located sites may contain no amphibians. Adaptive cluster
sampling can help focus sampling efforts in areas where animals are actually
being found, but do so in a repeatable way (Thompson and Seber 1996). In
adaptive cluster sampling, one starts by searching a randomly chosen quadrat
within the area of interest. If this plot satisfies some pre-determined criterion
(e.g. more than one amphibian found) then one begins to search adjacent plots.
This process continues until a network of plots has been searched and is sur-
rounded by so-called edge plots that fail to meet the criteria. Noon et al. (2006)
used adaptive cluster sampling to search leaf-litter plots in the Western Ghats
but found that the technique was actually less efficient than simple random sam-
pling for detecting individual species. They postulated that the technique might
have worked better if amphibians had been more abundant and more highly
aggregated; however, more research is needed before adaptive cluster sampling
can be recommended as a first-line approach to sampling.
14.5 Design issues: choice of sampling unit

There are two basic issues in the design of area-based studies: the size and shape
of the sampling units and the number of these units to sample. In terms of size,
small plots or quadrats are generally effective in situations where amphibians are
small and fairly abundant and sampling a small area will tend to lead to nonzero
counts. They are also useful when habitat heterogeneity is high because they can
be highly replicated across the study area. In contrast, larger plots will generally
be more effective when amphibians are less abundant, but are visible enough
that larger areas can be efficiently searched. Plot surveys have been commonly
used for searching low vegetation for frogs and searching coarse woody debris
for salamanders (Woolbright 1991; Bailey et al. 2004). In both of these cases,
small, randomly chosen quadrats might well contain no amphibians, and some
would not even contain appropriate vegetation or woody debris for sampling.
With respect to the shape of the sampling units, linear transects sometimes
offer advantages over rectangular plots. For one, transects are narrow and thus
allow the researcher to do all sampling in a single pass. Because plots are fairly
wide, it can be difficult for a single investigator to keep track of which areas have
already been searched, particularly during nocturnal surveys. Second, transects
cut across a good length of habitat, which may help to average out patchiness in
the environment. For example, when we conducted transect searches to deter-
mine the effects of forest roads on salamanders, we found a highly significant
relationship between distance from roads and salamander counts (Marsh and
Beckman 2004). However, when we later surveyed 8 m 8 m plots in similar
habitats we found no significant relationship. In these plot surveys, the patterns
in counts were almost identical, but the standard deviation in salamander counts
was more than twice as high because of patchiness in salamander distributions.
A further use of transects is when one wants to test the effects of an environ-
mental gradient such as elevation or distance from habitat disturbance. Transects
can be oriented parallel to the gradient to allow for a continuous analysis of the
effects of the gradient on amphibian abundance (e.g. Lehtinen et al. 2003). One
can also set up transects perpendicular to the gradient of interest (e.g. Marsh
and Pearman 1997); however, this approach has the disadvantage of restricting
analysis to a few specific points along the gradient. Sampling perpendicular to
the gradient may be reasonable when there are particular points of interest in
the gradient (Jaeger and Inger 1994), but should probably be avoided when the
gradient and its effects on amphibians are poorly understood.
Although transects may present some advantages over plots, there are also some
situations where plots are likely to be superior. First, when the area to be sampled
is small and round or rectangular, plots are far more practical than very short
transects. Also, plots may be useful when many observers are available for sam-
pling. Observers can cover a rectangular plot in parallel, making these more con-
venient to sample than several distinct transects. Finally, plots should generally be
selected when one is sampling amphibians in conjunction with mark–recapture
analysis. Even if amphibians move only a few meters between captures, they will
tend to wander off of transects and small quadrats, leading to low detection rates
and difficulty with estimating population parameters (Schaub et al. 2004).
14.5.1 Design issues: how many replicates?

After deciding between plots and transects the next question is usually “how
many?” Strictly speaking, the number of replicates needed depends on the mag-
nitude of differences in amphibian abundances one is hoping to detect and on
how much background variation exists between survey dates, sites, and obser-
vers. For example, to compare amphibians between two habitats in which the
standard deviation in counts is equal to 25% of the mean, one would need about
nine replicates per habitat to detect a 50% difference in counts between habitats
(at 90% power) and almost 50 replicates to detect a 20% difference.
In theory, larger plots or transects will tend to average out some habitat vari-
ation, so one would tend to need fewer plots than small quadrats. Typical levels of
replication in the literature are five to 30 for large forest plots, 10–50 for transects,
and 20–100 for small quadrats. To estimate how many replicates are needed,
one can do a few practice plots to calculate variance. From these estimates, the
R library pwr (R Development Core Team 2005) has a number of modules for
doing power analysis and the downloadable program Monitor.exe (Gibbs et al.
1998) can perform power calculations for detecting changes over time.
In some cases, time may be limiting and the more realistic answer to the ques-
tion of “how many plots?” is simply “as many as you can.” Thus, an alternative
approach is to do a few practice sites and estimate the time needed for one replicate.
Then consider how much total time can be committed to the project. Dividing the
latter by the former yields an estimate of how many replicates can be surveyed.
14.5.2 Reducing variation among replicates

More variation among replicates means surveying more sites, so anything that
reduces variation will usually be helpful. Standardization of sampling methods
is one critical aspect of reducing variation. For area-based surveys, standardiza-
tion can include things like always searching plots for the same amount of time,
always searching the same microhabitats within a plot, and always using the
same surveyors for each plot.
These suggestions are valid for all animals, but for amphibians there are
additional considerations. Because amphibian activity is highly dependent on
weather conditions, amphibian counts can vary by orders of magnitude from
day to day or even from hour to hour. There are two main approaches to deal-
ing with this kind of variation. The first is to restrict surveys to periods of time
when amphibians are most likely to be observed. This requires flexibility with
scheduling surveys but can dramatically reduce the amount of survey effort
needed. It also requires some initial surveys to get a sense of amphibian activity
patterns. The second approach is to carry out surveys in a variety of condi-
tions, but to collect data on weather-related variables that may be associated
with activity levels (e.g. rainfall within the past week, rainfall at the time of
the survey, temperature, humidity). These variables can then be included as
covariates in any analysis of amphibian counts, thereby controlling some of the
background variation.
In conjunction with either of these two approaches, surveys can be paired or
grouped in a way that ensures that background variation will not confound a
hypothesis being tested. For example, in a comparison of forest to pasture, each
forest plot could be paired with a pasture plot so that one of each (or two or three
of each) is surveyed on the same date. The pairing of these surveys ensures that
any day-to-day variation affects both habitats equally. Additionally, by using
survey date as a random effect in the statistical analysis, one can remove most of
this variation from the variable of interest.
14.5.3 How many times to survey each replicate?

The number of times each replicate should be surveyed will depend on how
destructive a method is to the microhabitats where amphibians are found. If a
sampling method tends to degrade the habitats of interest, repeat surveys will
result in lower counts and the appearance of population declines where none
would otherwise exist. In these cases (e.g. leaf-litter plots) one can sample a dif-
ferent, randomly chosen set of quadrats in each survey period. This will tend to
increase sampling error (since different plots will be surveyed each time) but will
eliminate any bias associated with habitat degradation from surveys.
However, even when there is little disturbance associated with sampling, one
is still faced with a trade-off between the number of times each site can be
surveyed and the total number of sites that can be covered. If amphibian abun-
dance is highly variable within an area, but there is not much variation from
one survey to the next, more replicate plots that are surveyed less frequently is
likely to be optimal (e.g. Smith and Petranka 2000). The exception to this is
when using repeat surveys with mark–recapture to estimate detection rates (see
Chapter 24).
14.6 An example of study design

Suppose one wanted to compare counts of a single amphibian species in for-
est habitat to counts in pasture. Assuming that only one forest area is available
for sampling and that it is entirely surrounded by pasture, one would have at
least four choices for area-based surveys: large plots, small quadrats, transects,
or some nested design. If a total of 1000 m2 could be surveyed in each of the two
habitats, one could sample five 20 m 10 m plots, 10 50 m 2 m transects, or 40
5 m 5 m quadrats (Figure 14.1a–c). Alternatively, one could use a nested design
(a) Large plots (b) Transects
Pasture
Pasture
Forest
Forest
(c) Small quadrats (d) Quadrats within plots
Pasture Pasture
Forest
Forest
Fig. 14.1 Four possible designs for area-based sampling to compare a single forest
area to surrounding pasture. Equal numbers of samples would be randomly located
within pasture but these are not shown. (a) Five plots of 200 m2. (b) Ten transects
of 50 m 2 m. (c) Forty 5 m 5 m quadrats. (d) Four 5 m 5 m quadrats nested
within each of ten 200 m2 plots.
(a) Transects
Pasture
Forest
Forest
Pasture
(b) Quadrats
Pasture
Forest
Forest
Pasture
Fig. 14.2 Two possible designs for area-based sampling to compare five forest
patches to surrounding pasture. (a) Two transects of 50 m 2 m within each forest
patch. (b) Eight 5 m 5 m quadrats within each forest patch.
with 10 20 m 10 m plots and four 5 m 5 m quadrats chosen for sampling

from within these (Figure 14.1d).
Any of these designs would permit a comparison of amphibian counts between
the single forest patch and the surrounding pasture. However, because only one
forest patch is sampled, inferences can only be made about that particular site.
Thus, an even better design would use replicated forest patches rather than the
single forest area, which would allow general inferences about amphibian counts
in forest within this region. Assuming that five forest patches could be located,
an area-based survey could employ a single 20 m 10 m plot in each forest
patch, two 50 m 2 m transects, or eight 5 m 5 m quadrats (Figure 14.2)
Each of these could be paired with a parallel set-up in the surrounding pasture.
14.7 Assumptions of area-based surveys

Area-based surveys, as with any other research technique, require some assump-
tions to draw inferences from the data collected. First, plots or transects are
assumed to be representative of the larger area of interest. In theory, a large sam-
ple of randomly chosen plots will closely approximate the larger area. However,
smaller numbers of plots, even if randomly chosen, might be spatially aggre-
gated or might under-represent particular microhabitats. When the larger area
is known to vary, stratifying the area into different habitat types, and then ran-
domly choosing plots within strata may be a better approach (Cochran 1977).
Second, area-based surveys generally yield count data with unknown detec-
tion rates. This is particularly true with more destructive techniques (e.g.
leaf-litter plots, stream-bed quadrats) since one cannot use repeat surveys and
mark–recapture to directly estimate detectability. Issues related to detectability
and the interpretation of count data are covered in Chapters 24 and 25. Here,
we simply point out that most uses of area-based surveys are aimed at comparing
relative abundance or species richness among sites or time periods. Thus, the key
assumption is not that all individuals are detected, but that detection rates are
relatively constant among habitats or surveys. For instance, if vegetation differs
between primary and secondary forest, counts from nocturnal transects in the
two habitats may not be directly comparable since detection rates will likely
depend on vegetation density. Furthermore, detectability is pretty much always
an issue if one wants to compare abundances of different species. Distinct spe-
cies tend to use different microhabitats, so area-based surveys will almost always
detect some species at higher rates than others. Thus, characterizing amphibian
communities usually requires multiple search techniques and some formal esti-
mate of detectability.
Finally, with area-based surveys one generally assumes that amphibians do

not move around much during the course of a survey. If amphibians are dis-
turbed by the survey techniques and consistently escape the sampling area, these
surveys will not be particularly effective.
14.8 Summary and recommendations

There is no single best approach for conducting area-based surveys. Which
approach is optimal will depend on the site, the species, and the hypothesis
being tested. That said, there are some general recommendations to keep in
mind when conducting area-based surveys.
1) Before beginning a study, always practice the survey techniques. Testing
out techniques is essential and usually leads to adjustments and improve-
ments. Ultimately, trial surveys save time, because data will be of higher
quality and there will be no need to repeat any surveys. Even two or three
trial plots or transects can go a long way towards identifying potential
problems and suggesting the necessary changes.
2) When in doubt, include more replicates. Almost nothing is as frustrating
as coming up a few plots or transects short of an answer. When faced with
a trade-off between the size of sampling units and the number of repli-
cates, we believe that a greater number of smaller replicates will usually
be optimal (the exception is when plots are so small that zero counts fre-
quently result). Again, practicing plots of various sizes can help steer one
to an appropriate-sized sampling unit.
3) Carefully consider the research question and how the survey design will
answer that question. Going to the extreme of writing out a sample data set
and attempting to analyze it can help identify errors or unstated assump-
tions in a sampling approach.
4) Know your amphibians. While this chapter has focused on the general
concepts of area-based surveys, the details of a species’ natural history will
ultimately determine whether a particular sampling approach is success-
ful. Understanding the basics of activity patterns and movement behavior
are critical for effective sampling.
14.9 References
Allmon, W. D. (1991). A plot study of forest floor litter frogs, central Amazon, Brazil.
Journal of Tropical Ecology, 7, 503–22.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004). Spatial and temporal variation
in detection probability of Plethodon salamanders using the robust capture-recapture
design. Journal of Wildlife Management, 68, 14–24.
Barr, G. E. and Babbitt, K. J. (2002). Effects of biotic and abiotic factors on the distribu-
tion and abundance of larval two-lined salamanders (Eurycea bislineata) across spatial
scales. Oecologia, 133, 176–85.
Cochran, W. G. (1977). Sampling Techniques. Wiley, New York.
Dodd, Jr, C. K. (1990). Line transect estimation of Red Hills salamander burrow density
using a Fourier series. Copeia, 1990, 555–7.
Duellman, W. E. (1995). Temporal fluctuations in abundances of anuran amphibians in
a seasonal Amazonian rainforest. Journal of Herpetology, 29, 13–21.
Fogarty, J. H. and Vilella, F. J. (2001). Evaluating methodologies to survey
Eleutherodactylus frogs in montane forests of Puerto Rico. Wildlife Society Bulletin,
29, 948–55.
Funk, W. C., Almeida-Reinoso, D., Nogales-Sornosa, F., and Bustamante, M. R. (2003).
Monitoring population trends of Eleutherodactylus frogs. Journal of Herpetology, 37,
245–56.
Gibbs, J. P., Droege, S., and Eagle, P. (1998). Monitoring populations of plants and animals.
Bioscience, 48, 935–40.
Gower, D. J., Loader, S. P., Moncrieff, C. B., and Wilkinson, M. (2004). Niche separ-
ation and comparative abundance of Boulengerula boulengeri and Scolecomorphus vittatus
(Amphibia: Gymnophiona) in an East Usambara forest, Tanzania. African Journal of
Gregory, R. D., Gibbons, D. W., and Donald, P. F. (2004). Bird census and survey
techniques. In W. J. Sutherland, I. Newton, and R. E. Green (eds), Bird Ecology and
Conservation: a Handbook of Techniques, pp. 17–55. Oxford University Press, Oxford.
Hairston, Sr, N. G. (1987). Community Ecology and Salamander Guilds. Cambridge
University Press, New York.
Hyde E. J. and Simons, T. R. (2001). Sampling plethodontid salamanders: sources of
variability. Journal of Wildlife Management, 65, 624–32.
Inger, R. F. (1980). Densities of floor–dwelling frogs and lizards in lowland forests of SE
Asia and Central America. American Naturalist, 115, 761–70.
Jaeger, R. and Inger, R. F. (1994). Standard techniques for inventory and monitoring:
quadrat sampling. In W. R. Heyer, M. A. Donnely, R. W. McDiarmid, L. C. Hayek,
and M. S. Foster (ed.), Measuring and Monitoring Biological Diversity. Standard Methods
for Amphibians, pp. 97–102. Smithsonian Institution Press, Washington DC.
Lehtinen, R. M., Ramanamanjato, J. B., and Raveloarison, J. G. (2003). Edge effects and
extinction proneness in a herpetofauna from Madagascar. Biodiversity and Conservation,
12, 1357–70.
Marsh, D. M. and Pearman, P. B. (1997). Effects of habitat fragmentation on the abun-
dance of two species of Leptodactylid frogs in an Andean montane forest. Conservation
Biology, 11, 1323–8.
Marsh, D. M. and Goicochea, M. A. (2003). Monitoring terrestrial salamanders: biases
due to frequent sampling and choice of cover objects. Journal of Herpetology, 37,
460–6.
Marsh, D. M. and Beckman, N. G. (2004). Effects of forest roads on the abundance and
activity of terrestrial salamanders in the Southern Appalachians. Ecological Applications,
14, 1882–91.
Measey, G. J. (2006). Surveying biodiversity of soil herpetofauna: towards a standard
quantitative methodology. European Journal of Soil Biology, 42, S103–10.
Measey, G. J., Gower, D. J., Oommen, O. V., and Wilkinson, M. (2003). Quantitative
surveying of endogeic limbless vertebrates – a case study of Gegeneophis ramaswamii
(Amphibia: Gymnophiona: Caeciliidae) in southern India. Applied Soil Ecology, 23,
43–53.
Noon, B. R., Ishwar, N. M., Vasudevan, K. (2006). Efficiency of adaptive cluster and
random sampling in detecting terrestrial herpetofauna in a tropical rainforest. Wildlife
Society Bulletin, 34, 59–68.
Pearman, P. B. (1997). Correlates of amphibian diversity in an altered landscape of
Amazonian Ecuador. Conservation Biology, 11, 1211–25.
R Development Core Team (2005). R: a Language and Environment for Statistical
Computing, Reference Index Version 2.2.1. www.R-project.org. R Foundation for
Statistical Computing, Vienna.
Rocha, C. F. D., Van Sluys, M., Alves, M. A.S., Bergallo, H. G., and Vrcibradic, D. (2000).
Activity of leaf-litter frogs: when should frogs be sampled? Journal of Herpetology, 34,
285–7.
Rocha, C. F. D., Van Sluys, M., Alves, M. A. S., Bergallo, H. G., and Vrcibradic, D.
(2001). Estimates of forest floor litter frog communities: a comparison of two methods.
Austral Ecology, 26, 14–21.
Schaub, M., Gimenez, O., Schmidt, B. R., and Pradel, R. (2004). Estimating survival
and temporary emigration in the multistate capture–recapture framework. Ecology,
85, 2107–13.
Smith, C. K. and Petranka, J. W. (2000). Monitoring terrestrial salamanders: repeatabil-
ity and validity of area-constrained cover object searches. Journal of Herpetology, 34,
547–57.
Thomas, L., Laake, J. L., Strindberg, S., Marques, F. F. C., Buckland, S. T., Borchers,
D. L., Anderson, D. R., Burnham, K. P., Hedley, S. L., Pollard, J. H. et al. (2006).
Distance 5.0. Release 2. www.ruwpa.st-and.ac.uk/distance/. Research Unit for Wildlife
Population Assessment, University of St Andrews, St Andrews.
Thompson, S. K. and Seber, G. A. F. (1996). Adaptive Sampling. John Wiley and Sons,
New York.
Toft, C. A. (1980). Feeding ecology of thirteen syntopic species of anurans in a seasonal
tropical environment. Oecologia, 45, 131–41.
Townsend, D. S. and Stewart, M. M. (1994). Reproductive ecology of the Puerto Rican
frog Eleutherodactylus coqui. Journal of Herpetology, 28, 34–40.
Vonesh, J. R. (2001). Patterns of richness and abundance in a tropical African leaf-litter
herpetofauna. Biotropica, 33, 502–10.
Woolbright, L. L. (1991). The impact of Hurricane Hugo on forest frogs in Puerto Rico.
Biotropica, 23, 462–7.
15
Rapid assessments of amphibian diversity
James R. Vonesh, Joseph C. Mitchell, Kim Howell, and
Andrew J. Crawford
15.1 Background: rapid assessment of

amphibian diversity
More than 6400 amphibian species are known worldwide, with more than 50
new species being described in just the first half of 2008 (AmphibiaWeb 2008).
Many of these species are threatened or declining and more than 150 may have
recently become extinct (IUCN 2006). Such rates of species loss are far greater
than the historic background extinction rate for amphibians (e.g. McCallum
2007; Roelants et al. 2007). Amphibians play diverse roles in natural ecosys-
tems, and their decline may cause other species to become threatened or may
undermine aspects of ecosystem function (Matthews et al. 2002; Whiles et al.
2006). Anthropogenic habitat loss and degradation, disease, introduced species,
and pollution or combinations of these factors are at the root of most declines.
As awareness of declines has increased, conservation groups, governments, and
land managers have become more interested in protecting amphibian diversity.
However, the lack of accurate data on amphibian distributions, particularly for
tropical regions where diversity and declines are concentrated (IUCN 2006),
is often a roadblock to effective conservation and management. Ideally, lack of
information on the amphibian fauna for a particular area would result in a thor-
ough inventory. Unfortunately, this is usually not realistic. Exhaustive inven-
tories are costly and may take decades to compile (e.g. Timm 1994). Given the
urgency of the current global biodiversity crisis, finite resources, and dynamic
socio-economic environments with respect to conservation, an approach for
rapidly gathering preliminary data on biodiversity is sometimes required.
Rapid biodiversity-assessment methods were developed in response to this
need. A Rapid Assessment (RA) is an accelerated, targeted, and flexible bio-
diversity survey, often focusing on species associated with particular vegetation
types or topographical features (Sayre et al. 2000). RAs consist of planning,

field sampling, and analytical and writing stages, and are implemented by teams
of biologists and resource managers with expertise in the region and taxonomic
group(s) of interest. Most RAs aim to be completed, from planning to report,
in less than a year. Although a variety of RA strategies have been proposed,
two of the better documented are the Rapid Ecological Assessment method-
ology developed by the Nature Conservancy (Sayre et al. 2000) and the Rapid
Assessment Program methodology developed by Conservation International
(Roberts 1991; www.conservation.org). RA approaches share a number of fea-
tures in common (modified from Sayre et al. 2000).
• Speed. Reducing project duration reduces costs and delivers results to
decision-makers quickly.
• Careful initial planning and training. Effective planning saves time and
money and increases the consistency of the information gathered.
• Use of mapping technologies. The use of Geographic Information Systems
(GIS), remote sensing, and global positioning systems (GPS) can greatly
facilitate planning, sampling design, and geo-referencing of data.
• Careful scientific documentation. Selecting classification, sampling, and
surveying methods that are most effective for the habitat, season, and taxa
being sampled can maximize assessment effectiveness.
• Capacity-building and partnerships. Developing collaborative relationships
among conservation, government, and academic partners in host countries
and internationally helps ensure that information generated will have a
local impact.
An RA is not an exhaustive faunal inventory, monitoring program, environ-
mental impact assessment, or management plan, nor does it seek to elucidate
species interactions or ecological processes, although an RA could serve as an
important prelude to any of these activities. Instead, an RA is a conservation
planning tool that can provide an efficient initial characterization of diversity
in areas for which relatively little is known (Sayre et al. 2000). Here we describe
in brief the steps involved in RA and discuss sampling methodologies likely to
be effective for assessing amphibian diversity. Due to the integrative nature of
RA, we often refer the reader to other chapters in this volume that cover material
important to the topic.
15.1.1 When is an RA needed?

Determining whether an RA approach is called for depends on how much is
already known about the area of interest and the urgency to obtain additional
15 Rapid assessments of amphibian diversity | 265
information. The best candidate sites are those that are only marginally sur-
veyed and are highly threatened (Sayre and Roca 2000). It is in these cases that
data resulting from an RA, though typically limited in scope, can have import-
ant implications for decision-making.
For example, the Atewa Range Forest Reserve of Ghana contains some of the
last remaining upland evergreen forest in the country and is home to endemic
plant and insect species (McCullough et al. 2007). In 2006, Conservation
International initiated a Rapid Assessment Program of amphibians (and other
taxa) within the Atewa Range Forest Reserve in response to proposed mineral
exploration in the reserve. Sampling sites were selected in areas suspected to sup-
port high biodiversity that also had large mineral deposits (McCullough et al.
2007). Field surveys were conducted over just 18 days at the beginning of the
rainy season. Nearly one-third of the 32 amphibian species observed were listed
as threatened on the IUCN Red List. One species was considered Critically
Endangered (Kouamé et al. 2007). Based in part on the amphibian assessment
results, the authors recommended that Atewa Range Forest Reserve be elevated
to national park status and that mineral exploitation be prohibited (Kouamé
et al. 2007; McCullough et al. 2007).
15.2 Planning an RA
When an RA is indicated, the next steps involve planning and preparation.
Because RAs may involve government and non-government participants from
multiple regions, are often conducted in remote, logistically challenging field
localities, and investigate unknown or poorly described faunas, substantial
advanced planning is necessary for success.
15.2.1 Developing objectives

Establishing clear and realistic objectives is the most important step of the plan-
ning process, as the objectives become the yardstick for subsequent activities
and resource allocation (Hayek 1994; Sayre and Roca 2000). Often amphibian
assessments will be conducted as part of a larger project involving specialized
teams looking at a variety of taxonomic groups, and objectives will need to be
coordinated (e.g. Kouamé et al. 2007). The objectives should define the tem-
poral, spatial, and taxonomic scope of the project, and in doing so determine the
sampling methods and intensity required. For example, the objective might be
to provide a preliminary species list of stream-associated taxa from a particular
watershed based on snapshot sampling of several sites during the rainy season.
Or perhaps the goal is to provide an initial inventory of all amphibian species
within an area of conservation interest based on sampling effort spread over

multiple seasons. Ideally, one’s objectives are focused, realistic, and quantifiable
(Chapter 2). Initial effort in reviewing and refining objectives is always worth
the investment.
15.2.2 Costs and funding

Consideration of the costs and time necessary to successfully complete an RA
is important and the decision on whether or not to proceed with an RA is often
affected by financial and time investments. Costs will depend upon the scope
of the objectives and logistics of field implementation, and can include sal-
ary expenditures, field equipment and supplies, landscape imagery purchases,
international and in-country travel expenses, collecting and research permits,
training expenses, contracts, laboratory equipment, museum supplies, software,
digital storage media, and publication costs (Sayre and Roca 2000). Advanced
consideration of these costs allows careful evaluation of trade-offs in the use of
finite resources. For example, the use of satellite images, aerial photographs, or
aerial videography to identify and map different habitat types and topographical
features can greatly facilitate selection of sampling localities and is a key feature
of some RA approaches (e.g. Rapid Ecological Assessment; Sayre et al. 2000).
However, acquisition and processing of remote imagery can be very expensive,
putting it beyond the scope of projects with more limited resources. Similarly,
additional field sampling across seasons may increase the number of species
observed, but also increases costs associated with returning to the site or main-
taining a team in the field. Once costs have been estimated and a short project
proposal developed, potential financial supporters can be approached. Options
for securing financial support include development banks, governments, devel-
opment agencies, conservation organizations, foundations, corporations, mili-
tary landowners, and private individuals (Sayre et al. 2000).
15.2.3 Team selection and training

Depending upon the scope of the project, RA teams range from a few to many
people, and may include senior scientists, students, technicians, field assist-
ants, guides, rangers, data managers, cartographers, and administrators from
multiple collaborating institutions (Sayre and Roca 2000; McCullough et al.
2007). For larger teams, an initial planning workshop is advisable, with the goal
of developing a work plan that clearly identifies the roles and responsibilities of
participating organizations and individuals. An example outline for a planning
workshop and an example working plan for Rapid Ecological Assessment is
given in Sayre and Roca (2000). International collaborations are recommended,
whenever possible. Foreigners planning an RA should collaborate with local

scientists to increase the chances of a successful project, and to make lasting
contributions to the host country. Nationals planning an RA should consider
involving foreign experts in important methodologies or taxonomic groups,
especially from neighboring countries with related faunas.
A training workshop specifically for field personnel can also be an import-
ant tool, by providing training related to animal identification, field sampling
techniques and data collection, use of GPS for geolocation, mapping and navi-
gation, and safety protocols. All team members in charge of data collection
should have at least a basic appreciation of the sampling design and statistical
analyses planned (Magnusson and Mourão 2004; Chapter 18). In addition,
amphibian pathogens and parasites can be carried between habitats on hands,
footwear, or field equipment, and could be spread to new localities. Thus,
it is important that initial training for those involved in RA include how to
minimize the spread of disease and parasites between study sites (Declining
Amphibians Population Task Force (DAPTF) code of practice, Chapter 26;
Aguirre and Lampo 2006). Typically, this training occurs at the study area at
the commencement of fieldwork. An overview of basic qualifications for field
biologists, training tips, project supervision, and quality assurance are given in
Fellers and Freel (1995).
15.2.4 Permits
Conducting an RA may require permits from local, regional, national, and inter-
national governing agencies. Most national governments, and many state and
local governments, require permits to study amphibians within their boundaries
and public lands. Many protected areas (e.g. national parks) require additional
permission. National agencies and international treaties (e.g. Convention on
International Trade in Endangered Species of Wild Fauna and Flora (CITES),
U.S. Fish and Wildlife Service, Convention on Biological Diversity) also regu-
late export and import of specimens and tissue samples. There is considerable
variation among regions in permit fees and the times required to process and
receive permits, ranging from weeks to well over a year. Thus, the permitting
process should be initiated well in advance of the planned fieldwork.
15.2.5 Data management

Although short in duration, RA projects can produce large amounts of data of
different types, include quantitative data that can be represented as numbers,
qualitative data that are not easily represented in numerical form, spatial data that
are linked to geographic coordinates, metadata documentation that accompanies
other data sets, and images or video or audio clips. Data management is the process
that helps ensure that diverse data sets can be efficiently collected, integrated and
analyzed, and archived so that they can be easily retrieved in the future (Gotelli
and Ellison 2004). Planning for good data management begins early in project
development by making a table of the types of data to be produced in the field;
for example habitat variables (Chapter 17), conditions, counts, specimens, and
other media such as photos, audio (Chapter 16), or video. This information can
be used to design protocols (e.g. field forms, labels, code numbers) for collection
and management of disparate types of data. Once data are being generated it is
important to review data in the field, and verify that team members are using iden-
tical protocols. Data should be transcribed regularly and transferred to a database
as soon as possible. If a portable computer is brought to the field, raw data should
still be maintained on paper in case of damage to the computer. Widely used
software for data management include Microsoft Access and the non-proprietary
OpenOffice BASE. Protocols need to be established for reviewing, cleaning, and
backing up data on a regular basis. Museum specimen data should be stored in a
manner compliant with the Distributed Generic Information Retrieval (DiGIR)
protocols, such as the database management software, Specify (http://www.
specifysoftware.org).
15.2.6 Developing the sampling plan

The sampling plan is a document that identifies areas to be sampled during field
work, the time that each area is to be sampled, the sampling techniques to be
used and designate the individuals or teams responsible for conducting the field
work. The sampling plan lays out the strategy for obtaining data that are repre-
sentative of the target area of interest (Sayre 2000).
15.2.6.1 Selecting sampling sites

RA, at its most basic level, involves the collection and characterization of taxo-
nomic and spatial information about biodiversity (Sayre et al. 2000). Although
identifying sampling localities is one of the first steps in developing a sampling
plan for any RA, protocols differ in their emphasis of initial landscape charac-
terization. In some cases, sampling sites may be identified based upon perceived
high biodiversity potential (e.g. perhaps as indicated by previous sampling of other
taxa), specific conservation threats, or access logistics (e.g. Kouamé et al. 2007).
In other cases, sites may be selected to capture clines in important abiotic features
(e.g. elevation, rainfall, vegetation) that may determine the boundaries of particu-
lar amphibian assemblages (e.g. Menegon and Salvidio 2005). Some projects use
remote sensing imagery to visually identify specific habitats of interest prior to
sampling. Finally, GIS (Chapter 19) coupled with ecological niche modeling pro-
vide a new tool for remotely identifying potential habitat for species of concern or
biodiversity hotspots in general (e.g. Raxworthy et al. 2003; Pawar et al. 2007).
Development and accessibility of new mapping technologies over the past two
decades have had important impacts on the way RAs are conducted. Spatial tech-
nologies, including GIS, remote sensing, and GPS are commonly used in defin-
ing project scope and establishing sampling localities. The Rapid Ecological
Assessment approach developed by Conservation International places strong
emphasis on using satellite and aerial imagery to classify landscapes of interest
into vegetation or land-use cover categories (Sayre et al. 2000). Sampling sites are
spread across these vegetation categories to establish the framework within which
field sampling is conducted and to facilitate linkages between fine-scale sampling
by field teams with landscape-level assessment of biodiversity (Nagendra and
Gadgil 1999; Sayre et al. 2000). Sayre et al. (2000) provide a detailed example of
how natural-colour and colour-infrared satellite and aerial fly-over imagery were
used to develop an initial landscape characterization of Parque Nacional del Este,
Dominican Republic, which was subsequently used to identify specific sampling
sites and generate detailed site maps. While the utility of remote sensing imagery
when planning an RA is readily apparent, costs can be considerable. Commercial
satellite imagery may cost US$3000–5000 per scene and an aerial photo acqui-
sition mission may cost between $20 000 and 120 000 (Sayre et al. 2000). For
projects with limited resources, Google Earth™ maps integrate data from satel-
lite imagery, aerial photography, and a GIS 3D globe to provide resolution of at
least 15 m for most terrestrial areas.
We draw a distinction between the ways sampling sites are determined in most
RA studies compared to traditional ecological inventory methods. Traditional
ecological inventories emphasized objective field sampling based on rand-
omized selection of replicated samples and substantial effort may be given to
issues such as defining sampling coverage and determining detection probabil-
ities (e.g. Buckland et al. 1993; 2001, 2004; Heyer et al. 1994; Williams et al.
2002). This emphasis on randomization, replication, and estimating detection
functions greatly increases the kinds of inferences that can be made. However,
for many RAs, sampling may be opportunistic, or determined by issues such as
logistics, access, efficiency, and a priori perceptions of areas likely to be high-
est in diversity or under the greatest threat. Although replicate units may be
sampled in some cases, the resulting data may not be rigorously applied to all
questions of interest to the ecologist or biogeographer (Sayre 2000). As a result,
the results from many RAs are best viewed as qualitative or semi-quantitative
(e.g. Kouamé et al. 2007).
15.2.6.2 Selecting sampling time

Scheduling sampling times will depend on latitude, elevation, seasonal patterns
of rainfall, and knowledge of species breeding phenology. For example, in Kibale
National Park, Uganda, pond-breeding Hyperolius reed frogs are most active
and breed during the two rainy seasons (Vonesh 2000). In contrast, Leptopelis
treefrogs breed during the dry seasons (J.R. Vonesh, unpublished results).
Successfully sampling both of these taxa might require fieldwork during wet
and dry seasons. However, when resources limit sampling to a single season,
more amphibians are likely to be sampled during wet months.
15.2.6.3 Selecting sampling techniques

A variety of techniques can be used to assess amphibian diversity, including
drift fences, pitfall traps, and cover boards (Chapter 13; Menegon and Salvidio
2005), quadrat sampling (aquatic: Chapter 4; terrestrial: Chapter 14), call sur-
veys (Chapter 16), road and trail censuses (Pearman et al. 1995; Rödel and Ernst
2004), aquatic sweep-net sampling (Chapter 4), bottle or minnow traps, and
visual surveys (e.g. Crump and Scott 1994). References on amphibian sampling
include Campbell and Christman (1982), Corn and Bury (1990), Heyer et al.
(1994), Olson et al. (1997), Lips et al. (2001), Howell (2002), Rueda-Almonacid
et al. (2006), and Williams et al. (2002).
Selecting the sampling methodologies for an RA requires careful thought.
Above all an RA must be rapid, yet our objective of a complete species list would
suggest that we employ multiple techniques because techniques differ in the set
of species they sample successfully. Often the strengths of different techniques
are readily apparent, for example litter quadrat sampling will be more effective at
sampling direct-developing species than visual encounter surveys (VESs) around
aquatic habitats (Vonesh 2000, 2001a, 2001b; J.R. Vonesh, unpublished results).
In other cases the trade-offs among techniques are more subtle and may change
seasonally (K. Howell, unpublished results). Techniques also vary in terms of
cost, effort, and other logistical requirements. Drift-fence arrays can be effect-
ive at sampling amphibians in a variety of situations (Chapter 13; K. Howell,
unpublished results) but they are relatively expensive and labor-intensive to set up
and maintain. Therefore, each potential field method should be evaluated for its
relative capture success, speed, bias, and the extent to which it can facilitate the
objectives of the RA (e.g. see Chapter 6, Table 4 in Heyer et al. 1994).
15.2.6.3.1 Visual encounter surveys (VESs)

VESs involves field personnel searching the focal habitat systematically for
a known period of time. The clock is stopped when not actively searching
(e.g. animals are being processed). The number of animals observed can then
be expressed in terms of animals observed per area (or distance) per person
searching, or per unit time per person. VES is an effective technique for build-
ing species lists rapidly (Crump and Scott, 1994; Rodda et al. 2007), requires
little equipment (e.g. minimally a headlamp and field notebook), and can
be implemented in a variety of habitat types. For these reasons it is perhaps the
most widely employed sampling method in RA of amphibian diversity, so we
provide a brief overview of the method here. For additional information see
Crump and Scott (1994), Corn and Bury (1990), Campbell and Christman
(1982), Rödel and Ernst (2004), and Rueda-Almonacid et al. (2006).
There is considerable variation among past studies in how VES is conducted.
In some instances (e.g. Mitchell 2006; Kouamé et al. 2007) experienced her-
petologists selectively search areas and microhabitats determined most likely to
yield amphibians. This approach has the benefit of potentially yielding more
animals and species per effort than randomized sampling approaches. However,
it is most subject to variation in the skill levels of the field personnel and is lim-
ited in its ability to generalize about relative abundances and habitat associations
because (micro-)habitats are searched in a biased manner. Alternatively, an area
may be searched via VES using a randomized-walk design (Crump and Scott
1994). In this case, prior to going to the field site the researcher generates a ran-
dom sequence of compass headings as well as a random distance to be searched
along each heading. A starting heading and distance is selected at random and
field personnel simply work through the list of headings and distances, search-
ing for animals, for a specified time. Since the path to be sampled is determined
randomly, this approach has the advantage that replicated random walks from
different areas could be compared statistically. However, it is important to appre-
ciate that differences in the observed number of animals among sites may be as
much due to differences in detection among sites as much as differences in true
abundance. If the sampling design requires greater effort per area, VES can be
conducted using a quadrat design (Crump and Scott 1994). In this case, quad-
rats of a known area are established and each is then sampled systematically by
searching parallel transects across the plot (e.g. Hairston 1980; Aichinger 1987;
Donnelly 1989). Crump and Scott (1994) recommend plot sizes of 10 m 10 m
or 25 m 25 m, depending upon amphibian densities. To statistically compare
areas of interest (e.g. different habitats) plots must be replicated and randomly
located within areas to be compared (e.g. using GPS coordinates or a trail grid
system). Finally, transect designs are often used in coordination with VES for
sampling across habitat gradients that may affect amphibian diversity and abun-
dance (e.g. moisture, elevation). Although these considerations may determine
transect direction, distances to be sampled and starting points are best deter-
mined a priori and multiple randomly located replicate transects will be needed
if areas are to be compared statistically.
It is also important to determine and carefully define the intensity of field
sampling in advance. Crump and Scott (1994) suggest three levels of sampling
intensity for sampling amphibians. The least intensive surveys are counts of
animals that are active on the surface (Hairston 1980; Vonesh 2000). This
approach is minimally invasive and thus is suitable for search areas with sensi-
tive faunas and also requires the least amount of time, increasing the amount
of area that can be covered. Intermediate-level searches count exposed ani-
mals but also turn over surface cover objects (note that cover objects must
be returned to their original position to minimize disturbance). This level
of VES can yield species often overlooked by low-intensity VES because of
their secretive and semi-fossorial life histories (e.g. many salamanders and
caecilians). When the most complete inventory possible is desired, all pos-
sible microhabitats are searched; surface objects are turned, decaying logs are
torn apart, the leaf litter is systematically raked, epiphytes and tree holes are
searched (Crump and Scott 1994), and aquatic habitats are searched for adults
and larvae (Chapter 4). Such intensity requires considerable labor per unit
area and causes the greatest disturbance to the habitat but may be more effect-
ive at sampling some rare species.
15.2.6.3.2 Assumptions and limitations of VESs

The analysis of VES data is based on the following assumptions: (1) all individu-
als are equally detectable; (2) individuals are recorded once during a survey; and
(3) there are no observer-related biases in sampling (Crump and Scott 1994). In
most instances, one or more of these assumptions will be violated. Amphibians
of different species, sizes, or life stages are generally not equally conspicuous.
Even the probability of detecting the same individual may vary due to temporal
(e.g. diel, seasonal) variation in behavior (Bailey et al. 2004; Schmidt 2005).
Similarly, differences in observer experience can often result in differences in
detection. Thus, the assumption of equal detection probability across time, space,
and individuals is unlikely to hold for most amphibian taxa or studies (Mazerolle
et al. 2007). This has important implications for simple count-based methods
like VES. Consider that the expected number of individuals or taxa counted
(C) in a survey area is E(C) PN, where P is the probably of detection and N is
the true number in the population (Williams et al. 2002). Without an estimate
of P, for any count of animals sampled, there is an infinite number of possible
true population sizes. Mazerolle et al. (2007) and Schmidt (2004, 2005) provide
lucid reviews of these issues and specifically focus on their application to her-
petological studies. Given these limitations, VES is best viewed as a qualitative
or semi-quantitative approach and should be used only when qualitative results
are acceptable (e.g. threatened species were observed; Kouamé et al. 2007). VES
may be a suitable sampling approach given the objectives of many RA projects.
However, when resources and objectives allow, other more powerful study designs
(e.g. distance sampling methods; Thomas et al. 2002; Funk et al. 2003; Fogarty
and Viletta 2001; Viletta and Fogarty 2005) should be considered.
15.2.6.4 Determining what data to collect

Generating lists of observed species is usually the principle type of data-collection
method used during an RA. In some cases, animals encountered in the field
may be identified in the field. In other cases, it may not be possible to identify all
species, requiring that specimens be held temporarily for identification (e.g. to
consult a key) before being released at the site of collection, or it may be necessary
to preserve voucher specimens that can be identified by taxonomic specialists
(Jacobs and Heyer 1994; McDiarmid 1994). Voucher specimens are what ultim-
ately define the identity and distribution of any amphibian species. Depending
on collecting permits obtained, a series of one or more voucher specimens per
sex, life stage, and species should be collected during any give RA, especially
for remote geographic locations, and deposited in a recognized natural history
museum. Voucher specimens allow independent researchers to confirm speci-
men identification at a later date, a very important consideration given the highly
unstable state of amphibian taxonomy. The systematic value of a voucher speci-
men increases as more ancillary data are included.
Depending upon the project’s goals, it may be important to collect add-
itional data for animals observed or collected. GPS can be used to provide a
geo-reference for each sample. Basic size measurements (snout–vent length,
mass), sex, reproductive status based on secondary sexual characteristics, notes or
photographs indicating coloration, and basic activity (e.g. calling, resting, mov-
ing, etc.) may be recorded. Tissue samples for DNA analyses may be obtained
from a toe clip, liver sample, or a buccal swab preserved in 95% ethanol or other
buffer (Seutin et al. 1991; Gonser and Collura 1996; Goldberg et al. 2003).
The benefits of integrating RA data with the DNA barcode of life campaign are
potentially enormous but so far have been under-utilized (Vences et al. 2005;
Fouquet et al. 2007; Ficetola et al. 2008; Smith et al. 2008). Additional types of
sample, such as skin swabs, may be required for disease detection (Chapter 26).
It is also important to record environmental characteristics, such as weather
conditions, lunar cycle, and vegetation or habitat characteristics (Chapter 17).
However, collecting additional data on individuals is time-consuming and thus

additional data on individuals may involve trading-off the amount of area that
can be searched. Field forms are useful to remind the field team what data are
to be collected and can help ensure consistency of data collection among team
members. Crump and Scott (1994) provide an example data sheet for VES that
includes some basic individual-level data. If designed with foresight, forms can
also expedite data entry.
15.2.6.5 Field logistics

Poorly organized trips typically result in poor-quality data collection. Careful
advance planning for the transportation of the team, sampling equipment, and
basic supplies will be required, particularly for large teams and remote localities.
Logistics also includes coordinating the safety of the team. Injuries and illness
in remote locations can be dangerous and hinder project progress. In advance
of fieldwork the field team should prepare a well-stocked medical kit, review
vaccines and allergies of each participant, and plan for various possible medical
emergencies that could arise.
15.3 In the field

After the planning and preparation has been completed, the team goes to
the field and begins sampling in the manner indicated by the sampling plan.
Situations will naturally arise that prevent the sampling protocols from being
followed exactly. RAs are intended to be flexible enough to allow field teams to
respond to unforeseen roadblocks as they arise. The field team should keep the
primary objectives of the project in mind if they are required to alter plans in the
field (Young et al. 2000).
15.4 Compiling data and interpreting results

Producing information that is useful to conservation biologists, land managers,
and policy-makers requires skillful synthesis of the large volume of field data
generated. If data-management strategies were developed and the data and
metadata have been entered into the database, work can now focus on ana-
lysis and interpretation. RA data summaries typically focus on species lists
(Kouamé et al. 2007). Species lists may be organized to highlight species
sampled in different primary sampling localities, particular habitat types
(e.g. stream, pond, litter), points along environmental gradients (e.g. eleva-
tion; Menegon and Salvidio 2005), indicator taxa (e.g. disturbance-tolerant
versus -intolerant; Kerr et al. 2000), endemicity or conservation status (e.g.

IUCN Red List status). The best summary approach will depend on the goals
of the RA. Because sampling effort often varies among sites and may have
relied to some degree on opportunistic sampling, it is often difficult to make
direct comparisons between species lists among sites across or even within
RA studies. In some cases, species accumulation or rarefaction curves can be
used to show how rapidly the number of species sampled at a specific site or
habitat increased with sampling effort (e.g. Chapter 18; Gotelli and Colwell
2001; Magurran 2003) to facilitate comparison at a given sampling effort.
Similarly, in some cases it may be possible to use total diversity estimators
to extrapolate from incidence sampling data to estimate the total number of
amphibian species at a site (e.g. Chapter 18; Chao et al. 2005). When using
VES data, however, investigators should clearly defi ne their sampling unit
and be aware how their choice could influence the diversity analyses.
15.5 Recommendations and reporting

Because RAs are typically motivated from a conservation or management per-
spective, recommendations will focus on how the information generated in
the RA can be used to maintain long-term faunal diversity of the study sites.
However, recommendations should be firmly based on the study results, not
on preconceived ideas about the sites importance, and should consider the
target audience and their ability to follow through on the recommendations.
Little benefit comes from recommendations for which the target audience has
no ability to implement (Young et al. 2000). Ways in which RA results are
typically used to guide decision-making include providing direction for future
land acquisition, identifying monitoring needs, establishing priorities for future
research, suggestions for how to zone for multiple use, and for identifying poten-
tial threats to species of interest (Young et al. 2000).
15.6 Summary
Rapid assessment is an approach that can be useful for conservation planning
because it can provide an efficient preliminary characterization of the amphibian
fauna. However, because the methods are often qualitative or semi-quantitative,
the RA approach is of most value when the fauna of the study areas is unknown
and when even qualitative information is needed urgently to help decision-
making. When resources, time, and project objectives allow, researchers should
consider more quantitative approaches.
15.7 References
Aguirre, A. A. and Lampo, M. (2006). Protocolo de bioseguridad y cuarentena para preve-
nir la transmisión de enfermedades en anfibios. In A. Angulo, J. V. Rueda-Almonacid,
J. V. Rodríguez-Mahecha, and E. La Marca (eds), Técnicas de Inventario y Monitoreo
para los Anfibios de la Región Tropical Andina, pp. 73–92. Conservación Internacional,
Bogotá.
Aichinger, M. (1987). Annual activity patterns of anurans of anurans in a seasonal
Neotropical environment. Oecologia, 71, 583–92.
AmphibiaWeb (2008). Information on Amphibian Biology and Conservation. AmphibiaWeb,
Berkeley, CA. http://amphibiaweb.org.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004). Estimating site occupancy
Buckland, S. T., Anderson, D. R., Burnham, K. P., and Laake, J. L. (1993). Distance
Sampling: Estimating Abundance of Biological Populations. Chapman and Hall,
London.
Buckland, S. T., Anderson, D. R., Burnham, K. P., Laake, J. L., Borchers, D. L., and
Thomas, L. (2001). Introduction to Distance Sampling: Estimating Abundance of
Biological Populations. Oxford University Press, New York.
Buckland, S. T., Anderson, D. R., Burnham, K. P., Laake, J. L., Borchers, D., and
Thomas, L. (2004). Advanced Distance Sampling: Estimating Abundance of Biological
Populations. Oxford University Press, Oxford.
Campbell, H. W. and Christman, S. P. (1982). Field techniques for herpetofaunal com-
munity analysis. In N. J. Scott, Jr (ed.), Herpetofaunal Communities, pp. 193–200.
Wildlife Research Report 13. U.S. Department of the Interior, Fish and Wildlife
Service, Washington DC.
Chao, A., Chazdon, R. L., Colwell, R. K., and Shen, T.-J. (2005). A new statistical
approach for assessing similarity of species composition with incidence and abundance
data. Ecology Letters, 8, 148–59.
Corn, P. S. and Bury, R. B. (1990). Sampling Methods for Terrestrial Amphibians and Reptiles.
General Technical Report, PNW-GTR-256. U.S. Forest Service, Fort Collins, CO.
Crump, M. L. and Scott, Jr, N. J. (1994). Visual encounter surveys. In W. R. Heyer,
M. A. Donnelly, R. W. McDiarmid, L. A. C. Hayek, and M. S. Foster (eds), Measuring
and Monitoring Biological Diversity, Standard Methods for Amphibians, pp. 84–92.
Donnelly, M. A. (1989). Demographic effects of reproductive resource supplementation
in a territorial frog, Dendrobates pumilio. Ecological Monographs, 59, 207–21.
Fellers, G. M. and Freel, K. L. (1995). A standardized protocol for surveying aquatic amphib-
ians. Technical Report NPS/WRUC/NRTR-95-01. U.S. National Park Service,
Davis, CA.
Ficetola, G. F., Miaud, C., Pompanon, F., and Taberlet, P. (2008). Species detection using
environmental DNA from water samples. Biology Letters, 4, 423–5.
Fogarty, J. H. and Vilella, F. J. (2001). Evaluating methodologies to survey Eleutherodactylus
frogs in montane forests of Puerto Rico. Wildlife Society Bulletin, 29, 948–55.
Fouquet, A., Gilles, A., Vences, M., Marty, C., Blanc, M., and Gemmell, N. J. (2007).
Underestimation of species richness in Neotropical frogs revealed by mtDNA analyses.
PLoS ONE, 2, e1109.
Funk, W., Almeida-Reinoso, D., Nogales-Sornosa, F., and Bustamante, M. (2003).
245–56.
Goldberg, C. S., Kaplan, M. E., and Schwalbe, C. R. (2003). From the frog’s mouth:
buccal swabs for collection of DNA from amphibians. Herpetological Review, 34,
220–1.
Gonser, R. A. and Collura, R. V. (1996). Waste not, want not: toe-clips as a source of
DNA. Journal of Herpetology, 30, 445–7.
Gotelli, N. J. and Colwell, R. K. (2001). Quantifying biodiversity: procedures and pitfalls
in the measurement and comparison of species richness. Ecology Letters, 4, 379–91.
Gotelli. N. J. and Ellison, A. M. (2004). Chapter 8: Managing and curating data. A
Primer of Ecological Statistics, pp. 207–36. Sinauer Associates, Sunderland, MA.
Hairston, N. G. (1980). The experimental test of an analysis of field distributions:
Competition in terrestrial salamanders. Ecology, 61, 817–26.
Hayek, L.-A. (1994). Research design for quantitative amphibian studies. In W. R. Heyer,
Heyer, W. R., Donnelly, M. A., McDiarmid, R. W., Hayek, L. A. C., and Foster, M. S.
(eds) (1994). Measuring and Monitoring Biological Diversity, Standard Methods for
Howell, K. M. (2002). Amphibians and reptiles: Herptiles. In G. Davies, (ed.), African
Forest Biodiversity, pp. 17–44. Earthwatch, London.
IUCN, Conservation International, and NatureServe (2006). Global Amphibian
Assessment. http://globalamphibians.org.
Jacobs, J. F. and Heyer, W. R. (1994). Collecting tissue for biochemical analysis. In
W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. A. C. Hayek, and M. S. Foster
(eds), Measuring and Monitoring Biological Diversity, Standard Methods for Amphibians,
pp. 103–7. Smithsonian Institution Press, Washington DC.
Kerr, J. T., Sugar, A., and Packer, L. (2000). Indicator taxa, rapid biodiversity assessment,
and nestedness in an endangered ecosystem. Conservation Biology, 14, 1726–34.
Kouamé, N. G., Bonteng, C. O., and Rödel, M.-O. (2007). A rapid survey of the amphib-
ians from Atewa Range Forest Reserve, Eastern Region, Ghana. In J. McCullough,
L. E. Alonso, P. Naskrescki, H. E. Wright, and Y. Osei-Owusu (eds), A Rapid Biological
Assessement of the Atewa Range Forest Reserve, Eastern Ghana. RAP Bulletin of Biological
Assessment 47, pp. 76–83. Conservation International, University of Chicago Press,
Chicago, IL.
Lips, K. R., Reaser, J. K., Young, B. E., and Ibáñez, R. (2001). Amphibian monitoring in
Latin America: a protocol manual/Monitoreo de Anfibios en América Latina: Manual
de protocolos. SSAR Herpetological Circular no. 30.
Magnusson, W. E. and Mourão, G. (2004). Statistics without Math. Sinauer Associates,
Sunderland, MA.
Magurran, A. E. (2003). Measuring Biological Diversity. Blackwell Publishing, Oxford.

Matthews, K. R., Knapp, R. A., and Pope, K. L. (2002). Garter snake distributions in
high-elevation aquatic ecosystems: is there a link with declining amphibian popula-
tions and non-native trout introductions? Journal of Herpetology, 36, 16–22.
(2007). Making great leaps forward: Accounting for detectability in herpetological field
McCallum, M. L. (2007). Amphibian declines or extinction? Current declines dwarf
background extinction rate. Journal of Herpetology, 41, 483–91.
McCullough, J., Alonso, L. E., Naskrecki, P., Wright, H. E., and Osei-Owusu, Y.
(eds) (2007). A Rapid Biological Assessment of the Atewa Range Forest Reserve, Eastern
Ghana. RAP Bulletin of Biological Assessment 47, 1–191. Conservation International,
Arlington, VA.
McDiarmid, R. W. (1994). Preparing amphibians as scientific specimens. In W. R. Heyer,
Menegon, M. and Salvidio, S. (2005). Amphibian and reptile diversity in the southern
Udzungwa Scarp Forest Reserve, south-eastern Tanzania. In B. A. Huber, B. J. Sinclair,
and K.-H. Lampe (eds), African Biodiversity. Molecules, Organisms, Ecosystems,
pp. 205–12. Springer, New York.
Mitchell, J. C. (2006). Inventory of Amphibians and Reptiles of Fredericksburg and
Spotsylvania National Historic Battlefield. Technical Report/NERCHAL/NRTR-
200600/000. U.S. National Park Service, Fredericksburg, VA.
Nagendra, H. and Gadgil, M. (1999). Biodiversity assessment at multiple scales: linking
remotely sensed data with field information. Proceedings National Academy of Sciences
USA, 96, 9154–8.
Olson, D. H., Leonard, W. P. and Bury, R. B. (eds) (1997). Sampling Amphibians in Lentic
Habitats: Methods and Approaches for the Pacific Northwest: Northwest Fauna No. 4.
Pawar, S., Koo, M. S., Kelley, C., Ahmed, M. F., Chaudhuri, S., and Sarkar, S. (2007).
Conservation assessment and prioritization of areas in Northeast India: priorities for
amphibians and reptiles. Biological Conservation, 136, 346–61.
Pearman, P. B., Velasco, A. M., and Lopez, A. (1995). Tropical amphibian monitoring:
a comparison of methods for detecting inter-site variation in species composition.
Raxworthy, C. J., Martinez-Meyer, E., Nussbaum, R. A., Schneider, G. E., Ortega-
Huerta, M. A., and Peterson, A. T. (2003). Predicting distributions of known and
unknown reptile species in Madagascar. Nature, 426, 837–41.
Roberts, L. (1991). Ranking the rainforests. Science, 251, 1559–60.
Rodda, G. H., Campbell, E.W, Fritts, T. H., and Clark, C. S. (2007). The predictive
power of visual searching. Herpetological Review, 36, 259–64.
Rödel, M.-O. and Ernst, R. (2004). Measuring and monitoring amphibian diversity in
tropical forests. I. An evaluation of methods with recommendations for standardiza-
tion. Ecotropica, 10, 1–14.
Roelants, K., Gower, D. J., Wilkinson, M., Loader, S. P., Biju, S. D., Guillaume, K.,
Moriau, L., and Bossuyt, F. (2007). Global patterns of diversification in the history
of modern amphibians. Proceedings of the National Academy of Sciences USA, 104,
887–92.
Rueda-Almonacid, J. V., Castro, F., and Cortez, C. (2006). Técnicas para el inventario
y muestreo de anfibios: Una compilación. In A. Angulo, J. V. Rueda-Almonacid,
J. V. Rodríguez-Mahecha, and E. La Marca (eds), Técnicas de Inventario y
Monitoreo para los Anfibios de la Región Tropical Andina, pp. 135–71. Conservación
Internacional, Bogotá.
Sayre, R. (2000). Overview: Rapid Ecological Assessment after ten years. In R. Sayre,
E. Roca, G. Sedaghatkish, B. Young, S. Keel, R. L. Roca, and S. Sheppard (eds), Nature
in Focus Rapid Ecological Assessment, pp. 1–18. Island Press, Washington DC.
Sayre, R. and Roca E. (2000). Careful planning: a key to success. In R. Sayre, E. Roca,
G. Sedaghatkish, B. Young, S. Keel, R. L. Roca, and S. Sheppard (eds), Nature in Focus
Rapid Ecological Assessment, pp. 33–44. Island Press, Washington DC.
Sayre, R., Roca, E., Sedaghatkish, G., Young, B., Keel, S., Roca, R. L., and
Sheppard S. (eds) (2000). Nature in Focus: Rapid Ecological Assessment. Island
Press, Washington DC.
Schmidt, B. (2004). Declining amphibian populations: the pitfalls of count data in the
study of diversity, distributions, dynamics, and demography. Herpetological Journal,
14, 167–74.
Schmidt, B. (2005). Monitoring the distribution of pond breeding amphibians when spe-
cies are detected imperfectly. Aquatic Conservation: Marine and Freshwater Ecosystems,
15, 681–92.
Seutin, G., White, B. N., and Boag, P. T. (1991). Preservation of avian blood and tissue
samples for DNA analyses. Canadian Journal of Zoology, 69, 82–90.
Smith, M. A., Poyarkov, Jr, N. A., and Hebert, P. D. N. (2008). CO1 DNA barcod-
ing amphibians: take the chance, meet the challenge. Molecular Ecology Resources, 8,
235–46.
Thomas, L., Buckland, S. T., Burnham, K. P., Anderson, D. R., Laake, J. L., Borchers, D. L.,
and Strindberg, S. (2002). Distance sampling. In A. H. El-Shaarawi and W. W. Piegorsch
(eds), Encyclopedia of Environmetrics, vol. 1, pp. 544–52. John Wiley & Sons, Chichester.
Timm, R. M. (1994). The mammal fauna. In L. A. McDade, K. S. Bawa, H. A. Hespenheide,
and G. S. Hartshorn (eds), LaSelva: Ecology and Natural History of a Neotropical Rainforest,
Vences, M., Thomas, M., Bonett, R. M., and Vieites, D. R. (2005). Deciphering
amphibian diversity through DNA barcoding: chances and challenges. Philosophical
Transactions of the Royal Society of London Series B Biological Sciences, 360, 1859–68.
Vilella, F. and Fogarty, J. (2005). Diversity and abundance of forest frogs (Anura:
Leptodactylidae) before and after Hurricane Georges in the Cordillera Central of
Puerto Rico. Caribbean Journal of Science, 41,157–62.
Vonesh, J. R. (2000). Dipteran predation on the arboreal eggs of four Hyperolius frog spe-
cies in western Uganda. Copeia, 2000, 560–6.
Vonesh, J. R. (2001a). Patterns of richness and abundance in a tropical African leaf-litter
herpetofauna. Biotropica, 33, 502–10.
Vonesh, J. R. (2001b). Natural history and biogeography of the amphibians and reptiles
of Kibale National Park, Uganda. Contemporary Herpetology. http://www.nhm.ac.uk/
hosted_sites/ch/.
Whiles, M. R., Lips, K. R., Pringle, C. M., Kilham, S. S., Bixby, R. J., Brenes, R.,
Connelly, S., Colon-Gaud, J. C., Hunte-Brown, M., Huryn, A. D. et al. (2006). The
effects of amphibian population declines on the structure and function of Neotropical
stream ecosystems. Frontiers in Ecology and the Environment, 4, 27–34.
Animal Populations. Academic Press, New York.
Young, B., Sedaghatkish, G., and Roca, R. (2000). Fauna surveys. In R. Sayre, E. Roca,
G. Sedaghatkish, B. Young, S. Keel, R. L. Roca, and S. Sheppard (eds), Nature in Focus
Rapid Ecological Assessment. pp. 93–117. Island Press,Washington DC.
16
Auditory monitoring of anuran populations
Michael E. Dorcas, Steven J. Price, Susan C. Walls, and
William J. Barichivich
16.1 Introduction
Because anurans rely on vocalization for most communication, detection of
species-specific calls provides relatively efficient mechanisms for studying and
evaluating the status of anuran populations. Consequently, most amphibian
monitoring programs focus on anurans and use call detection as their sole or pri-
mary monitoring technique. The overall goal of these programs is to determine
and monitor the status of populations over time. Some monitoring programs
may actually attempt some form of quantification of anuran population size or
density, but this is often difficult when using only calling data. In this chapter we
describe approaches for monitoring anuran populations based solely on auditory
techniques. These approaches include manual calling surveys (MCS), automated
recording systems (ARS), or some combination of the two. MCS can be used by
researchers to monitor multiple amphibian populations, often over large spatial
scales. ARS can be used by researchers interested in intensively monitoring popu-
lations of anurans at a single or a few locations. In this chapter, we describe both
approaches and the potential costs and benefits of each. We also explain how
data resulting from ARS can be used to optimize manual survey protocol and to
interpret data resulting from MCS. Our goal is to provide an overview of these
techniques and questions that can be addressed using them.
16.2 MCS
Generally, MCS simply involve observers listening to the vocalizations of male
frogs and recording all species detected for the duration of the survey or for a
given area. In many surveys, observers also score abundance of each species
heard vocalizing. Researchers have long used MCS to conduct inventories of

anurans in a given area. For example, Wright and Wright (1949) understood
the importance of MCS when they stated that “If one knows the frog notes,
he can in one night do more work on frog distribution than he might other-
wise do in years.” However, only within the last 30 years, when concerns of
declining amphibian populations prompted the development of formal moni-
toring programs, have MCS been widely used for large-scale investigations of
anuran populations. The widespread use of MCS for monitoring anurans has
led numerous investigations into survey protocol, study design, and efficiency
as a technique to detect anurans.
16.2.1 History and current status of MCS

Many historic and recent small-scale investigations have used MCS to detect
anurans for ecological, behavioral, and conservation-related investigations (i.e.
Martof 1953; Blair 1961; Woolbright 1985; Knutson et al. 1999). In this chapter,
we emphasize MCS approaches that have specific objectives to monitor the pro-
portion of sites where a given anuran species is observed, accounting for the poten-
tial effects of imperfect detection and other biases (site and sampling covariates
such as habitat type and weather conditions). This metric is known as occupancy
(MacKenzie et al. 2002, 2006).
Most MCS programs use surveys in which observers listen to anuran vocaliza-
tions at several points along a roadside transect or at designated sampling loca-
tions. Since the early 1980s, the use of roadside surveys to detect calling anurans
has become used widely by many local and regional environmental monitoring
programs. At least five Canadian provinces and 28 US states conduct, or have
conducted, roadside surveys for calling anurans (Weir and Mossman 2005). If
similar protocols are used, MCS data collected at the local or regional level (e.g.
state or province) can be incorporated into a centralized database allowing for the
evaluation of anuran populations at the national level. Two US programs initi-
ated by the Department of Interior’s US Geological Survey, the North American
Amphibian Monitoring Program (NAAMP; www.pwrc.usgs.gov/naamp/; Weir
and Mossman 2005), and the Amphibian Research and Monitoring Initiative
(ARMI; http://armi.usgs.gov/), are good examples of this. Many US Geological
Survey scientists use both MCS and ARS (see section 16.3) to meet their moni-
toring objectives, primarily on federal lands. Additional examples of North
American MCS include the Marsh Monitoring Program in the Great Lakes
region (Weeber and Vallianatos 2000), Ontario Backyard Frog Survey (de Solla
et al. 2005), and FrogWatch USA (Inkley 2006). MCS are also used to monitor
trends in anuran populations in Australia (Australia Frog Census; Walker 2002),
16 Auditory monitoring of anurans | 283
as well as some European (e.g. Anthony 2002; Pellet and Schmidt 2005; Schmidt
2005; Scott et al. 2008), and Central American countries (e.g. Kaiser 2008).
16.2.2 Study objectives: what can MCS tell us?

The objective of any contemporary MCS is typically to conduct an inventory
of an area of interest to determine the status of anurans in that region and/or
to monitor populations and communities, over time, to address how species are
responding to habitat change, climatic variation, or some other ecological or
management issue. An objective of many monitoring programs, including those
using MCS that are rapidly gaining recognition, is that of determining site occu-
pancy (Marsh and Trenham 2008).
16.2.3 Survey design

Establishment of sampling sites is a critical component of all monitoring pro-
grams for which it is essential that sites are sampled according to a probabilistic
scheme, so that statistical inference may be extrapolated to a defined area of
interest. According to the NAAMP protocol (Weir and Mossman 2005), sur-
veys are conducted along routes that generally are established along roads, dikes,
waterways, or other points of access. Routes consist of 10 stops that are either
placed at least 0.8 km apart or are stratified by habitat. Although NAAMP is
often used as a template for designing MCS, roads are not typically established
in a randomized fashion, thus illustrating one of the weaknesses of the NAAMP
approach. One compromise for conducting road surveys and meeting this cri-
terion is to, for example, randomly select stops along a given route that are at
least 0.8 km apart. In this modification the area of inference would be limited
to the route, rather than to a larger area of interest (e.g. a large protected area).
Determining the number of routes and their associated sites/stops is another
critical element if the resulting data are destined for occupancy analysis. The
number of sites to be sampled varies depending upon the distribution and nat-
ural history of the anuran species being monitored, the question of interest, and
other factors. But, generally speaking, a good rule of thumb in determining the
number of sites to sample is usually as many as possible, typically at least 50,
assuming that sites are sampled for a minimum of four seasons (see Figure 7.3 of
MacKenzie et al. 2006).
MCS are usually conducted at night, starting one-half hour after sunset, and
are completed by 01:00 h. For many anuran species, especially those in temperate
regions, peak calling does occur within this sampling time. Bowers et al. (1998)
found that the majority of species in their surveys ended peak calling at midnight.
To minimize the potential effect of anthropogenic noise (see section 16.2.4),
researchers typically allow 1 min between arriving at each site and beginning the
chorus survey. All anuran species heard in a specified time frame (usually 5 min,
but see section 16.2.4) are recorded; each calling anuran species is given a score,
known as an amphibian calling index (ACI), ranging from 1 to 3, where 1 means
distinct calls of individuals that can be counted and have no overlapping calls,
2 means calls of individuals that can be distinguished but have some overlap-
ping calls, and 3 means a full chorus, with calls of individuals indistinguishable.
Generally, MCS should not be conducted during heavy rain, high wind, or other
inclement weather that could affect the detection of calling anurans. At each site,
air temperature, relative humidity, barometric pressure, wind speed, and other
variables are often measured immediately after recording call data to be used as
sample covariates in occupancy analysis. These covariates can be used to help
explain variation in site occupancy or detection probabilites.
In regions where anurans vocalize in a somewhat predictable manner, many
MCS, such as NAAMP, recommend sampling each stop along a roadside route
during a specified sampling period, such as during early spring, late spring, and
summer. The NAAMP protocol recommends a single survey, based on conveni-
ence to volunteer observers, despite studies that have shown that significant vari-
ation in calling behavior does occur within the specified sampling periods (e.g.
Todd et al. 2003; Gooch et al. 2006; Kirlin et al. 2006). For example, Gooch
et al. (2006) conducted three MCS during the NAAMP-specified “summer”
sampling period in the western Piedmont region of North Carolina, USA, and
found that detection probabilities increased for some species (e.g. Acris crepitans,
Rana catesbeiana) and decreased for other species (e.g. Rana clamitans) from
survey 1 to survey 3 (Figure 16.1). Variation in calling behavior with prescribed
sampling periods may cause the observer to not detect a species if single sampling
occasions are employed. At least two surveys are required to calculate detection
probabilities (preferably a minimum of three; J. Nichols, personal communi-
cation) during a sampling period (see MacKenzie et al. 2002 and Table 6.1 of
MacKenzie et al. 2006 for more details).
16.2.4 Other survey-design issues to consider:

the efficiency of MCS
The widespread use of MCS has led to numerous investigations that focus on
the efficiency of MCS to detect anurans. In fact, few if any amphibian survey
methodologies have been scrutinized to the extent of MCS. A central issue in all
wildlife-monitoring programs, including MCS, is that of imperfect detection
(MacKenzie et al. 2002, 2006). For example, even during the peak breeding
season for a given anuran species, calling does not occur each night; variations
1.0 Survey 1
0.9 Survey 2
Survey 3
0.8
Detection probability (p)
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
Acris crepitans Hyla chrysoscelis Rana catesbeiana Rana clamitans
Fig. 16.1 Detection probability, p ( 1 standard error), for four summer-breeding

anurans in the western Piedmont region of North Carolina, USA. MCS were
conducted from 10 June through 13 July 2004. Detection probabilities were
calculated using a model with survey-specific p and constant occupancy estimate,
ψ (i.e. ψ·p(t)), allowing for the calculation of p for each species during each of the
three surveys within the sampling period. Note the influence of sampling occasion
on p for Acris crepitans and Rana clamitans. Adapted from Gooch et al. (2006).
in species-specific calling behaviors and abiotic and biotic conditions can lead to
the inference of absence despite the presence of a species. Aspects of survey proto-
col and observer bias can also influence detection probabilities. Fortunately,
recent advances in statistical techniques have allowed for calculation of species-
specific detection probabilities (e.g. program PRESENCE; MacKenzie et al.
2002, 2006; Chapter 24), which can greatly aid in inferring population status
and potentially long-term population trends.
Inter- and intraspecific variation in anuran calling behavior may affect detec-
tion probability and should be considered when conducting a MCS or evaluat-
ing MCS data. Anurans exhibit a vast array of acoustic properties (Duellman
and Trueb 1986) which influence probability of detection by observers. Some
species have calls that can carry long distances (e.g. 1 km), whereas calls of
other species cannot be detected until the observer is 100 m or less from the
breeding site. Many species, such as R. catesbeiana, may call sporadically every
few minutes. Other species (e.g. Pseudacris crucifer) call more continuously. In
species-rich communities, louder, higher-pitch calls of one species may interfere
with the detection of other, quieter species (Droege and Eagle 2005) or inhibit
calling in another sympatric species (Littlejohn and Martin 1969). In general,
MCS are best suited for regions where all species vocalize during a somewhat
predictable breeding season. Species that breed in response to localized heavy

rains (e.g. Spea and Scaphiopus), call quietly or not at all (e.g. Ascaphus truei),
call infrequently (e.g. Rana capito), or have relatively short breeding (i.e. vocal-
izing) seasons (e.g. Rana sylvatica) should not be monitored exclusively by MCS.
Droege and Eagle (2005) suggest that MCS are suitable for detecting and infer-
ring population trends for approximately 55 of the 103 anuran species in North
America.
Abiotic factors also influence calling behavior in many anuran species (Blair,
1961) and thus influence detection probability via MCS. Because anurans are
ectotherms, air temperature has been shown to influence calling, especially for
species that vocalize during winter and early spring (e.g. Todd et al., 2003;
Weir et al. 2005; Kirlin et al. 2006). Weir et al. (2005) found that air tempera-
ture explained calling variation in eight of 10 species studied; however, five
species displayed preference for particular temperatures, indicating an optimal
temperature for detection. Additionally, Pellet and Schmidt (2005) found that
air temperature was a predictor of calling in Hyla arborea in Switzerland, with
greater detections during warm temperatures. For winter-breeding species in
temperate zones, calling behavior may be more sporadic and limited to day-
light or early evening hours when temperatures are moderate (Todd et al. 2003;
Kirlin et al. 2006; Saenz et al. 2006).
Abiotic factors other than air temperature have also been shown to influ-
ence anuran calling. For example, Oseen and Wassersug (2002) suggested that
water temperature was the overall most important predictor of calling behav-
ior of some Canadian anurans. Precipitation can also influence calling because
many species are known to call more intensely after periods of heavy rain (Blair
1961; Oseen and Wassersug 2002). However, heavy rainfall during MCS is
not recommended as it may obfuscate anuran vocalizations and decrease prob-
ability of detection. Other factors, such as wind speed (Weir et al. 2005; Oseen
and Wassersug 2002; Johnson and Batie 2001), humidity of air (Oseen and
Wassersug 2002), barometric pressure (Oseen and Wassersug 2002), and
moonlight (Weir et al. 2005), may also influence anuran detection probabil-
ity. However, anuran species respond differently to abiotic factors (Saenz et al.
2006). Knowledge of the relationship between anuran breeding strategies and
abiotic factors prior to conducting MCS may allow for increased efficiency and/
or detection.
Anthropogenic noise also likely impacts anuran detection probability. Most
calling anurans will respond to anthropogenic noise disturbance. For this rea-
son, some MCS programs recommend that after arriving at a stop the obser-
ver must wait for a few minutes prior to conducting the survey. However, few
studies have been conducted on the effects of noise on anuran calling behavior.
Weir et al. (2005) found that only three of 10 species decreased calling behavior
due to increased traffic. Regardless, anthropogenic noise can affect the obser-
ver’s ability to detect frogs and should be recorded during MCS. Additionally,
anthropogenic noise can also reduce a species proclivity to call, thus lowering its
detectability (Sun and Narins 2005; C. Steelman, personal communication)
Several studies have investigated the effects of survey length on anuran detec-
tion. The majority of MCS range from 3 to 10 min per stop. Pierce and Gutzwiller
(2004) found that 15 min was required to detect 90% of all species known to be
present, whereas Shirose et al. (1997) found that 3 min surveys were adequate to
detect most species. Gooch et al. (2006) found that 94% of summer-breeding
anurans in the North Carolina Piedmont region were detected within the first
5 min of the MCS and detection probabilities were slightly higher as observers
spent longer listening (3 min compared to 10 min). However, these and other
investigations highlight that some species may go undetected even if surveys are
extended up to an hour. The length of survey should be determined based on
the specific objectives of the MCS; however, for detecting long-term population
trends for most species, 3–5 min appears to be adequate.
Variation among observers can also substantially influence the quality of
MCS data. Some large-scale MCS (e.g. NAAMP) rely heavily on volunteers,
who may vary in experience and/or hearing ability. As a result, observers may
not detect all species present, include species not present, or incorrectly identify
vocalizations, resulting in flawed assessments of anuran populations (Lotz and
Allen 2007). Weir et al. (2005) found that volunteer experience may influence
species detection. However, studies by Genet and Sargent (2003), Shirose et al.
(1997), and Lotz and Allen (2007) suggest that even relatively inexperienced
volunteers were reliable in their abilities to determine species, yet estimating
abundance categories often differed among observers. Pierce and Gutzwiller
(2007) showed 79% agreement among nine observers conducting MCS in cen-
tral Texas, USA, and stressed the importance of accounting for interobserver
variation in data analysis. NAAMP currently requires all volunteers to pass an
online anuran call test (available at www.pwrc.usgs.gov/frogquiz/) prior to con-
ducting MCS. In general, training of volunteers will likely increase the prob-
ability of ensuring quality data.
16.2.5 Limitations of MCS data

MCS are excellent tools for monitoring changes in anuran occupancy or for
inventorying which species occur in an area. At a given site, however, these
surveys only yield qualitative data on abundance as obtained through the use of
the ACI (section 16.2.3). These types of abundance data are of limited utility in
assessing population densities. However, Nelson and Graves (2004) compared
population estimates (via mark–recapture) of male R. clamitans with ACI col-
lected at the same sites and found abundance to be correlated positively with
ACI. In contrast, Corn et al. (2000) found that another measure of call fre-
quency failed to reflect “a relatively large population” of Bufo woodhousii and
only weakly distinguished among different-sized populations of Pseudacris
maculata. The need to be able to use indices of relative abundance, such as the
calling index, in occupancy models has been conceptualized and is currently an
active area of research (Royle and Nichols 2003; Dorazio 2007).
Because call surveys are based upon the vocalizations of adult male anurans,
they do not provide complete information on population structure; that is,
non-calling females and subadults are not assessed by this method (Stevens and
Paszkowski 2004). Similarly, this method is not useful for other non-calling
amphibians, such as salamanders and caecilians. The males of many species of
anurans will vocalize in contexts not necessarily related to breeding; moreover,
the MCS does not consider the presence of egg masses, tadpoles, or metamorph-
osing juveniles in the population. Thus, the MCS provides no information
about whether there was successful reproduction at a given site. Depending on
the questions of interest in a given study, the MCS is therefore most useful
when used in conjunction with other survey methods, such as visual encoun-
ter surveys and the use of traps or dipnets to assess the larval component of the
population.
16.3 ARS
In addition to MCS, automated systems (ARS) can be used to detect anuran
vocalizations and can be useful in monitoring many anuran populations
(Peterson and Dorcas 1994). Typically, ARS (or so-called frogloggers) are used
to collect data intensively at a single or a few locations, whereas MCS provide
more superficial data but for a larger number of sites. ARS can be used to survey
for anuran species in places difficult to access for MCS and can be left in the field
for extended periods of time, thus increasing the probability of detecting a given
species. ARS may be the only practical way to reliably detect species that have
very short or unpredictable breeding seasons, such as R. capito. ARS minimize
disturbance to calling anurans and provide a permanent sampling record that
can be evaluated by multiple experts if required (Mohr and Dorcas 1999; Todd
et al. 2003). When combined with information on environmental variation,
data from ARS can be incorporated into models that can be used to optimize
monitoring programs based on MCS (Bridges and Dorcas 2000; Oseen and
Wassersug 2002).
16.3.1 Sources for ARS

At the time Peterson and Dorcas (1994) published on building and using ARS,
there were few options for their procurement. Although not specifically intended
for anuran monitoring, the Cornell Laboratory of Ornithology builds robust ARS
units that have been used in a wide variety of tasks ranging from the detection of
ivory-billed woodpeckers (Campephilus principalis) to monitoring whales around
the world’s oceans. Bedford Technical (www.frogloggers.com) and Wildlife
Acoustics (www.wildlifeacoustics.com) have both produced ARS systems with
a variety of options that make them attractive for monitoring frogs. Although
their software has not been tested on anuran vocalizations, both Cornell Lab of
Ornithology and Wildlife Acoustics produce products, Raven and Song Scope
respectively, which may prove useful in automated call detection.
16.3.2 Construction, deployment, and retrieval of data

Although sources for ARS now exist, some biologists may still choose to build
their own. This may be a logistical hurdle for some investigators; however, con-
struction of an ARS can have the added benefit of familiarizing the user with the
inner workings of their equipment. Construction of an ARS generally requires
combining several simpler components: a recorder, timer/controller, micro-
phone, power supply, and housing (Peterson and Dorcas 1994; Barichivich
2003). If the investigator is not familiar with electronics, we recommend work-
ing with someone skilled in building electrical devices.
The core component of an ARS is the recorder. Devices that have been suc-
cessfully used include analog cassette and various digital (e.g. MPEG-1 Audio
Layer 3, or MP3) recorders. Selection of a recorder is dependent on the required
recording quality, capacity, and budget. Recordings produced on inexpensive
analog cassette recorders are sufficient for manual listening, but higher sam-
pling rates and frequency responses may be required in some circumstances.
Additionally, storage capacity can usually be greatly increased when using
digital recordings.
In most cases, researchers choose to make recordings at specific times of day
(e.g. dusk) and on set intervals rather than continuously. Some recording devices
(e.g. PDAs and digital voice recorders) have built-in timers or clocks. The add-
ition of a timer/controller allows programming of the desired recording schedule.
A wide variety of devices, from mechanical (K. Wharton, personal communica-
tion) to computer microprocessors (Acevedo and Villanueva-Rivera 2006), have
been used for this purpose. Timers generally work in one of two ways: controlling
the function of the recorder or interrupting the power supply to the recorder. In
addition to time-based triggers, environmental triggers can be used to activate an
ARS. For example, to detect explosive breeders like Spea or Scaphiopus, a tipping-
bucket rain gauge with a reed valve could be used so a rainfall event would trip
an ARS into service.
The choice of microphones is a crucial decision, as recordings will only be as
good as the microphone, regardless of the recording device. At least nine types
of microphone are available but condenser or dynamic varieties are those most
commonly used. Unlike dynamic microphones, condenser microphones require
a power supply. Condenser microphone power sources are often small (e.g. a
single AA battery) and may not be sufficient if the ARS is deployed for extended
periods of time without maintenance. This issue can be addressed by using the
timer/controller to control the microphone as well as the recorder. Another con-
sideration in selecting a microphone is the directional or acceptance cone. Cones
can range from a 360° circle around the microphone (omnidirectional) to just a
few degrees in front of the microphone (unidirectional or shotgun). To capture
the vocalizations of species that call while partially or completely submerged
(e.g. Rana sevosa or Rana subaquavocalis), a hydrophone would be more appro-
priate to use than a microphone (Platz 1993).
The final considerations of building an ARS are power supply and pro-
tection. Rechargeable batteries generally provide the best option for most
researchers. Cost, power, size, and weight should be considered when selecting
batteries. In environments with sufficient sunlight, rechargeable batteries can
be supplemented by a solar panel and can generally operate without interrup-
tion. With the exception of the microphone, all components of an ARS should
be firmly mounted inside a protective case. The case should provide adequate
environmental protection and be large enough to hold all the components,
yet small enough to be reasonably portable for ease of field deployment. For
most field deployments, waterproof boxes (e.g. Otter or Pelican brands) work
well, but less costly options exist. These include surplus ammunition cans,
polyethylene coolers, plastic tool boxes, and plastic watertight marine boxes.
Consideration should be given to the environment in which the ARS will be
deployed. In areas frequented by people, the ARS can be locked, hidden, or
even buried. Animals, such as raccoons and bears, may damage equipment
and thus more rugged cases might be needed in some situations (Corn et al.
2000). Microphones should also be protected by a windscreen to reduce wind
noise, along with a cover to shield the windscreen and microphone from the
environment.
Before deploying an ARS, careful consideration should be given to placing

and programming the ARS to capture the data germane to the question(s) of
interest. Although the program should be restrictive enough to minimize the
collection of superfluous data, sampling only during perceived peak times could
lead to erroneous interpretation of data (Figure 16.2). Typically an ARS is placed
near anuran breeding habitats with the microphone mounted facing the water.
After the recordings have been retrieved, data need to be transcribed, either
manually (i.e. by human ear) or by computer recognition. Manual assessment
of recordings requires a trained observer to listen to and review the recordings
while transcribing the data. Calling activity can be scored much like that in
MCS and is typically semi-quantified as an ACI. Depending on observer experi-
ence and the complexity of the calling choruses, 1.5 h per hour of recording are
required to review analog cassette tapes (Corn et al. 2000). Slightly less time
3.0 3.0
Hyla cinerea Rana clamitans
Mean calling intensity
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
3.0 3.0
Hyla gratiosa Rana sphenocephala
Mean calling intensity
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
Hour of day Hour of day
Fig. 16.2 Daily calling pattern of Hyla cinerea, Hyla gratiosa, Rana clamitans, and
Rana sphenocephala recorded using an ARS at Carolina Bay near Aiken, SC, USA.
Mean calling activity was calculated by averaging the recorded calling activity levels
for each 30 min time recording period over all days of the study (from 16 June to
12 July 1997). Error bars denote 1 standard deviation. Note that calling in both
species of treefrog peaked during the time period when manual calling surveys are
recommended (dusk to midnight), but peak calling in Rana peaked well after midnight
and R. sphenocephala called almost exclusively after midnight, and would thus likely be
missed on most calling surveys. Adapted from Bridges and Dorcas (2000).
is usually required to review digital recordings. While in its infancy, computer

recognition of high-quality recordings shows great promise for decreasing the
time, and cost, of manually listening to ARS recordings and potentially increas-
ing the accuracy of call recognition. Though not an automated process, one
study showed that in a mixed-species chorus, including the acoustically domin-
ant coqui (Eleutherodactylus coqui), several species were omitted or significantly
underrepresented using only a manual data review (Villanueva-Rivera 2007).
By adding a simple visual analysis of the spectrograms of digital call recordings
using Adobe Audition, the calls of the less acoustically dominant species were no
longer eclipsed (Villanueva-Rivera 2007). Research on birds demonstrates the
potentially significant time savings when using computer recognition software.
In 12 h, Agranat (2007) scanned more than 250 h of field recordings for bird
vocalizations using Song Scope software. This process allowed a human obser-
ver to review 1552 potential calls of the cerulean warbler (Dendrocia cerulea) in
under 1 h.
16.3.3 Monitoring environmental data

Automatically monitoring environmental data, while simultaneously monitor-
ing anuran calling activity using ARS, can allow interpretation of how envir-
onmental variation affects calling activity (Dorcas and Foltz 1991; Peterson
and Dorcas 1992, 1994; Saenz et al. 2006) and detection probability. Typically,
investigators use dataloggers to monitor variables such as air and water tempera-
tures, relative humidity, solar radiation, precipitation, wind speed, and baromet-
ric pressure. Numerous types of dataloggers of varying quality and capabilities
are available (e.g. Onset Computer Corp., Pocasset, MA, USA; Campbell
Scientific, Logan, UT, USA). Like ARS, some investigators find dataloggers to
be particularly challenging to learn to use. However, recent advances in software
interfaces for nearly all dataloggers make learning to use them simple enough for
even novice researchers.
16.3.4 Answering questions using ARS

ARS can be used by researchers to address basic questions such as whether a spe-
cies is present at a given location or to conduct more detailed investigations of
factors affecting calling activity and population fluctuations (Corn and Muths
2002; Todd et al. 2003). Generally, investigators using ARS for monitoring
purposes are simply interested in detecting whether a species is present. As such,
many ARS systems can be set to only come on and record at certain times of the
day when a particular species is most likely to vocalize (e.g. dusk until midnight),
thus minimizing the number of recordings that must be evaluated. However,
one must be careful not to assume that anurans call primarily during times trad-
itionally recognized as peak calling periods (Bridges and Dorcas 2000).
In some cases, researchers use ARS to attempt to detect a species that is
either rare or has very unpredictable calling patterns. Anurans such as Spea and
Scaphiopus that generally only call under certain conditions (i.e. during or after
heavy rains) may be particularly hard to detect using MCS and ARS may offer
the best opportunity to detect their populations. For some extremely rare species
(e.g. R. sevosa) detecting every known population is important for proper man-
agement and thus, increasing detectability of populations using ARS can play a
vital role in conservation efforts.
Because ARS can collect data at regular and precise intervals, mathematical
models can be developed that allow prediction of when and under what condi-
tions species are likely to call, thus providing data that can be used to optimize
MCS. Typically, data collected simultaneously with anuran calling data, such as
temperature, precipitation, wind speed, and time of day, are used as independ-
ent variables in a logistic regression to predict the optimal conditions to conduct
MCS (i.e. the times when detectability is maximized; Oseen and Wassersug
2002). Anurans may respond differently to environmental variables across their
ranges, and thus development of models should be done for particular regions
as needed. Such models were developed for three species of winter-breeding
anurans (Pseudacris crucifer, Pseudacris feriarum, and Rana sphenocephala) in the
western Piedmont region of North Carolina (Steelman and Dorcas, in press).
Models showed that for P. crucifer day of year, time, precipitation, and water
temperature positively influenced calling and air temperature negatively influ-
enced calling; for P. feriarum time, precipitation, air temperature, and water
temperature positively influenced calling, and day of year negatively influenced
calling; for R. sphenocephala day of year, time, precipitation, and air temperature
positively influenced calling, and higher water temperature negatively influ-
enced calling. The models described for the winter-breeding species above were
tested using previously collected data from MCS along with spot measurements
of environmental data (Kirlin et al. 2006). Models accurately predicted whether
a species was calling approximately 70% of the time.
In addition to using models to predict the best times and conditions to manu-
ally sample anurans, ARS can be used to interpret data collected previously,
assuming sufficient environmental data are collected at the time of the surveys.
Fortunately, nearly all existing MCS programs require collection of at least some
environmental data. To do this, an investigator would insert the spot measure-
ments of environmental variables collected by volunteers at the time the surveys
were conducted into the model equation for each species of interest to generate a
likelihood of calling (e.g. ranging from 0 to 1). Calling likelihoods can then be
used to interpret previously collected calling survey results. If a species was not
heard at a particular location but the model indicates a high likelihood of calling
if it was present (e.g. 0.9), then the investigator can have a higher confidence in
concluding that the species was not present, rather than it simply not vocalizing
and being undetectable.
16.3.5 Limitation of ARS

Despite the advantages of using ARS, there are some disadvantages when com-
pared to surveys based on MCS (Corn et al. 2000; Penman et al. 2005). For
many investigators, ARS can be expensive compared to volunteer-based MCS.
Additionally, ARS can typically only be deployed at one or a few sites, thus
decreasing the number of anuran populations that can be monitored. This can
have major implications for statistical analyses of data when each site is a rep-
licate. The possibility of wildlife damage, vandalism, or theft of equipment is
another issue that may limit investigators’ ability to use ARS in some localities
(e.g. Corn et al. 2000). Although ARS can be hidden relatively easily, one of us
(M.E.D.) has had equipment stolen even in remote locations. Some investiga-
tors find using automated systems particularly challenging, especially if they
consider themselves technologically challenged. However, we have found that
nearly anyone can be taught to use automated systems effectively and for those
that are apprehensive about doing so we suggest initial consultations with a
biologist experienced with the equipment being used.
16.4 Conclusions
Auditory monitoring of anuran populations, either by MCS or by ARS, is a use-
ful tool for assessing the status and trends of populations, as well as the responses
of populations and communities to environmental and anthropogenic change.
The use of auditory monitoring has increased tremendously since the realiza-
tion that amphibian populations are undergoing global population declines,
and has become a standard approach to monitoring in many state, national,
and international programs. Auditory monitoring has limitations in that (1) its
utility in estimating abundance is restricted, (2) this method provides little
information about the complete structure of a population (i.e. the status of
non-calling adult females and subadults) or about whether a given popula-
tion has experienced successful reproduction, and (3) this method cannot be
used on non-calling amphibians, such as salamanders and caecilians. The two
means by which auditory monitoring may be conducted (manual or automated)
differ in their relative costs and benefits, and the appropriate approach for a
given study ultimately depends on the study objectives and resources available.
Nevertheless, MCS can provide vital information on changes in occupancy
states of various sites, which can be overlaid with environmental data to assess
potential causes of change in occupancy for a given species or community. Data
from ARS can be used to optimize the effectiveness of MCS or interpret data
based on MCS. As such, auditory monitoring has great utility in assessing the
extent of declines of anuran amphibians, a topic of heightened concern in ecol-
ogy and conservation biology.
We thank P. Stephen Corn, Erin Muths, and Charlotte K. Steelman for com-
ments that improved the manuscript. Manuscript preparation was partially sup-
ported by the Department of Biology at Davidson College, Duke Power, and
National Science Foundation grant (DEB-0347326) to M.E.D and the U.S.
Geological Survey’s Amphibian Research and Monitoring Initiative to SCW
and WJB. The use of trade or product names does not imply endorsement by
US government agencies.
16.6 References
Acevedo, M. A. and Villanueva-Rivera, L. J. (2006). Using automated digital recording
systems as effective tools for the monitoring of birds and amphibians. Wildlife Society
Bulletin, 34, 211–14.
Agranat, I. D. (2007). Automatic Detection of Cerulean Warblers Using Autonomous
Recording Units and Song Scope Bioacoustics Software. www.wildlifeacoustics.com/
Anthony, B. P. (2002). Results of the first batrachian survey in Europe using road call
counts. Alytes, 20, 55–66.
Barichivich, W. J. (2003). Appendix IV: Guidelines for building and operating remote
field recorders. In C. K. Dodd, Jr, Monitoring Amphibians in Great Smoky Mountains
National Park, pp. 87–96. U.S. Geological Survey Circular no. 1258. U.S. Geological
Survey, Tallahassee, FL.
Blair, W. F. (1961). Calling and spawning seasons in a mixed population of anurans.
Ecology, 42, 99–110.
Bowers, D. G., Anderson, D. E., and Euliss, Jr, N. H. (1998). Anurans as indicator of
wetland condition in the Prairie Pothole Region of North Dakota: an environmen-
tal monitoring and assessment program pilot project. In M. Lannoo (ed.), Status and
Conservation of Midwestern Amphibians, pp. 369–78. University of Iowa Press, Iowa
City, IO.
Bridges, A. S. and Dorcas, M. E. (2000). Temporal variation in anuran calling behavior:
implications for surveys and monitoring programs. Copeia, 2000, 587–92.
Corn, P. S. and Muths, E. (2002). Variable breeding phenology affects the exposure of
amphibian embryos to ultraviolet radiation. Ecology, 83, 2958–63.
Corn, P. S., Muths, E., and Iko, W. M. (2000). A comparison in Colorado of three methods
to monitor breeding amphibians. Northwestern Naturalist, 81, 22–30.
de Solla, S. R., Shirose, L. J., Fernie, K. J., Barrett, G. C., Brousseau, C. S., and Bishop, C. A.
(2005). Effect of sampling effort and species detectability on volunteer based anuran moni-
toring programs. Biological Conservation, 121, 585–94.
Dorazio, R. M. (2007). On the choice of statistical models for estimating occurrence and
extinction from animal surveys. Ecology, 88, 2773–82.
Dorcas, M. E. and Foltz, K. D. (1991). Environmental effects on anuran advertisement
calling. American Zoologist, 31, 3111A.
Droege, S. and Eagle, P. (2005). Evaluating calling surveys. In M. Lannoo (ed.), Amphibian
Declines: The Conservation Status of United States Species, pp. 314–19. University of
Duellman, W. E. and Trueb, L. (1986). Biology of Amphibians. McGraw-Hill Publishing
Company, New York.
Genet, K. S. and Sargent, L. G. ( 2003). Evaluation of methods and data quality from a
volunteer-based amphibian call survey. Wildlife Society Bulletin, 31, 703–14.
Gooch, M. M., Heupel, A. H., Price, S. J., and Dorcas, M. E. (2006). The effects of survey
protocol on detection probabilities and site occupancy estimates of summer breeding
anurans. Applied Herpetology, 3, 129–42.
Inkley, D. B. (2006). Final report assessment of utility of Frogwatch USA data
1998–2005. www.nwf.org/frogwatchUSA/FinalRptFW-ScienceAssess.pdf.
National Wildlife Federation for U.S. Geological Survey, Washington DC.
Johnson, D. H. and Batie, R. D. (2001). Surveys of calling amphibians in North Dakota.
The Prairie Naturalist, 33, 227–47.
Kaiser, K. (2008). Evaluation of a long-term amphibian monitoring protocol in Central
America. Journal of Herpetology, 42, 104–10.
Kirlin, M., Gooch, M. M., Price, S. J., and Dorcas, M. E. (2006). Predictors of winter
anuran calling activity in the North Carolina Piedmont. Journal of the North Carolina
Academy of Science, 122, 10–18.
Knutson, M. G., Sauer, J. R., Olsen, D. A., Mossman, M. J., Hemesath, L. M., and
Lannoo, M. J. (1999). Effects of landscape composition and wetland fragmentation
on frog and toad abundance and species richness in Iowa and Wisconsin, U.S.A.
Littlejohn, M. J. and Martin, M. M. (1969). Acoustic interaction between two species of
leptodactylid frog. Animal Behaviour, 17, 785–91.
Lotz, A. and Allen, C. R. (2007). Observer bias in anuran call surveys. Journal of Wildlife
MacKenzie, D. L., Nichols, J. D., Lachman, G. B., Droege, S., Royle, J. A., and Langtimm,
C. A. (2002). Estimating site occupancy rates when detection probabilities are less than
one. Ecology, 83, 2248–55.
Occurrence. Academic Press, San Diego, CA.
Marsh, D. M. and Trenham, P. C. (2008). Current trends in plant and animal population
monitoring. Conservation Biology, 22, 647–55.
Martof, B. S. (1953). Territoriality in the green frog, Rana clamitans. Ecology, 34, 165–74.
Mohr, J. R. and Dorcas, M. E. (1999). A comparison of anuran calling patterns at two
Carolina bays in South Carolina. Journal of the Elisha Mitchell Scientific Society, 115,
63–70.
Nelson, G. L. and Graves, B. M. (2004). Anuran population monitoring: comparison
of the North American Amphibian Monitoring Program’s calling index with Mark-
Recapture estimates for Rana clamitans. Journal of Herpetology, 38, 355–9.
Oseen, K. L. and Wassersug, R. J. (2002). Environmental factors influencing calling in
sympatric anurans. Oecologia, 133, 616–25.
Pellet, J. and Schmidt, B. R. (2005). Monitoring distributions using call surveys: esti-
mating site occupancy, detection probabilities and inferring absence. Biological
Penman, T. D., Lemckert, F. L., and Mahony, M. J. (2005). A cost-benefit analysis of
automated call recorders. Applied Herpetology, 2, 389–400.
Peterson, C. R. and Dorcas, M. E. (1992). The use of automated data acquisition techniques
in monitoring amphibian and reptile populations. In D. McCullough and R. Barrett
(eds), Wildlife 2001: Populations, pp. 369–78. Elsevier Applied Science, London.
Peterson, C. R. and Dorcas, M. E. (1994). Automated data acquisition. In W. Heyer,
R. McDairmid, M. Donnelly, and L. Hayek (eds), Measuring and Monitoring Biological
Diversity. Standard Methods for Amphibians, pp. 47–57. Smithsonian Institution Press,
Washington DC.
Pierce, B. A. and Gutzwiller, K. J. (2004). Auditory sampling of frogs: detection effi-
ciency in relation to survey duration. Journal of Herpetology, 38, 495–500.
Pierce, B. A. and Gutzwiller, K. J. (2007). Interobserver variation in frog call surveys.
Platz, J. E. (1993). Rana subaquavocalis, a remarkable new species of Leopard Frog
(Rana pipiens complex) from southeastern Arizona that calls under water. Journal of
Royle, J. A. and Nichols, J. D. (2003). Estimating abundance from repeated presence-
absence data or point counts. Ecology, 84, 777–90.
Saenz, D., Fitzgerald, L. A., Baum, K. A., and Conner, R. N. (2006). Abiotic correl-
ates of anuran calling phenology: the importance of rain, temperature and season.
Herpetological Monographs, 20, 64–82.
Schmidt, B. R. (2005). Monitoring the distribution of pond-breeding amphibians
when species are detected imperfectly. Aquatic Conservation: Marine and Freshwater
Ecosystems, 15, 681–92.
Scott, W. A., Pithart, D., and Adamson, J. K. (2008). Long-term United Kingdom
trends in the breeding phenology of the common frog, Rana temporaria. Journal of
Shirose, L. J., Bishop, C. A., Green, D. M., MacDonald, C. J., Brooks, R. J., and Helferty,
N. J. (1997). Validation tests of an amphibian call count survey technique in Ontario,
Canada. Herpetologica, 53, 312–20.
Steelman, C. K. and Dorcas, M. E. (in press). Anuran calling survey optimization:
developing and testing predictive models of anuran calling activity. Journal of
Herpetology.
Stevens, C. E. and Paszakowski, C. A. (2004). Using chorus-size ranks from call sur-
veys to estimate reproductive activity of the wood frog (Rana sylvatica). Journal of
Sun, J. W. C. and Narins, P. M. (2005). Anthropogenic sounds differentially affect
amphibian call rate. Biological Conservation, 121, 419–27.
Todd, M. J., Cocklin, R. R., and Dorcas, M. E. (2003). Temporal and spatial variation in
anuran calling activity in the western Piedmont of North Carolina. Journal of the North
Carolina Academy of Science, 119, 103–10.
Villanueva-Rivera, L. J. (2007). Digital recorders increase detection of Eleutherodactylus
frogs. Herpetological Review, 38, 59–63.
Walker, S. J. (2002). Frog Census 2001: Community monitoring of water quality and habitat
condition in South Australia using frogs as indicators. Environment Protection Authority,
Adelaide.
Weeber, R. C. and Vallitanios, M. (2000). The Marsh Monitoring Program 1995–1999:
Monitoring Great Lakes wetlands and their amphibian and bird inhabitants. www.
bsc-eoc.org/mmpreport.html. Bird Studies Canada with the U.S. Environmental
Protection Agency, Ontario.
Weir, L. A. and Mossman, M. J. (2005). North American amphibian monitoring pro-
gram (NAAMP). In M. Lannoo (ed.), Amphibian Declines: The Conservation Status of
United States Species, pp. 307–13. University of California Press, Berkeley, CA.
Weir, L. A., Royle, J. A., Nanjappa, P., and Jung, R. E. (2005). Modeling anuran detection
Woolbright, L. L. (1985). Patterns of nocturnal movement and calling by the tropical
frog, Eleutherodactylus coqui. Herpetologica, 41, 1–9.
Wright, A. H. and Wright, A. A. (1949). Handbook of Frogs and Toads of the United States
and Canada. 3rd edn. Comstock Publishing Associates, Ithaca, NY.
17
Measuring habitat
Kimberly J. Babbitt, Jessica S. Veysey, and George W. Tanner
17.1 Introduction
Understanding wildlife–habitat relations is fundamental to sound management
and conservation and provides important basic information on species’ ecology.
Obtaining accurate information on amphibian habitat associations can be chal-
lenging as many species have complex life cycles requiring very different habitat
types during different life-history stages, and many species are highly cryptic
during much of the year. However, numerous amphibian species are experien-
cing significant threats from habitat loss and degradation. Better information
on habitat requirements and habitat use is necessary to understand how these
changes affect habitat suitability for different species and to improve manage-
ment targeted at enhancing amphibian habitat.
17.2 Habitat selection

Although this chapter is mainly concerned with measuring habitat, it is import-
ant to keep in mind that how an animal “selects” habitat is a complicated pro-
cess involving multiple ecological and behavioral factors. Habitat selection also
occurs at a variety of scales (Johnson 1980). The broadest scale (termed first-
order selection) involves selection that determines the geographic distribution
of a species. Second-order selection occurs at the level of general features in the
landscape and dictates home range. Third-order selection is finer-scale selection
for specific sites within a home range. Most habitat assessments focus on second-
and third-order selection.
It is also important to understand differences between habitat selection, pref-
erence, and use. Habitat selection is defined as the process whereby an individ-
ual chooses a habitat among available alternatives. Habitat preference occurs
when habitats are selected independent of availability. However, information
on preference is difficult to obtain as certain conditions, such as equal access

to resources provided on an equal basis, are a prerequisite for determining pref-
erence. Thus, most field research involving habitat assessments measure use or
selection. It is assumed that animals select habitats that confer fitness advan-
tages (i.e. increased survival or reproduction); however, only rarely are meas-
ures of fitness actually conducted along with habitat assessments. It is tempting,
but not correct, to assume that habitat use implies selection. Garshelis (2000)
provides a review of the problems inherent in making such an interpretation.
Over-interpretation of habitat-use information can lead to erroneous assump-
tions about habitat selection, and thus improper recommendations regarding
habitat management.
17.3 Spatial and temporal scale

The scale at which one examines habitat use can greatly influence the inferences
drawn from the data. Features that have a significant relationship with habitat
use at one scale may be of little importance at another scale (Wiens 1989; Levin
1992). It is important to collect data at a scale appropriate for the research ques-
tion being asked. Although certain research questions may address only one
scale, increased insight into habitat relations can be gained from examination
of multiple scales (e.g. Welsh and Lind 2002). However, multi-scale approaches
may be prohibitively costly. Spatial and temporal patterns of resources avail-
ability and abundance also influence habitat use and suitability and should be
incorporated into study design (Southwood 1977). Year-to-year or seasonal
changes in habitat use can reflect the availability and distribution of resources as
modified by biotic (e.g. predators) and abiotic (e.g. weather) factors. Behaviors
inherent in many species (e.g. seasonal breeding migration, over-wintering, esti-
vation) and/or highly variable population dynamics (Pechman et al. 1991) also
influence habitat use and detectability of individuals. Knowledge of natural
history helps to guide appropriate timing for sampling.
17.4 Approaches for examining habitat selection

In this section we provide a brief introduction into the most common approaches
for examining habitat selection. A detailed review of habitat selection theory,
design, and analysis is beyond the scope of this chapter. More detailed infor-
mation, reviews of study designs and their inherent strengths and weaknesses,
and suggested statistical approaches can be found in Thomas and Taylor (1990,
2006), Garshelis (2000), Manly et al. (2002), and Morrison et al. (2006).
17 Measuring habitat | 301
Use-availability designs are the most common approaches for examining habi-
tat selection (Thomas and Taylor 1990, 2006; Garshelis 2000). These approaches
compare the proportional use of each habitat type by an organism to the relative
area of each habitat. Use-availability studies can be classified into four general
designs (Thomas and Taylor 1990, 2006). In design 1, use is determined for all
known individuals in a population but the individuals are not identified. Habitat
is considered equally available to all individuals. This design is an assessment of
habitat use at the population level and does not allow examination of variation
in habitat selection among individuals. This design would be appropriate for
studies where individuals are documented using visual encounter surveys or area-
constrained searches but where no individuals are marked.
Design 2 is used when individuals are uniquely marked and, as in design 1,
habitat availability is considered equivalent for all individuals. In design 3, use
is determined for individually marked organisms but, unlike designs 1 and 2,
available habitat is estimated for each individual animal. This design is appro-
priate when, for example, individual home ranges have been estimated and only
areas within each home range are deemed available. Design 4 examines use of
habitat by uniquely marked individuals at multiple time periods (e.g. each time
an animal is located) and pairs measurements of used habitat with measure-
ments of available habitat taken at each interval. This design is particularly use-
ful for examination of small-scale habitat (i.e. microhabitat).
Use-availability studies generally focus on habitat selection based on broad
habitat types, or multiple habitat parameters that are analyzed individually
(Garshelis 2000). An alternative approach to use-availability studies is the site-
attribute approach. This approach differs from use-availability studies in that
the focus is on measuring habitat features that potentially influence use in areas
where an individual has been documented and comparing those values to meas-
urements made on the same features in areas that are not used. Differences
between used and non-used areas are then assumed to reveal which features are
responsible for selection for (or against) habitat use. Thus, the focus is not on the
amount of use in one habitat compared to another, but rather whether the site
was used or not and what specific features appear to determine use. Often, site-
attribute studies are used to examine habitat variables at biologically important
sites (e.g. breeding sites).
17.5 Determining availability

One of the most difficult aspects of examining habitat selection is determining
habitat availability. Because we do not share the same perspective as the organisms
we study, it is certain that some working assumptions will go into any determin-
ation of availability. At the broadest level, areas beyond the study area are not
examined, and so by default are unavailable. If we delineate the home range of
an organism we can define unavailable habitat as that outside the home-range
boundary. However, within a home range, intraspecific interactions may mean
that a portion of the home range is not just unused but unavailable. In contrast,
some habitat may go unused during much of the year but yet be both available
and critical (e.g. over-wintering sites, breeding sites). Knowledge of the organism’s
natural history can help to address some questions about what is unavailable and
what is unused, but cautious interpretation is always warranted because improper
designation of habitat availability can bias results.
17.6 What to measure

The list of variables one could measure is potentially quite long, but sampling
is usually limited by time, personnel, and fiscal constraints. Previously pub-
lished studies can provide initial guidance as to key parameters that should
be measured. However, the specific research question may lead to selection of
parameters that are not often measured. Further, because the importance of a
habitat parameter may vary with scale, measurement of variables that have not
previously appeared important may need to be measured when working at a dif-
ferent scale. The same can be argued when working in different areas to allow
for regional comparisons. A reasonable compromise between measuring every-
thing one can think of versus a small number of variables that appear most crit-
ical is to conduct a pilot study using a broader set of variables and then to reduce
the number of variables by eliminating those that appear least important, or
those that are correlated with one another. Because resource availability, abun-
dance, and importance can change temporally, it is important to keep in mind
that a pilot study conducted in one season may not provide useful guidance for
selecting parameters in a different season.
17.7 Weather variables

As ectothermic organisms with semipermeable skin, amphibians are very sen-
sitive to changes in abiotic conditions. This fact has two important influences
on how habitat assessments should be conducted. First, abiotic habitat meas-
ures are critically important for understanding habitat use and selection because
these features are often the key factors influencing habitat–amphibian associ-
ations. Second, study timing and location must be carefully selected because the
abiotic environment can greatly influence levels of amphibian activity, thereby

affecting our interpretation of habitat use. For example, movement patterns of
pond-breeding amphibians are often tied to seasonal temperature and rainfall
conditions. Thus, for many species assessment of breeding habitat use must be
timed to match seasonal weather conditions rather than the calendar per se. For
fossorial species, rainfall can trigger movement or surface activity, and interpret-
ation of habitat use may vary greatly with precipitation patterns (Taub 1961).
It would be hard to imagine a research project that would not require meas-
urement of basic weather data such as temperature and rainfall. Ambient air
temperature can be measured with a standard handheld thermometer. This low-
tech approach provides only a spot check of temperature but should be measured
as it provides a record of specific conditions at the time and point of sampling.
Temperature probes are available for water and soil temperature measurements.
Most water-based meters measure temperature along with other parameters.
Probes are useful for specific measures in space or time, but are not the best
approach for providing integrative or frequent measurements. To provide a
comprehensive assessment of abiotic and climate conditions, most projects will
require additional temperature observations. Min/max thermometers, which
can be used on land or in water, provide a relatively inexpensive way to obtain
the range of temperature conditions an organism may experience, and measure-
ments can be taken as frequently as one wants (e.g. over a 24 h period, over a
week). If more detailed temperature conditions are desired, dataloggers should
be used as these devices can be programmed to measure temperature at specific
intervals throughout the day.
Rainfall data are easily collected with rainfall gauges. Gauge types range from
simple plastic devices that are manually emptied at specified times intervals to
electronic devices such as tipping buckets that record rainfall amounts on data
sheets, on data drums, or directly to a computer and then empty themselves.
Both collection frequency (i.e. hourly, daily, weekly) and gauge placement influ-
ence the data collected. For simple measurements of rainfall amounts, gauges
should be located in open areas. However, to measure the amount of precipita-
tion that reaches the forest floor (i.e. rain throughfall), rain gauges placed under
the canopy provide more relevant data.
Relative humidity (and temperature) influences water loss rates and therefore
affects both amphibian activity and habitat suitability. Similarly, wind currents
across the skin surface affect water loss. Differences in habitat structure and
aspect among sites, even at a local level, can influence these parameters making
them potentially important to measure. Relative humidity can be measured
using a sling psychrometer or a hygrometer. An anemometer is used to measure
wind speed, and handheld versions are very inexpensive. Wind speed can also be
estimated using a Beaufort scale of winds, which provides descriptive measures
of wind effects that are given a number from 1 to 12. Wind direction can be
determined at the same time as wind velocity by taking a compass reading in the
direction of wind flow.
All the weather parameters mentioned can be measured using relatively inex-
pensive equipment or electronic data devices that require more money but less
personnel time. Further, meteorological stations are maintained throughout
many areas of the world and may provide free access to data. Availability may
differ with geographic location, but researchers should check into availability
before purchasing expensive equipment. These stations are an excellent source
of climate information, but the utility of station data depends on the scale of
the research project. If the purpose of an investigation is to understand the role
of local habitat relations then weather-station data will likely be insufficient for
examining among-site differences.
17.8 Aquatic habitat

Aquatic amphibian habitat includes lentic (i.e. pond or wetland) and lotic (i.e.
stream) environments. Spatial and temporal variability, and systemic connec-
tions between aquatic habitat and the surrounding terrestrial landscape, are key
features of aquatic habitats that structure amphibian communities. Aquatic sites
constitute obligate habitat for some species, and are critical breeding habitat for
species with complex life cycles involving aquatic egg or larval development.
Because many amphibian species undergo ontogenetic shifts in habitat, gener-
ally from aquatic to terrestrial habitat, different foci and approaches are needed
to understand habitat use for each life-history stage (i.e. eggs, larvae, juveniles,
adults).
17.9 Physical habitat variables

Physical features of aquatic habitats provide a dynamic structure within which
amphibians interact with their biotic and abiotic environments. Those physical
habitat parameters that prescribe broadly whether aquatic habitats are suitable
for amphibians are discussed in detail below. Many other physical variables
influence the finer-scale suitability of aquatic habitat. Some of the more com-
monly measured of these are presented in Table 17.1. Welsh et al. (1997) and
Welsh and Ollivier (1998) provide an extensive list of relevant stream habitat
variables.
Table 17.1 Additional aquatic variables for assessing amphibian habitat.
Habitat variable Relevance to amphibians Measurement technique
Ultraviolet B radiation Associated with altered growth, Polysulfone plastic dosimeter

survival, and behavior Pyranometer+datalogger
Interacts synergistically with other variables Spectrometer
Sun exposure Through effects on water temperature, Light at surface
evaporation, and plant productivity, Pyranometer
influences amphibian development rate; Pyroheliometer
and oxygen, food, and water availability Quantum sensor
Light at depth
Submersible spectroradiometer or comparison
between underwater and surface photocells
See sections 17.11.1–17.11.3
Turbidity Alters water transparency, with impacts on predator and amphibian behavior Secchi disk
Related to sedimentation and pollution loads, which can affect fitness Nephelometer
Turbidity meter
Stream order Related to current intensity, water temperature, and microhabitat suitability Determined via examination of maps or GIS
See section 17.9
Stream reach type Flow regime and substrate interact to create particular reach types Classify and map reach types (see Welsh et al. 1997
(e.g. riffle, pool, run) Amphibian species may be closely associated with particular reach types for review of reach classification systems)
Water volume (e.g. of bromeliads, Related to reproductive potential within a microhabitat Remove water from microhabitat and quantify
tree holes and other microhabitats) (e.g. in syringe or graduated cylinder), return
water to microhabitat
Substrate depth Related to the refuge-potential of the substrate Ruler; meter stick; or long, graduated pole
Habitat age and type (i.e. natural, Related to vegetative and substrate development, Examine chronological series of topographic
artificial, or restored) (also applies to microhabitat quality, and the likelihood of maps or aerial photos
terrestrial habitats) amphibian use, immigration, and survival Examine historical records or management reports
Degree of human disturbance (in and Associated with sedimentation, pollution, Categorize according to land-use type and intensity
adjacent to the habitat) invasive species, and predators
GIS, Geographic Information Systems.
Morphometry, the habitat’s basic shape and structure, limits the community
types possible in a habitat, through its influence on water temperature, current
strength, sediment and vegetation patterns, water-holding capacity, substrate,
and weathering rates. Morphometric variables include contours, perimeter,
length, width, area, depth, slope, and surficial geology. For streams, morph-
ometry can be assessed for the wetted or dry channel, or for individual habi-
tat reaches. Channel slope is particularly important for streams because slope
affects dispersal potential and flow velocity.
Many morphometric variables can be mapped using global positioning sys-
tem (GPS) or professional survey equipment. Perimeter, length, and width can
also be assessed with pacing or a measuring tape. Wet contours and water depth
can be charted with an echo sounder, weighted line, or meter stick. Area can be
calculated from length and width; or from maps, aerial photographs, or GIS lay-
ers. The reliability of map-derived area measurements increases with basin size
(Skidds et al. 2007). Slope can be measured with a clinometer. Surficial geology
can be determined from geologic maps or field-verified from test or gravel pits.
Hydroperiod, or the timing and duration that a habitat holds water during a
given year, strongly influences community composition (Wellborn et al. 1996;
Babbitt et al. 2003). For example, in ephemeral waterbodies, invertebrates are
often the dominant predators and amphibians must have a short larval stage
to complete metamorphosis before the habitat dries. Conversely, in perman-
ent waterbodies, amphibians can over-winter as larvae and fish are typically
major predators. Hydroperiod can be categorized for lentic habitats as tempor-
ary, semi-permanent, or permanent; and for lotic habitats as intermittent or
perennial. Such categorizations can be made by examining vegetation, or water
level and substrate, towards the end of the dry season. Detailed data can be
obtained by recording changes in the extent of inundation (Skidds et al. 2007),
or by tracking water depth and noting the drying date. Hydroperiod can vary
widely from year to year in seasonal waterbodies, so hydroperiod data should be
collected over several years.
Hydrologic connectivity (i.e. the extent to which aquatic environments are
linked) influences both hydroperiod and the inputs to, and outputs from, a
habitat. Connectivity may be observed by inspection for inflows and outflows
during the rainy season, or detected from contour maps or historic aerial photo-
graphs. Monitoring wells or a hydrologic analysis may be necessary to establish
groundwater connectivity.
For streams, flow intensity (e.g. velocity, variability) also strongly influences
habitat suitability. Flow affects dispersal, predator distribution, scouring, and
sedimentation, and the ability of amphibians, and their sperm (i.e. for external
fertilization) and food (i.e. invertebrates and algae), to maintain channel pos-
ition. Velocity is measured with a flow meter, a pitot tube, or by calculating the
time required for a buoyant object or dye to travel a specified distance. Discharge,
or flow volume per unit time, is equal to the stream’s cross-sectional area times
the velocity. Variability can be described by flow variance per unit time, or by the
return interval of specific events (e.g. 100-year floods). Flow timing determines
desiccation and scouring risks, and can be understood by examining velocity and
discharge records.
Water temperature impacts amphibians at multiple ecological scales. For
instance, temperature affects decomposition and photosynthesis, which influ-
ence food availability and thus amphibian abundance. Amphibian development
rates also vary with water temperature. Temperature measurement is discussed
in section 17.7. Drainage area, elevation and aspect, and canopy closure all affect
water temperature. Drainage area and elevation can be determined from topo-
graphic maps. Elevation can also be measured with an altimeter. Canopy clos-
ure is discussed in section 17.11.
Aquatic substrates can be organic or inorganic. Organic substrates provide
over-wintering habitat and are integral to aquatic food webs. Inorganic sub-
strates (e.g. cobbles) provide oviposition sites and refuges for amphibians and
their invertebrate and algal food. Substrate particle size is a critical stream-
habitat variable, as particle size affects size of interstitial hiding places which
affects susceptibility to predation (Barr and Babbitt 2002). Small particles (i.e.
32 mm) can be quantitatively described via sieve analysis. Visual estimates of

percent cover of each substrate type and size class (ranging from leaf litter and
muck, to fine sediment up to boulders) can also be used to classify substrate
composition and particle size (Welsh and Ollivier 1998).
17.10 Chemical variables

The chemical makeup of aquatic sites can significantly affect amphibian
growth, survival, and behavior. Pollutants can be toxic or can act as endocrine
disrupters and alter many life processes (e.g. sexual development; Orton et al.
2006). Therefore, most aquatic habitat studies include measurements of some
chemical parameters. Chemical potency varies with a wide variety of factors,
ranging from chemical formulation (e.g. Howe et al. 2004) to amphibian life
stage (e.g. Griffis-Kyle 2005). Exposure to some chemicals is greatest in the
water column, while other substances accrue in the substrate (e.g. Jofre and
Karasov 1999). Finally, synergistic effects between chemicals and other abi-
otic or biotic variables are common (e.g. Relyea and Mills 2001). Therefore,
it is important to consider context (e.g. site history) when planning chemical

sampling.
Given the inherent variability of aquatic environments, a composite sample,
drawn from multiple points in a habitat, is necessary for the sample to be repre-
sentative of general water conditions. To characterize specific sites (e.g. riffles),
however, single samples should be collected from point locations. While some
variables (e.g. pH, conductivity) can be measured in the field, many chemicals
require laboratory analysis. Sampling protocols can vary significantly between
chemical variants, so protocols should be researched or laboratory personnel
contacted prior to sampling. In particular, it is important to know whether and
when a sample should be filtered or cooled; sample bottle type; and how quickly
following collection a sample must be analyzed. Sampling methods for some
commonly measured parameters are discussed below.
pH is a measure of water acidity. pH can be assessed with a pH meter or paper
pH strips. Strips are less accurate than pH meters, however. Alkalinity, a meas-
ure of the site’s ability to neutralize acids, is also often sampled, since it reflects
the habitat’s resistance to sudden changes in pH. Alkalinity is measured via
titration with a strong acid.
Dissolved oxygen is required for respiration by aquatic amphibian life stages.
Dissolved oxygen can be a strongly limiting variable for cold-water-adapted spe-
cies (Jackson 2007). Dissolved oxygen also affects the redox potential and solu-
bility of some metals and nutrients. Dissolved oxygen is usually measured in
the field, using either a variation of the Winkler titration method or an oxygen
meter. Samples can be titrated in the laboratory, but only if properly preserved
while still in the field.
Conductivity, a measure of the water’s ionic strength, is an easily sampled vari-
able that can be used as a proxy for parameters that are more difficult to assess.
Ions occur naturally in aquatic habitats as a result of weathering and nutrient
fluxes, but are also produced by human activities (e.g. winter road salting). Very
high conductivity levels, or significant changes in conductivity through time,
suggest poor water quality and are cause for further analysis of more-specific
chemical variables. Conductivity can be measured with a conductivity meter.
Nutrients influence amphibians through two main pathways: nutrients
are necessary for growth and survival of, and can be toxic to, most life forms.
Amphibians seem particularly sensitive to nitrogen, especially in the forms of
nitrate, nitrite, and ammonia. Commonly sampled nutrients include: NH4-N,
NH3-N, NO3-N, NO2-N, total Kjeldahl N, dissolved organic N, total ortho-
PO4, total P, and dissolved P. Dissolved organic carbon, which can attenuate
ultraviolet radiation, is also of interest (e.g. Brooks et al. 2005). Other humic
substances can also strongly impact amphibian biochemistry and communi-

ties (e.g. reduced plant photosynthesis), and should be considered for analysis
(Steinberg et al. 2006). Important cationic micronutrients (e.g. calcium, mag-
nesium) may also be analyzed.
Aluminum, heavy metals, polycyclic aromatic hydrocarbons (PAHs), poly-
cyclic chlorinated hydrocarbons (PCHs), polychlorinated biphenyls (PCBs),
and a variety of pesticides are potentially important chemicals for measurement
as they have demonstrated lethal and sublethal impacts to amphibians and can
significantly alter habitat suitability. Nutrients, dissolved organic carbon, cati-
ons, metals, pesticides, hydrocarbons, and other chemical samples are usually
analyzed in the laboratory. Use of nationally certified laboratories is recom-
mended for many analyses.
17.11 Measuring vegetation

Vegetation is a major component of most terrestrial habitat. Vegetation strongly
influences habitat suitability through its effects on microclimate. Vegetation
can affect thermal gradients by providing shade, impacting relative humidity
and moisture retention, adding litter (from leaves to large woody debris) to the
ground stratum, and altering wind flows and snow accumulation. Vegetation,
as living individuals, snags, and down logs, also contributes to habitat suitability
by providing amphibians foraging substrates, protective cover from predators,
and structures for egg (leaves) and larval (bromeliads) development.
In aquatic habitats, living and decaying plants affect the productivity, and
structural and chemical suitability of amphibian habitat. Vegetation provides
refuge, oviposition, calling, nesting, and over-wintering sites. Vegetation can
change the abiotic environment (e.g. water temperature), with effects on amphib-
ian development and fitness. Aquatic, shoreline, and proximate terrestrial vege-
tation can all influence different aspects of aquatic amphibian habitat.
The techniques discussed below apply to terrestrial and aquatic environ-
ments, though some techniques must be modified for aquatic sites. In particu-
lar, for aquatic assessments of mid- and understory vegetation, emergent and
submergent plants are often distinguished and branch density, which reflects
oviposition site potential, may be of interest. Further, any examined micro-
habitat parameters will be habitat-specific. For example, four-toed salamanders
(Hemidactylium scutatum) often brood nests within wetland-plant hummocks.
For this species, it might be useful to know not only the abundance, distribu-
tion, and species composition, but also the depth, moisture content, and height,
of each plant hummock (Wahl et al. 2008).
In the following sections we discuss measurement of overstory, midstory, and

ground-stratum parameters. Many techniques can be used to quantify more
than one of these categories. In some ecosystems, it may also be important to
measure characteristics of vinous and epiphytic vegetation (see Bandoni de
Oliveira and Navas 2004).
Vegetative parameters can be measured using quadrats, line-intercept, or
plotless techniques. Quadrat shapes vary from circular, to square, to rectangu-
lar. Circular plots have less edge and can be permanently marked with one point.
Depending on research objectives, quadrats may be placed within or across a
moisture or elevation gradient.
Plant distribution often is spatially structured or patchy (Levin 1992) lead-
ing to a condition known as spatial autocorrelation. This condition complicates
the measurement and statistical descriptions of species diversity; however, field
sampling designs and statistical analyses have been developed to address this
condition (e.g. Goslee 2006).
17.11.1 Tree (overstory) measurements

Parameters typically quantified for the tree stratum include species composition
(i.e. richness and frequency of occurrence), density, total height, crown depth,
diameter, and basal area. Since species richness and frequency are a function of
presence or absence within sampling units, these parameters often are partial
components of an overall sampling protocol that may be directed at determin-
ing plant density and canopy cover. Species composition and absolute density
(number of trees/area) can be measured using quadrat or distance (plotless)
techniques.
Appropriate quadrat size will vary inversely with the natural abundance of
individuals in the sampled community. Conversely, adequate quadrat size will
vary directly with increasing species richness following species/area relation-
ships. Species richness may be described in terms of number of species/unit
area or the total number of species encountered/area sampled. Frequency of
occurrence is estimated by the proportion of sampling units within which a
given species was found and describes how common a species is within the plant
community. Density is estimated as the number of trees (either summed over all
species or by individual species)/unit area.
While several distance measurement techniques exist, the point-center-quarter
(PCQ) technique is commonly used (Higgins et al. 2005). This technique
works best when describing structural parameters of just one or two species
(Cottam and Curtis 1956). The PCQ method provides acceptable estimates of
density of randomly dispersed trees. Variants of this method, either the angle-
order (Morista 1957) or the corrected-point-distance (Batcheler 1973), are rec-
ommended for use when non-random distributions of individuals occur.
Tree height, up to 10 m, can be measured using a telescoping stage stadia rod
elevated from the base of the tree to the highest point of the canopy crown. A
manual or automated clinometer can be used to measure the height of any tree
for which the base of the stem and top of the canopy (apex) can be seen. Stadia
rods and clinometers also can be used to measure height to base of canopy and
thus compute canopy thickness.
Tree canopy cover is defined as the vertical projection of the total areal extent
of the canopy upon the ground and is typically expressed as m2/ha or more com-
monly as a percentage. Changes in canopy cover can alter suitability of aquatic
sites (e.g. via effects on temperature and hydroperiod), resulting in changes in
amphibian community composition (Skelly et al. 1999). Light interception and
irradiance at a point on the ground, however, are influenced by both the vertical
and angular aspects of the canopy (Nuttle 1987). Several ground-based methods
have been developed to measure tree canopy cover including, but not limited to,
line-interception, moosehorn, spherical densitometer, and hemispherical pho-
tography (Fiala et al. 2006). The line-intercept method employs multiple line
transects (each 15–20 m long) distributed within a stand. Tree canopy cover is
estimated as the percentage of each line lying beneath the vertical projection of
the canopy disregarding any breaks within the outline of the canopy.
The moosehorn and the spherical densitometer are both sighting devices that
employ mirrors with etched markings. The moosehorn has a flat mirror to reflect
the canopy directly above the sampling location onto the reflective grid. Percent
canopy cover is determined by counting the number of cross-hair intersections
or points intercepting the reflected canopy divided by the total number of points
on the grid. The spherical densitometer uses a convex or concave spherical mir-
ror. Canopy measurements are taken while facing in all four cardinal directions
from a single point. Canopy cover is estimated by averaging the proportion of
the 24 squares etched on the reflective spherical surface that is contacted by can-
opy. A simpler approach to measuring canopy cover is to use a tube and ocularly
estimate percent canopy in broader categories (e.g. 0, 5, 25, 50, 75, 100%).
Horizontal photographs, when taken at a designated height with a hemi-
spherical wide-angle lens, record the extent of the overhead canopy dependent
upon the angle of view of the lens. Photographs taken at bright midday condi-
tions should be avoided (Fiala et al. 2006). Photogrammetric data obtained
using hemispherical lenses are analyzed following Rich (1989). These data are
an estimate of light penetration through angular openings in the canopy and

are not direct measurements of canopy cover (Fiala et al. 2006). From above the
canopy, remotely-sensed data using recently developed laser altimetry (lidar)
provides three-dimensional geospatially-referenced data that is useful in model-
ing structural attributes of terrestrial ecosystem, especially at the landscape level
(Vierling et al. 2008).
The standard height above ground to measure tree diameter is 1.37 m and is
called diameter at breast height (dbh). If a tree has a fluted or buttressed base, the
diameter is measured immediately above this area. Flexible diameter tapes (that
transform circumference to diameter) or calipers are typically used to measure
dbh. Forest basal area is the proportion of ground occupied by the bases of trees,
expressed as m2/ha. Clear plastic prisms with known degrees of light refraction,
called Basal Area Fraction (BAF), can be used to view and count the num-
ber of individual trees within a complete circle around a sampling point whose
stems, when viewed through the prism, are not completely separated (Salek and
Zahradnik 2008). This count is then multiplied by the prism’s BAF (units are in
multiples of m2/ha) to estimate basal area in the vicinity of that sampling point.
Individual trees whose stem displacements just touch are counted as one half.
Amber-colored prisms aid in sighting under low-light conditions.
17.11.2 Shrub (midstory) measurements

The shrub or midstory stratum is composed mostly of woody plants that may
never enter the overstory or of juvenile individuals of species that will mature
into the overstory. Midstory structural and compositional parameters that may
interest those studying amphibians are similar to those of the overstory, namely
species composition, density, and cover.
Species composition can be estimated using quadrat and line-intercept methods
(see section 17.11.1). When using quadrats, density of midstory plants is deter-
mined by counting the number of individual stems within the quadrat. Many
shrubs and juvenile trees will produce multiple stems per plant (i.e. basal and
coppice sprouts). Therefore, midstory plant density is characterized as the num-
ber of stems per unit area without regard to clonal associations. Quadrat size for
estimating midstory plant density may range from 4 to 10 m2 (Irwin and Peak
1979).
Midstory canopy cover can be estimated using the line-intercept method
described for overstory canopy cover. The proportion of a fi xed-length line
transect intersected by the vertical projection of midstory canopy provides a
measure of percent canopy cover. Since the canopy of midstory (compared to
overstory) plants is closer to the observer, canopy cover of individual species
can be determined. Canopies of different species may overlap; therefore, total

midstory canopy cover may be less than the sum of the individual species’
covers and should be measured as a separate parameter. Measuring maximum
height of midstory plants at systematic or randomly located points along each
line transect will provide a description of midstory plant-height dynamics.
In environments where shrubs are low-growing (i.e. less than 1 m tall), can-
opy and basal cover can be sampled using the point-intercept method (Bonham
1989). Sharp metal pins are inserted vertically or on an incline through the
vegetation at random or systematically-placed locations. Canopy and basal
cover are calculated as the proportion of inserted pins that intercept, respect-
ively, any piece of vegetation (leaf or stem) or the base of a plant at ground level
(Higgins et al. 2005). Point intercepts are often recorded along a line transect;
however, note that the line is the sample unit. Measuring fewer points along a
greater number of lines improves statistical power (Bonham 1989).
17.11.3 Ground (understory) measurements

Understory components may be vegetative or non-vegetative (e.g. rocks, bare
ground), and can provide critical refuges for ground-dwelling amphibians.
Typically measured characteristics are plant density, frequency, and percent and
type of cover. Standing-crop biomass and moisture content may also be of inter-
est, especially in fire-prone ecosystems.
Density of understory plants can be estimated using direct counts within
quadrats. Quadrats typically vary in area from 0.5 to 1.0 m2 (Higgins et al.
2005). Horizontal cover of vegetative and non-vegetative components can be
assessed using ocular estimates within quadrats, intercepted distances along line
transects, or vertical digital photographs (Luscier et al. 2006). Ocular estimates
require observers to place estimated cover values into bracketed cover classes;
the midpoints of those classes are used for statistical analyses. Line-intercept
cover measurements are more precise and accurate but require more time to
collect. Vertical digital photographs of the understory also can provide accur-
ate estimates of object groups (e.g. grasses, forbs, litter, bare ground); however,
this technology cannot easily differentiate among plant species (Luscier et al.
2006).
Coarse woody debris (CWD) consists of dead trees and tree parts that are in
contact with the ground. Measurement of CWD is important in both terrestrial
and aquatic environments. Typically measured structural attributes of CWD
include areal cover and volume. CWD areal cover can be quantified using the
line-intercept method while measuring other understory parameters. The pro-
portion of lines intercepting the vertical projections of CWD can be presented as
percent cover or translated to m2/ha (Baillie et al. 1999). Using quadrats, CWD
volume (m3/ha) can be estimated from the density of downed logs and diameter
measurements at the large and small ends of each log. CWD decomposition stage
may be of interest and be classified according to Maser et al. (1979).
Litter depth and moisture content can greatly influence microhabitat suit-
ability. Litter depth to mineral soil is easily measured using a ruler. To estimate
biomass (g of dry weight/m2) and moisture content (%), litter can be collected
in the field, weighed, and oven dried to a constant weight (Bonham 1989). Dry
weight and moisture content of understory plants can be determined by clip-
ping the vegetation at ground level within quadrats and weighing, drying, and
re-weighing the vegetative samples.
17.12 Edaphic features

Many amphibian species spend a significant portion of their lives on or beneath
the soil surface. Chemical and structural features of the edaphic environment can
strongly influence habitat suitability. Soil pH, moisture, temperature, and texture
are important parameters to measure. Soil pH outside the physiological toler-
ance of amphibians can be an important limiting factor (Wyman and Hawksley-
Lescault 1987). Soil pH can be assessed by inserting a probe into an emulsion of
sieved, dried soil in a 1:10 solution of distilled water (Davey and Conyers 1988).
Soil moisture can be determined with a probe or gravimetrically, whereby the
sample is weighed wet, dried to a constant weight, and reweighed (also known as
weigh-dry-weigh). Soil thermometers allow measurement at specific depths. To
characterize retreat temperatures, it may be useful to measure temperature at the
soil surface and at greater depths. Texture can be assessed with a sieve analysis.
Availability of underground retreats is an important habitat component for many
fossorial amphibians. For example, mammal burrows provide important retreats
for ambystomatid salamanders (Madison 1997), whereas spider burrows provide
refuges for desert anurans (Dayton and Fitzgerald 2006). Burrow measurements
can include dimension (e.g. length and diameter), distribution, and whether bur-
rows are aligned vertically or horizontally.
17.13 Conclusion
Characterization of amphibian habitat has lagged behind many other taxa.
Increased research on habitat suitability will provide valuable data that will
assist in management and conservation of amphibian species. Habitat suitabil-
ity is influenced by both the features of the environment outlined in this chapter
as well as changes that occur at the landscape level. Researchers are encouraged
to design studies that include approaches outlined in this chapter as well as those
detailed in Chapter 20.
17.14 References
Babbitt, K. J., Baber, M. J. and Tarr, T. L. (2003). Patterns of larval amphibian distribution
along a wetlands hydroperiod gradient. Canadian Journal of Zoology, 81, 1539–52.
Baillie, B. R., Cummins, T. L., and Kimberley, M. O. (1999). Measuring woody debris in
the small streams of New Zealand’s pine plantations. New Zealand Journal of Marine
and Freshwater Research, 33, 87–97.
Bandoni de Oliveira, F. and Navas, C. A. (2004). Plant selection and seasonal patterns of
vocal activity in two populations of the bromeligen treefrog Scinax perpusillus (Anura,
Hylidae). Journal of Herpetology, 38, 331–9.
Barr, G. E. and Babbitt, K. J. (2002). Effects of biotic and abiotic factors on distribution
and abundance of larval two-lined salamanders (Eurycea bislineata): changes across
spatial scales. Oecologia, 133, 176–85.
Batcheler, C. L. (1973). Estimating density and dispersion from truncated or unre-
stricted joint point-distance nearest-neighbor distances. Proceeding of the New Zealand
Ecological Society, 20, 131–47.
Bonham, C. D. (1989). Measurements for Terrestrial Vegetation. John Wiley and Sons,
New York.
Brooks, P. D., O’Reilly, C. M., Diamond, S. A., Campbell, D. H., Knapp, R., Bradford, D.,
Corn, P. S., Hossack, B., and Tonnessen, K. (2005). Spatial and temporal variability in
the amount and source of dissolved organic carbon: implications for ultraviolet expos-
ure in amphibian habitats. Ecosystems, 8, 478–87.
Cottam, G. and Curtis, J. T. (1956). The use of distance measures in phytosociological
sampling. Ecology, 37, 451–60.
Davey, B. G. and Conyers, M. K. (1988). Determining the pH of acid soils. Soil Science,
146, 141–50.
Dayton, G. H. and Fitzgerald, L. E. (2006) Habitat suitability models for desert amphib-
ians. Biological Conservation, 132, 40–9.
Fiala, A. C. S., Garman, S. L., and Gray, A. N. (2006). Comparison of five cover estima-
tion techniques in the western Oregon Cascades. Forest Ecology and Management, 232,
188–97.
Garshelis, D. L. (2000). Delusions in habitat evaluation: measuring use, selection,
and importance. In L. Boitani and T. K. Fuller (eds), Research Techniques in Animal
Ecology: Controversies and Consequences, pp. 111–64. Columbia University Press,
New York.
Goslee, S. C. (2006). Behavior of vegetation sampling methods in the presence of spatial
autocorrelation. Plant Ecology, 187, 203–12.
Griffis-Kyle, K. L. (2005). Ontogenic delays in effects of nitrite exposure on tiger salaman-
ders (Ambystoma tigrinum tigrinum) and wood frogs (Rana sylvatica). Environmental
Toxicology and Chemistry, 24, 1523–7.
Higgins, K. F., Jenkins, K. J., Clambey, G. K., Uresk, D. W., Naugle, D. E., Norland, J. E.,
and Barker, W. T. (2005). Vegetation sampling and measurement. In C. E. Braun
(ed.), Techniques for Wildlife Investigations and Management, 6th edn, pp. 524–54. The
Wildlife Society, Bethesda, MD.
Howe, C. M., Berrill, M., Pauli, B. D., Helbing, C. C., Werry, K., and Veldhoen, N.
(2004). Toxicity of glyphosate-based pesticides to four North American frog species.
Irwin, L. L. and Peek, J. M. (1979). Shrub production and biomass trends following five
logging treatments within the cedar-hemlock zone of northern Idaho. Forest Science,
24, 415–26.
Jackson, D. C. (2007). Temperature and hypoxia in ectothermic tetrapods. Journal of
Thermal Biology, 32, 125–33.
Jofre, M. B. and Karasov, W. H. (1999). Direct effect of ammonia on three species of
North American anuran amphibians. Environmental Toxicology and Chemistry, 18,
1806–12.
Johnson, D. H. (1980). Comparison of usage and availability measurements for evaluat-
ing resource preference. Ecology, 61, 65–71.
Levin, S. A. (1992). The problem of pattern and scale in ecology. Ecology, 73, 1943–67.
Luscier, J. D., Thompson, W. L., Wilson, J. M., Gorham, B. E., and Dragut, L. D. (2006).
Using digital photographs and object-based image analysis to estimate percent ground
cover in vegetation plots. Frontiers in Ecology and Environment, 4, 408–13.
Madison, D. M. (1997). The emigration of radio-implanted spotted salamanders,
Ambystoma maculatum. Journal of Herpetology, 31, 542–51.
Manly, B. F. J., McDonald, L. L., Thomas, D. L., McDonald, T. L., and Erickson, W. P.
(2002). Resource Selection by animals: Statistical Design and Analysis for Field Studies,
2nd edn. Kluwer, New York.
Maser, C., Anderson, R. G., Cromack, Jr, K., Williams, J. T., and Martin, R. E. (1979).
Dead and down woody material. In J. W. Thomas (ed.), Wildlife Habitats in Managed
Forests—The Blue Mountains of Oregon and Washington, pp. 78–95. Agricultural
Handbook 553. US Department of Agriculture, Washington DC.
Morista, M. (1957). A new method for the estimation of density by the spacing method
applicable to non-randomly distributed populations. Physiological Ecology, 7, 134–44.
Morrison, M. L., Marcot, B. G., and Mannan, R. W. (2006). Wildlife-habitat Relationships:
Concepts and Applications, 3rd edn. Island Press, Washington DC.
Nuttle, T. (1997). Densiometer bias? Are we measuring the forest or the trees? Wildlife
Orton, F., Carr, J. A., and Handy, R. D. (2006). Effects of nitrate and atrazine on lar-
val development and sexual differentiation in the northern leopard frog Rana pipiens.
Pechmann, J. H. K., Scott, D. E., Semlitsch, R. D., Caldwell, J. P., Vitt, L. T., and
Relyea, R. A. and Mills, N. (2001). Predator induced stress makes the pesticide car-
baryl more deadly to gray treefrog tadpoles (Hyla versicolor). Proceedings of the National
Rich, P. M. (1989). A Manual for Analysis of Hemispherical Canopy Photography.

Publication LA-11733-M. Los Alamos National Laboratory, Los Alamos, NM.
Salek, L. and Zahradnik, D. (2008). Wedge prism as a tool for diameter and distance
measurement. Journal of Forest Science, 54, 121–4.
Skelly, D. K., Werner, E. E., and Cortwright, S. A. (1999). Long-term distributional
dynamics of a Michigan amphibian assemblage. Ecology, 80, 2326–37.
Skidds, D. E., Golet, F. C., Paton, P. W.C., and Mitchell, J. C. (2007). Habitat correlates
of reproductive effort in wood frogs and spotted salamanders in an urbanizing water-
shed. Journal of Herpetology, 41, 439–50.
Southwood, T. R. E. (1977). Habitat, the template for ecological strategies? Journal of
Animal Ecology, 46, 337–65.
Steinberg, C. E. W., Kamara, S., Prokhotskaya, V. Y., Manusadzianas, L., Karasyova, T. A.,
Timofeyev, M. A., Jie, Z., Paul, A., Meinelt, T., Farjalla, V. F. et al. (2006). Dissolved
humic substances – ecological driving forces from the individual to the ecosystem level?
Freshwater Biology, 51, 1189–1210.
Taub, F. B. (1961). The distribution of red-backed salamanders, Plethonon c. cinereus,
within the soil. Ecology, 42, 681–8.
Thomas, D. L. and Taylor, E. J. (1990). Study designs and tests for comparing resource
use and availability. Journal of Wildlife Management, 54, 322–30.
Thomas, D. L. and Taylor, E. J. (2006). Study designs and tests for comparing resource
use and availability II. Journal of Wildlife Management, 70, 324–36.
Vierling, K. T., Vierling, L. A., Gould, W. A., Martinuzzi, S., and Clawges, R. M. (2008).
Lidar: shedding new light on habitat characterization and modeling. Frontiers in Ecology
and Environment, 6, 90–8.
Wahl, III, G. W., Harris, R. N., and Nelms, T. (2008). Nest site selection and embryonic
survival in four-toed salamanders, Hemidactylium scutatum (Caudata: Plethodontidae).
Wellborn, G. A., Skelly, D. K., and Werner, E. E. (1996). Mechanisms creating com-
munity structure across a freshwater habitat gradient. Annual Review of Ecology and
Systematics, 27, 337–63.
Welsh, Jr, H. H. and Ollivier, L. M. (1998). Stream amphibians as indicators of ecosystem
stress: a case study from California’s redwoods. Ecological Applications, 8, 1118–32.
Welsh, Jr, H. H. and Lind, A. J. (2002). Multiscale habitat relationships of stream amphib-
ians in the Klamath-Siskiyou Region of California and Oregon. Journal of Wildlife
Welsh, Jr, H. H., Ollivier, L. M., and Hankin, D. G. (1997). A habitat based design
for sampling and monitoring stream amphibians with an illustration from Redwood
National Park. Northwestern Naturalist, 78, 1–16.
Wiens, J. A. (1989). The Ecology of Bird Communities. Vol. 1, Foundations and Patterns.
Cambridge University Press, Cambridge.
Wyman, R. L. and Hawksely-Lescauly, D. S. (1987). Soil acidity affects distribution,
behavior and physiology of the salamander Plethodon cinereus. Ecology, 68, 1819–27.
Part 5
Amphibian communities
18
Diversity and similarity
C. Kenneth Dodd, Jr
18.1 Introduction
Determining how many species are present in a particular habitat, their abun-
dance, their importance, and how communities are related are critical questions
in ecology and conservation biology. Ecologists use measures of diversity and
similarity to understand community structure and function, particularly in
terms of habitat use, food webs, predator–prey relationships, estimating how
many species can co-exist within a community, energy flow, and nutrient cyc-
ling. Conservation biologists need to estimate diversity to identify areas for pro-
tection and management (Scott et al. 1987; Snodgrass et al. 2000), and to assess
the effects of habitat change through time. Knowledge of these variables also is
very practical for consultants or others doing rapid assessments, such as in the
preparation of environmental impact assessments. Species diversity (richness,
heterogeneity, evenness) is fairly simple to estimate, yet the use of these indices
in amphibian studies is generally less than in many other areas of field research,
and most field amphibian researchers have yet to explore the utility of measures
of similarity.
There are many indices used to express various aspects of diversity and simi-
larity; some of the most common are listed in Table 18.1. Magurran (1988,
2003), Krebs (1999), and Clarke and Warnick (2001) provide good discussions
of some diversity and similarity indices, and these references can serve as a start-
ing point for understanding what the indices do and how they are calculated,
and for where to find more information about them. Whatever index is chosen,
it is important to understand the assumptions surrounding the computation,
the biases of the index, and the way the index should be interpreted. It is very
common for biologists to sacrifice rigor (by violating assumptions) for ease of
computation or precedence of use by other biologists. Some of the easiest indices
to use (for example, Margalef) are also some of the most informative, whereas
Table 18.1 Indices and measures of species diversity commonly

used in ecological studies.
Measure of . . . Test
Species richness Rarefaction

Menhinick (DMn)
Jacknife
Richness (S)
Bootstrap
Margalef (DMg)
Species-area curve estimates
McIntosh (U)
Heterogeneity (logarithmic)
Brillouin’s Index (HB)
(log normal)
Fisher’s
Simpson’s Index (1-D)
Shannon-Wiener Function (H)
Evenness and dominance Simpson’s (D)
Berger–Parker (d)
Camargo (E)
Hill’s Ratio (N1)
Smith and Wilson (EQ)
Pielou’s (E)
Modified Nee
McIntosh (D)
other indices (for example, Shannon) are popular but not very informative by
themselves.
18.2 Data transformation

Measures of diversity and similarity give differing weight to abundance; that is,
some indices are more sensitive to the presence of rare species, whereas others
are more sensitive to the abundance of the more common species. Thus, using
the same data set but analyzing it with different indices could result in very dif-
ferent profiles of community diversity (Bloom 1981). Inclusion of rare species in
diversity and similarity estimates might indicate that a community is far richer
than it normally is, especially if the species is migratory or an extremely unusual
find. On the other hand, metamorphosis could heavily weight an amphibian
community index toward a species that happened to transform during the sam-
pling period.
18 Diversity and similarity | 323
Table 18.2 The effect of transformation on perceptions of abundance. Transformation

normalizes data, and allows for variable weighting in studies of amphibian diversity. As
you move to the right in the table, rare species assume greater importance in the analysis
of community similarity/dissimilarity.
No transformation Square root Double square root Log(1+x)1 Presence/absence
(N) (√) (√√)
0 0 0 0 0
1 1 1 0.3 1
10 3.3 1.7 1 1
100 10 3.3 2 1
1000 33 5.7 3 1
10 000 100 10 4 1
1
Because the log(0) ∞, 1 is added to the value inasmuch as the log(1y) 0 when y 0.
One way to change weighting would be to transform the data. Table 18.2
shows how data transformation can weight the ‘importance’ of the less abun-
dant or rare species, and thus allow them to have more influence in perceptions
of community patterns. Data transformation also helps in clustering and ordin-
ation methods. Thus, it should be apparent that the objectives of the researcher
become very important in the selection of diversity and similarity indices.
Knowing the limitations and assumptions of the indices helps avoid confusion
and aids in the interpretive process.
A second reason for transforming data is that sampling usually involves large
numbers of zero captures; that is, all species are not caught on each day of sam-
pling. Zeros present complications in data analysis, especially when using para-
metric measures which assume normal distributions. Transformation offers a
means to validate the statistical assumptions of parametric techniques prior to
using them. One must always remember to verify that the data transformation
actually corrects the original issue with the data set. If it does not, transformation
is not warranted. Of course, an alternative approach would be to avoid paramet-
ric statistics, and many of the diversity and similarity indices commonly used
are based on non-parametric techniques.
18.3 Species diversity

18.3.1 Sampling considerations
There are a number of considerations to remember if one of the goals of sampling
is to develop an estimate of species diversity within an area. First, estimates of
diversity are only as accurate as the detection probabilities of finding individual
species during the sampling period. This is especially important inasmuch as

most species have heterogeneous detection probabilities due to a variety of both
natural and observer-related causes (Boulinier et al. 1998). Second, all sampling
techniques have biases, as numerous chapters in this volume have discussed.
Estimates of species diversity will be biased to the extent that sampling tech-
niques are biased. Examples of such biases include the following.
• Methods. If the techniques used will not detect the species, estimates
of diversity within an area will be inaccurate. If the goal is to estimate
amphibian community diversity, then multiple and appropriate tech-
niques must be employed, especially if rare species are suspected to be in
the study area.
• Timing. Sampling must extend throughout the activity period of all the
species in question, or at least it must include representative subsamples
throughout the activity period. For example, the timing of sampling has
been shown to alter estimates of species richness by changing the shape of
the species accumulation curve. If continuous sampling is not possible,
biologists might consider randomization (for example, by randomizing the
start of weekly sampling periods throughout an activity season).
• Location. Estimates of community diversity will only be accurate if all
habitats (arboreal, aquatic, terrestrial, fossorial) are sampled. Species diver-
sity estimates should not be extrapolated beyond the area sampled (that is,
the statistical area of inference).
• Stochastic events. Weather conditions (heat, cold, drought, precipitation)
and unexpected disturbances (catastrophic storms, earthquakes) all have
the potential to influence sampling and thus estimates of diversity. Where
necessary, biologists need to apply caveats to their estimates to acknow-
ledge the limitations of their data.
18.3.2 Species richness

The number of species within a community, species richness, is the simplest
way to describe local and regional diversity (Magurran 1988). Although species
richness is a natural measure of diversity, it is also an elusive quantity to measure
accurately (May 1988; Gotelli and Colwell 2001), especially if species have dif-
ferent detection probabilities during the sampling period (de Solla et al. 2005).
The number of species observed is frequently an underestimate of actual species
richness if based on counts conducted over limited periods of time. Estimates of
species richness are very sensitive to the sample size and methods of sampling on
which they are based (Smith and van Belle 1984).
Species richness does not imply anything about abundance, although the
term frequently is used inaccurately as a synonym of diversity (see below). In
some situations, species richness can be estimated in advance using field guides
or museum specimens, coupled with knowledge of habitat requirements; this
method is termed interpolation. For example, the number of species found
within a forested area of a national park might be predicted from published lit-
erature and museum records. Interpolation can be very inaccurate, depending
on scale. Knowledge of species richness in one watershed might be useful for
predicting species occupancy in an adjacent watershed, but not on a regional
scale. Thus, interpolation does not supplant the necessity of field sampling for
an accurate understanding of richness since myriad factors may influence occu-
pancy at a particular location.
The number of species within an area results from a complex interaction
of resource availability, habitat complexity, biogeography, land-use history,
and phylogenetic history. Habitats with large numbers of amphibian species
include tropical rainforests and areas with diverse topography and, as might be
expected, amphibian species richness is inversely correlated with latitude and,
usually, elevation (Duellman 1999). Even within lowland rainforests, how-
ever, amphibian species richness changes spatially, so sampling considerations
become paramount when estimating how many species are present. An estimate
of species richness is only as accurate as the reliability of the sampling methods
on which the estimate is based.
18.3.3 Species accumulation curves

One measure of species richness in a region is the rate at which new species are
added to an inventory (Soberón and Llorente 1993). If repetitive samples are
taken using standardized techniques, the rate at which species are detected and
the point at which detection of new species levels off should give an indication
of the number of species within an area. Such information is useful when little
or nothing is known a priori concerning species richness.
Of course, sampling every species within a habitat may not be temporally or
logistically possible, and rarely would all species be captured with 100% con-
fidence within a limited sampling period. Thus, it usually becomes necessary
to estimate statistically the number of species within the area of interest. This
is done by plotting the increasing number of species captured (on the y axis,
termed the ordinate) by either effort (assuming sampling is regularly spaced
through time) or by the cumulative number of individuals observed or captured
(on the x axis, termed the abscissa). Effort is defined quantitatively, such as by
trap-days, hours searched, or number of pitfalls checked. Thus, an estimate of
species richness is derived through cumulative sampling, preferably over a sub-

stantial period of time.
Species accumulation curves can be used in two primary ways. First, esti-
mates of the number of amphibian species within the sampling area during the
time of the survey can be derived (Thompson et al. 2003; Dodd et al. 2007),
especially when that number is imprecisely known, such as during an inven-
tory of tropical forest habitats. Second, estimates can be derived of the amount
of effort needed to sample a particular area with a preset degree of confidence
(usually within 80–90% of the asymptote) (Thompson and Thompson 2007;
Thompson et al. 2007). During the initial phases of sampling, a species accu-
mulation curve will rise steeply, especially in species-rich communities (Cam
et al. 2002). When the curve levels off, it is likely that additional sampling
will not detect many more species, and thus sampling could be curtailed. This
technique helps avoid unnecessary sampling, especially when it is logistically
difficult or expensive.
Species-richness estimates are of two types: extrapolative (inferring species
richness based on subsamples of a data set) and interpolative (as above, infer-
ring species richness based on comparisons with other areas or data sets). The
former method is widely used, with three distinct classes of statistical approaches
(Chazdon et al. 1998): (1) extrapolation of either species accumulation curves or
species-area curves to an asymptotic value; (2) fitting data on the relative abun-
dance of species in a single sample to a parametric distribution (e.g. log-series,
log-normal, Poisson log-normal); and (3) non-parametric estimators.
The failure to detect rare species can dramatically underestimate the true
local species richness. If, however, a limited fraction of a specific taxonomic
group is sampled quantitatively, sampling bias can be reduced by using statis-
tical extrapolation to estimate species richness (Colwell and Coddington 1994).
Several non-parametric estimators either have been developed specifically for
estimating species richness from samples, adapted to do so from mark–recapture
applications, or were developed for the general class-estimation problem (Smith
and van Belle 1984; Colwell and Coddington 1994). These non-parametric esti-
mators only require the number of samples in which each species is found, rather
than any parametric information about their abundance (Brose et al. 2003).
Some of them can be reduced to a very simple form: S estimated Sobserved R, where
R is an estimate based on whether rare species are present or undetected in the
samples. Overall, non-parametric estimators appear to be less biased and more
precise than the other two approaches.
The shape of species-accumulation curves are influenced by both abundance
and diversity (Thompson and Withers 2003). If rare species are present, or if
there are few species with high abundance, accumulation curves have low shoul-
ders and long trajectories to the asymptote. Conversely, areas with large num-
bers of abundant species have steep trajectories and reach asymptotes quickly.
Species diversity is positively correlated with the initial slope of the trajectory of
the accumulation curve.
Both incidence-based (that is, occupancy) and abundance-based (that is,
incorporating diversity) species accumulation curves can be generated. A com-
puter program (such as the Mao Tau of EstimateS, see below) might use 1000
iterations of a data set to predict the expected range of shapes of the accumu-
lation curve. This allows biologists to incorporate confidence intervals (95%)
around the accumulation curve, which provides a better estimate of the actual
numbers of species likely within the area than a simple curve. An example of
species-accumulation curves based on intensive pitfall and other sampling tech-
niques for amphibians is shown in Figure 18.1.
There are few applications of these methods dealing with amphibians. Pineda
and Halffter (2004) used them to verify the completeness of inventories at
both local and regional scales and determined the sampling effort needed for
reaching a plateau. Heyer et al. (1999) tested the utility of museum collections
for conservation decisions, focusing on frogs of the genus Leptodactylus from
Amazonia. The results indicated that the data set was adequate in terms of sam-
pling effort and useful for conservation decisions, at least for amphibians. Dodd
et al. (2007) used species accumulation curves to compare changes in species
presence through time as a consequence of habitat changes. They showed that
long-term changes in habitat management resulted in decreases in species rich-
ness within the amphibian community.
18.3.4 Heterogeneity
The concept of heterogeneity (also sometimes termed ‘species diversity’) pro-
vides a way of expressing both the number of species and a measure of counts or
abundance into a single index. What is more diverse: a community with many
rare species or one with fewer but much more abundant species? What if two
communities have the same number of species, but at different abundances?
In computations, actual (rarely measured) or relative (based on counts) abun-
dances are combined with species richness to measure the heterogeneity of the
community. The result is an index which allows the observer to compare the
heterogeneity of one or more communities (for example, Dodd 1992). An index
by itself tells little; it gains value when it is used comparatively.
Amphibians are sampled as one might for other measures of diversity; that
is, via a variety of techniques that provide capture histories through time.
35
30
Number of amphibian species
25
20
15
10
Historic
Recent
5
0
2 4 6 8 10 12
Number of samples
40
35
Number of reptile species
30
25
20
15
Historic
Recent
10
5
2 4 6 8 10 12
Number of samples
Fig 18.1 Species accumulation curves, with 95% confidence intervals, for amphibians
(top) and reptiles (bottom) sampled at 12 study sites at St. Marks National Wildlife
Refuge, Florida, USA. The historic curve shows relationships in 1977–9, whereas the
recent curve shows the relationships in 2002–5. These species accumulation curves
suggested that fewer species of amphibians were expected in the 2000s than in the
1970s. In the 1970s, the curves predicted that 29 species of amphibians might be
present throughout the sampling sites, but only 19 in the 2000s. The actual values
were 29 (1970s) and 24 (2000s). Reprinted from Dodd et al. (2007).
Abundance is usually and somewhat erroneously expressed in terms of counts

taken during the course of a sampling period. Counts and abundance are not
the same thing. A count is simply that: the number of individuals captured or
observed during the course of a sampling period. On the other hand, true abun-
dance is estimated by the equation:
Nˆ C / pˆ
where N̂ is abundance, C is the number of animals counted, and p̂ is the prob-

ability of detecting an individual. Thus, counts may or may not be a relative
index of abundance and, hence, statistical indices based on counts may or may
not be accurate (see Chapter 23). Given the unlikelihood of obtaining true esti-
mates of abundance for all members of a community, counts will have to suffice
if diversity indices are used.
18.3.5 Evenness and dominance

Evenness and dominance refer to the influence a species has in numbers on the
community, particularly the most abundant species. Evenness is an important
variable to measure, since high evenness is generally considered synonymous
with high diversity (Magurran 1988); the most diverse community is thought
of as one that has a large number of species and a large number of individuals
uniformly abundant among samples. Like other measures of the proportional
abundance of species, dominance and evenness indices incorporate estimates
of richness and abundance. The result is a number that can be used to interpret
whether a community is dominated by one or more species, or if species are more
equitably distributed within the community. The biologist then uses this infor-
mation to examine the data and identify the species responsible for dominance.
Representative evenness/dominance indices are listed in Table 18.1.
18.4 Similarity
Measures of similarity are used to examine data from a number of sampling
areas in order to compare community similarity, or more correctly, dissimilar-
ity among those sampling areas. The result is a matrix of numbers representing
paired comparisons which can be difficult to interpret without a visual context.
However, these matrix-based comparisons can be fitted into a graphic depiction
of similarity, aligning those communities or sampling areas most similar to one
another into a cluster dendrogram. Cluster dendrograms then can be compared
to see how communities change through time and how variables change from
one community to another. One of the advantages of using similarity measures

is that, depending on the measure, various types of data can be compared, such
as species richness, abundance, biomass, or ecological parameters. Thus, these
measures can be used to compare species diversity as well as some of the factors
that might influence community structure or function.
Distance coefficients are often used to calculate the indices of similarity
for some of the most commonly used approaches (for example, Bray–Curtis,
Canberra), and computationally these are actually measures of dissimilar-
ity. Thus, the measure of similarity is the reciprocal of the calculated value.
When communities are exactly equal in the variables being compared, d=0; d
approaches 1 as the communities become more dissimilar. Distance coefficients
are computed in two ways, termed the Euclidean and Manhattan metrics. Think
of walking between two points a few blocks apart in a city. Because of build-
ings and road structure, walking a direct line (Euclidean) is not possible, so it
becomes necessary to go up and down streets in a somewhat non-direct pathway
(Manhattan). Using a similar logic, popular metrics such as the Bray–Curtis
Similarity Index use the Manhattan metric to estimate dissimilarity among all
the possible comparisons within the data matrix, although other indices employ
a somewhat less informative Euclidean metric.
Of course, it helps to know which species are responsible for perceptions of
similarity or dissimilarity. There is a routine (SIMPER) in the statistical pro-
gram Primer-E which can be used to compute the overall percent contribution
that each species (or variable) makes to the dissimilarity between groups in a
dendrogram cluster analysis (see below). The outcome is a list of species (or vari-
ables) in a decreasing order of importance in discriminating between all possible
pairs of dissimilarity coefficients within the cluster (Clarke and Gorley 2006).
Similarity indices are just like any other indices as they have no real signifi-
cance except comparatively. The indices will not tell the observer why certain
communities are more similar or dissimilar to one another, or tell him or her
why communities have changed in diversity through time. Instead, they help
direct the observer toward the reasons for change, and allow for the development
of hypotheses which can be tested using other data, future experimentation, or
habitat manipulation. Some of the common similarity indices (see Magurran
1988, 2003; Krebs 1999; Clarke and Warnick 2001) include the following.
• Jaccard (C) uses binary data in a simplistic formula (Cjk p/p m) where p
is the number of species in both communities j and k, and m is the number of
species in one community but not the other. This is not a distance measure.
• Morisita (C) is a true similarity index formulated for counts of individuals only.
• Horn is a modification of the Morisita C that uses proportions instead

of counts. It is thus not affected by sample size, and can be used for vari-
ables other than count or abundance data. However, it is not as robust as
Morisita.
• Bray–Curtis (B) is a distance measure of dissimilarity, whereby coefficients
are weighted toward abundant species, with rare species adding very lit-
tle to the value. In addition to counts, matrices of catch-per-unit-effort or
environmental variables can be analyzed for dissimilarity. See Beals (1984)
and Clarke et al. (2006).
• Canberra (C) is another distance measure of dissimilarity, but one that is
not as influenced by abundant species. This index gives more weight to rare
species than the Bray–Curtis Index.
• Other similarity/dissimilarity indices include the Squared Euclidean
Distance, Percent Similarity, Mountford, Dice, Kendall, Simplified
Morisita, Sørensen, Product Moment Correlation Coefficient, Raup–
Crick, and Whittaker.
In an example of the way similarity measures can be used, my colleagues and
I conducted extensive field sampling for amphibians at St. Marks National
Wildlife Refuge, Florida, USA, 30 years after similar sampling had occurred
there. We attempted to use the same techniques at exactly the same sites. Based
on capture results, we computed Bray–Curtis Similarity Indices for each com-
munity and sampling period, square-root-transformed the Bray–Curtis indices,
and examined cluster dendrograms to visually compare community similar-
ity. Changes in the dendrograms indicated significant changes in community
similarity through time (Figure 18.2). We then re-examined weather condi-
tions, habitat types, succession stages, and management practices during the
two sampling periods. This led us to hypothesize four possible causes, acting
interactively, that led to changes in species diversity within the amphibian com-
munity (Dodd et al. 2007).
18.5 Software
The program EstimateS does most computations required for species accumu-
lation curves and non-parametric analyses of species richness, diversity, and
dominance. A detailed description of the estimators computed can be found in
Colwell and Coddington (1994), Chazdon et al. (1998), and Colwell (2006).
EstimateS (version 8.0) computes randomized species accumulation curves,
statistical estimators of true species richness (S), and a statistical estimator of
20
40
Similarity
60
80
100
WBF HYR SPC EYR NAT SBL SYR CLH LPH PRS BSF UBF
20
40
Similarity
60
80
100
LPH PRS HYR WBF SPC EYR NAT SBL SYR CLH BSF UBF
Fig 18.2 Cluster dendrograms based on square-root-transformed Bray–Curtis

Similarity Index values showing the relationship among habitat types and the
amphibian community at 12 study sites (coded on the x axis) at St. Marks National
Wildlife Refuge, Florida, USA. The top cluster shows relationships in 1977–9,
whereas the cluster at the bottom shows the relationships in 2002–5. The
amphibian communities in the 1970s were most similar to one another within
management units, regardless of habitat type or ongoing management practice.
The pattern changed somewhat in the 2000 sampling period. Instead of a similarity
based on east-to-west location, amphibian community similarities appeared to be
based on dominant community type. Reprinted from Dodd et al. (2007).
the true number of species shared between pairs of samples, based on species-by-
sample (or sample-by-species) incidence or abundance matrices. It can be used
to compute an expected species accumulation curve, the Mao Tau (with 95%
confidence limits). The Mao Tau is a sample-based rarefaction curve which pro-
vides a graphic estimate of expected species accumulation (Colwell et al. 2004;
Figure 18.1). The program also can be used to compute both incidence-based
coverage estimates (ICE) and abundance-based coverage estimates (ACE) of
species richness among sampling sites. The derivation and use of these estima-
tors is discussed by Chazdon et al. (1998) and Colwell (2006).
EstimateS further allows computation of Fisher’s , Shannon, and Simpson
diversity indices; the Chao, Jacknife, ICE, ACE and other species richness esti-
mators for abundance and incidence data; modified versions of the Sørensen
and Jaccard similarity indices based on abundances, including the effects of
unseen shared species (Chao et al. 2005); and classic Jaccard, Bray–Curtis, and
Morisita–Horn (both incidence-based and abundance-based) similarity esti-
mators. EstimateS can be downloaded free of charge from http://viceroy.eeb.
uconn.edu/estimates.
Ecological software to accompany Kreb’s Ecological Methodology is available
from Exeter Software (Setauket, New York, USA; www.exetersoftware.com/
cat/ecometh/ecomethodology.html). Version 6.1.3 sells for US$150 (as of May
2008). The software covers only the topics and analyses discussed in the book,
which include various measures of richness, heterogeneity, and equitability, in
addition to a wide range of other ecological analyses. These include binary coef-
ficients, Euclidean distance coefficients, Bray–Curtis metric, Canberra met-
ric, percentage similarity, Morisita’s Index of Similarity, Horn’s Index, Species
Richness (Rarefaction method, Jackknife method for counts), logarithmic ser-
ies, log-normal distribution, Simpson’s Index of Diversity, Shannon–Wiener
measure, Brillouin’s Index of Diversity, and evenness measures. The program
comes with a manual describing the analyses, and includes some information on
the use and theory behind the various indices.
Primer 6 for Windows is another powerful tool for analysis of ecological data,
including both diversity and dominance indices. Information on the program
can be obtained from the website (www.primer-e.com/). Primer 6 allows uni-
variate, graphical, and multivariate analyses of species, abundance, biomass,
and physio-chemical data. The program facilitates grouping data into clusters,
allows identification of species that are responsible for discrimination among
two sample clusters, and graphs species abundance distributions through ordin-
ation and multidimensional scaling plots. As of May 2008 the cost is US$500
for a single-user license for academic research and is available from Primer-E
(Ivybridge, Devon, UK). Primer-E holds advanced workshops, the cost of

which includes the manual (Clarke and Gorley 2006) and a book demonstrat-
ing the use of Primer 6 programs on data from the Plymouth Marine Laboratory
(Clarke and Warnick 2001). Information on forthcoming workshops may be
obtained from the website.
18.6 Summary
Indices of diversity and similarity offer a means of critically examining cur-
rent ecological patterns and changes in community composition, especially
when estimates of site occupation across a sufficient number of habitats are not
available. In particular, the Bray–Curtis Similarity Index has proven useful in
assessing the effects of habitat changes on herpetofauna and other taxa during
monitoring programs (Pawar et al. 2004; Pieterson et al. 2006; Dodd et al.
2007), comparing stream and forest faunas (Parris and McCarthy 1999; Huang
and Hou 2004), measuring the success of restoration efforts (Ruiz-Jean and
Aide 2005), prioritizing areas for conservation (Seymour et al. 2001), assess-
ing dietary differences (Whitfield and Donnelly 2006), and in analyses of geo-
graphic differences in diversity patterns (Urbina-Cardona and Londroño-M
2003; Menegon and Salvidio 2005; Smith-Vaniz et al. 2006).
An index-based approach offers insights into potential, if not definitive causes
of community change. Once potential causes are identified, research can be
designed to test hypotheses related to changes in species composition and rela-
tive abundance. Community ecology, conservation, and monitoring programs
should incorporate an evaluation of habitat variables into data-collection proto-
cols, and researchers must be aware of the potential importance of stochastic
or periodic environmental disturbances, such as storms and flooding, when
interpreting species presence/not detected and abundance data. Counting indi-
vidual animals or determining the percentage of site occupancy neglects much
information needed to understand changes in community composition through
time. The use of diversity and similarity indices offers further insight into the
comparative structure of amphibian communities.
18.7 References
Beals, E. W. (1984). Bray-Curtis ordination: an effective strategy for analysis of multi-
variate ecological data. Advances in Ecological Research, 14, 1–56.
Bloom, S. A. (1981). Similarity indices in community studies: potential pitfalls. Marine
Ecology Progress Series, 5, 125–8.
Boulinier, T., Nichols, J. D., Sauer, J. R., Hines, J. E., and Pollock, K. H. (1998).
Estimating species richness: the importance of heterogeneity in species detectability.
Ecology, 79, 1018–28.
Brose, U., Martinez, N. D., and Williams, R. J. (2003). Estimating species richness: sen-
sitivity to sample coverage and insensitivity to spatial patterns. Ecology, 84, 2364–77.
Cam, E., Nichols, J. D., Sauer, J. R., and Hines, J. E. (2002). On the estimation of species
richness based on the accumulation of previously unrecorded species. Ecography, 25,
102–8.
Chao, A., Chazdon, R. L., Colwell, R. K., and Shen, T.-J. (2005). A new statistical
approach for assessing similarity of species composition with incidence and abundance
data. Ecology Letters, 8, 148–59.
Chazdon, R. L., Colwell, R. K., Denslow, J. S., and Guariguata, M. R. (1998). Statistical
methods for estimating species richness of woody regeneration in primary and second-
ary rainforests of northeastern Costa Rica. In F. Dallmeier and J. A. Comisky (eds),
Forest Biodiversity Research, Monitoring and Modelling: Conceptual Background and Old
World Case Studies, pp. 285–309. Smithsonian Institution/Man and Biosphere No. 20.
Smithsonian Institution, Washington DC.
Clarke, K. R. and Warnick, R. M. (2001). Change in Marine Communities: an approach to
statistical analysis and interpretation, 2nd edn. Primer-E, Plymouth.
Clarke, K. R. and Gorley, R. N. (2006). PRIMER v6: user manual/tutorial. Primer-E,
Plymouth.
Clarke, K. R., Somerfield, P. J., and Chapman, M. G. (2006). On resemblance measures
for ecological studies, including taxonomic dissimilarities and a zero-adjusted Bray-
Curtis coefficient for denuded assemblages. Journal of Experimental Marine Biology and
Ecology, 330, 55–80.
Colwell, R. K. (2006). EstimateS 8.0 user’s guide. http://viceroy.eeb.uconn.edu/
EstimateSPages/EstSUsersGuide/EstimateSUsersGuide.htm.
Colwell, R. K. and Coddington, J. A. (1994). Estimating terrestrial biodiversity through
extrapolation. Philosophical Transactions of the Royal Society London Series B Biological
Sciences 345, 101–18.
Colwell, R. K., Mao, C. X., and Chang, J. (2004). Interpolating, extrapolating, and com-
paring incidence-based species accumulation curves. Ecology, 85, 2717–27.
de Solla, S. R., Shirose, L. J., Fernie, K. J., Barrett, G. C., Brousseau, C. S., and Bishop, C. A.
(2005). Effect of sampling effort and species detectability on volunteer based anuran moni-
toring programs. Biological Conservation, 121, 585–94.
Dodd, Jr, C. K. (1992). Biological diversity of a temporary pond herpetofauna in north
Florida sandhills. Biodiversity and Conservation, 1, 125–42.
Dodd, Jr, C. K., Barichivich, W. J., Johnson, S. A., and Staiger, J. R. (2007). Changes
in a northwestern Florida Gulf Coast herpetofaunal community over a 28-y period.
American Midland Naturalist, 158, 29–48.
Duellman, W. E. (1999). Global distribution of amphibians: patterns, conservation, and
future challenges. In W. E. Duellman (ed), Patterns of Distribution of Amphibians. A
Global Perspective, pp. 1–30. Johns Hopkins University Press, Baltimore, MD.
Gotelli, N. J. and Colwell, R. K. (2001). Quantifying biodiversity: procedures and pitfalls
in the measurement and comparison of species richness. Ecology Letters, 4, 379–91.
Heyer, R. W., Coddington, J., Kress, J. W., Acevede, P., Cole, D., Erwin,T. L., Meggers,
B. J., Pogue, M. G., Thorington, R. W., Vari, R. P., Weitzman, M. J., and Weitzman,
S. H. (1999). Amazonic biotic data and conservation decisions. Ciencia e Cultura,
Journal of the Brazilian Association for the Advancement of Science, 51, 372–85.
Huang, C.-Y. and Hou, P.-C. L. (2004). Density and diversity of litter amphibians in a
monsoon forest of southern Taiwan. Zoological Studies, 43, 795–802.
Krebs, C. J. (1999). Ecological Methodology, 2nd edn. Benjamin-Cummings, Menlo
Park, CA.
Magurran, A. E. (1988). Ecological Diversity and its Measurement. Princeton University
Press, Princeton, NJ.
Magurran, A. E. (2003). Measuring Biological Diversity. Blackwell Publishing, Oxford.
May, R. M. (1988). How many species are there on Earth? Science, 241, 1441–9.
Menegon, M. and Salvidio, S. (2005). Amphibian and reptile diversity in the south-
ern Udzungwa Scarp Forest Reserve, south-eastern Tanzania. In B. A. Huber,
B. J. Sinclair, and K.-H. Lampe (eds), African Biodiversity. Molecules, Organisms,
Ecosystems, pp. 205–12. Springer, New York.
Parris, K. M. and McCarthy, M. A. (1999). What influences the structure of frog assem-
blages at forest streams? Australian Journal of Ecology, 24, 495–502.
Pawar, S. S., Rawat, G. S., and Choudhury, B. C. (2004). Recovery of frog and lizard
communities following primary habitat alteration in Mizoram, northeast India. BMC
Ecology, www.biomedcentral.com/1472-6785/4/10.
Pieterson, E. C., Addison, L. M., Agobian, J. N., Brooks-Solveson, B., Cassani, J., and
Everham III, E., M. (2006). Five years of the southwest Florida frog monitoring
network: changes in frog communities as an indicator of landscape change. Florida
Scientist, 69, 117–26.
Pineda, E. and Halffter, G. (2004). Species diversity and habitat fragmentation: frogs in a
tropical montane landscape in Mexico. Biological Conservation, 117, 499–508.
Ruiz-Jean, M. C. and Aide, T. M. (2005). Restoration success: how is it measured?
Restoration Ecology, 13, 569–77.
Scott, J. M., Csuti, B., Jacobi, J. D., and Estes, J. E. (1987). Species richness. A geographic
approach to protecting future biological diversity. BioScience, 37, 782–8.
Seymour, C. L., De Klerk, H. M., Channing, A., and Crowe, T. M. (2001). The bio-
geography of the Anura of sub-equatorial Africa and the prioritisation of areas for their
conservation. Biodiversity and Conservation, 10, 2045–76.
Smith, E. P. and van Belle, G. (1984). Nonparametric estimation of species richness.
Biometrics, 40, 119–29.
Smith-Vaniz, W. F., Jelks, H. L., and Rocha, L. A. (2006). Relevance of cryptic fishes
in biodiversity assessments: a case study at Buck Island Reef National Monument,
St. Croix. Bulletin of Marine Science, 79, 17–48.
Snodgrass, J. W., Komorosski, M. J., Bryan, Jr, L., and Burger, J. (2000). Relationships
among isolated wetland size, hydroperiod, and amphibian species richness: implica-
tions for wetland regulation. Conservation Biology, 14, 414–19.
Soberón, M. J. and Llorente, B. J. (1993). The use of species accumulation functions for
the prediction of species richness. Conservation Biology, 7, 480–8.
Thompson, G. G. and Withers, P. C. (2003). Effect of species richness and relative abun-
dance on the shape of the species accumulation curve. Austral Ecology, 28, 355–60.
Thompson, G. G. and Thompson, S. A. (2007). Using species accumulation curves to

estimate trapping effort in fauna surveys and species richness. Austral Ecology, 32,
564–9.
Thompson, G. G., Withers, P. C., Pianka, E. R., and Thompson, S. A. (2003). Assessing
biodiversity with species accumulation curves; inventories of small reptiles by pit-
trapping in Western Australia. Austral Ecology, 28, 361–83.
Thompson, G. G., Thompson, S. A., Withers, P. C., and Fraser, J. (2007). Determining
adequate trapping effort and species richness using species accumulation curves for
environmental impact assessments. Austral Ecology, 32, 570–80.
Urbina-Cardona, J. N. and Londoño-M, M. C. (2003). Distribución de la comunidad
de herpetofauna asociada a cuatro áreas con diferente grado de perturbación en la Isla
Gorgona, Pacífico Colombiano. Revista de la Academia Colombiana de Cienias Exactas
y Fisicas Naturales, 27, 105–13.
Whitfield, S. M. and Donnelly, M. A. (2006). Ontogenetic and seasonal variation in the
diets of a Costa Rican leaf-litter herpetofauna. Journal of Tropical Ecology, 22, 409–17.
19
Landscape ecology and GIS methods
Viorel D. Popescu and James P. Gibbs
19.1 Introduction
19.1.1 Relevance of landscape ecology to
amphibian biology and conservation
The field of landscape ecology “deals with the effects of the spatial configuration
of mosaics on a wide variety of ecological phenomena” (Wiens et al. 1993). The
field has emerged from recognition of the need to link ecological processes to
landscape configuration and composition (Turner et al. 2001). Another stimu-
lus has been the revelation that population processes play out over much larger
areas and time frames than previously assumed. Landscape ecology is a rela-
tively young discipline that emerged in Central and Eastern Europe (Naveh
and Lieberman 1994). The field has exploded over the past two decades largely
because of technological advances (Turner et al. 2001). On the data side, a
sudden wealth of spatial data from global positioning systems (GPS), satellite
imagery, and high-resolution aerial photography has materialized, coupled with
the Internet as a medium for its quick distribution. On the analysis side, com-
puter processing speeds have increased exponentially (Moore 1975) and a large
suite of software packages for manipulating and analyzing spatial data (that is,
Geographic Information Systems or GIS) has evolved that capitalize on these
vastly faster processing speeds. The result is a synergism between more data and
faster computers for processing it, thereby creating opportunity for tackling
increasingly complex questions in landscape ecology.
Technological advances have also occurred in a societal context in which envir-
onmental problems have become more and more pervasive. This, in turn, has
forced recognition that conservation actions must be implemented over broad
spatial scales that have not been addressed within the domain of classic ecological
research; that is, conducted on a few hectares at most (Turner et al. 2001). As a
result, the “landscape perspective” has become widespread in ecology and GIS
has become an essential component of the ecologist’s toolbox.
In the realm of amphibian biology, Douglas Gill’s studies of the metapopula-
tion ecology of red-spotted newts (Gill 1978a, 1978b) were among the first to
catalyze this shift in scale in how we conceptualize population processes. These
studies prompted recognition that the fates of individual, pond-based breeding
populations of amphibians—the focus of nearly all prior research—are often
tightly linked to the fates of other populations nearby as mediated by the flow
of migrants among them. This landscape-as-population rather than pond-
as-population perspective has been the subject of much subsequent research
(Sjögren-Gulve 1994; Pope et al. 2000; Marsh and Trenham 2001; Hels 2002;
Gamble et al. 2007; Richter-Boix et al. 2007; Werner et al. 2007) that has radic-
ally expanded the temporal and spatial scales at which we conceive of planning
approaches for conserving amphibians (Semlitsch 2000; Trenham et al. 2003;
Smith and Green 2005).
Landscape ecology is an increasing focus of amphibian research
(Figure 19.1). The goal of this chapter is to assist amphibian biologists in
0.25
Amphibian(s) and
landscape(s)
0.2
Fraction of total citations
Amphibian(s) and
habitat(s)
0.15
Amphibian(s) and
genetic(s)
0.1
Amphibian(s) and
communit(y, ies)
0.05
Amphibian(s) and
population(s)
0
2002 2003 2004 2005 2006 2007
Year
Fig 19.1 Fraction of amphibian studies published since 1985 featuring keywords
that combine “amphibian(s)” with “landscape(s),” “habitat(s),” “genetic(s),”
“communit(y, ies),” and “population(s),” indicating a recent surge in interest
in landscape-level studies of amphibians relative to that on habitats, genetics,
communities, or populations (Source: Thomson Scientific’s Web of Science; data
plotted for last 5 years only).
19 Landscape ecology and GIS methods | 341
grasping the extraordinary opportunities (and significant limitations) for

applying GIS to questions in amphibian biology. We emphasize GIS applica-
tions to research in the context of the broad spatial scales required to achieve
amphibian conservation over the long term.
19.1.2 Defining a “landscape” from an amphibian’s perspective

Landscape ecology is highly relevant for amphibian conservation because it deals
with mosaics and patches, which very much define the biphasic life histories of
the vast majority of amphibian species that require the use of different habitats/
patches at different life stages. For landscape ecology to be informative, how-
ever, we must scale our analysis of landscape heterogeneity relative to the per-
ception of the environment by the focal organism (Wiens 1989). For example,
the “pond-as-patch” view of spatial organization of amphibian populations and
metapopulations (Marsh and Trenham 2001) has as its advantage simplifying
scientific investigations (i.e. binary landscapes of “good” versus “bad” habitat),
but is not always appropriate. A growing body of evidence suggests that the habi-
tat matrix surrounding ponds plays a critical role in shaping migration among
pools and hence the resilience of amphibian metapopulations. What happens
to an amphibian population when forest habitat used for migration is no longer
contiguous but fragmented? What are the implications of replacing forest or
wetland with agricultural or urban development? Answers to these questions
cannot be provided unless traditional ecological research intertwines with
description and modeling of spatial processes at broad spatial scales. Landscape
ecology facilitates amphibian biologists to focus on generating a spatially expli-
cit representation of real-world processes, such as metapopulation dynamics,
gene flow, and spatial organization of amphibian populations (Figure 19.2), to
help understand how actual populations operate in the landscape.
The concept of landscape connectivity is highly relevant to amphibian con-
servation because amphibian migration is so critical for maintaining viable
amphibian populations. Most amphibians are generally regarded as highly
philopatric and having poor dispersal abilities (Semlitsch 2000) despite evidence
of long-distance movement in frogs and toads (Lemckert 2004). Landscape
connectivity implies two complementary aspects: physical and functional. For
example, reforesting an upland area between two breeding ponds restores phys-
ical connectivity for forest-dwelling amphibians, but is it enough to ensure the
movement and hence the functional connectivity between the two? Moreover,
the impact of habitat fragmentation on amphibian communities is difficult to
detect because the negative effects of fragmentation occur with a time lag. Thus,
a physical description of numbers and sizes of populations and the environment
24
17
2 5
5 3
1 6
3
1 1
3
1
1
1
1 2 2
1
2
1
2
3 1
1 500 m
1
2 1
Fig 19.2 In spite of their comparatively low dispersal abilities, productivity of

amphibians combined with movement patterns can result in large absolute numbers of
amphibians moving across landscape. The figure depicts a network of ponds (n 14)
and numbers of dispersing first-time and experienced breeding marbled salamanders
(Ambystoma opacum) during 1999 through 2005, in western Massachusetts, USA,
displayed by origin and destination ponds. Two-ended arrows indicate an equal
number of individuals dispersing in both directions (reprinted from Gamble et al.
2007, with permission from Elsevier).
they occupy might imply stability when in fact they are experiencing continuous
decline and are functionally extinct (Löfvenhaft et al. 2004).
19.1.3 GIS
A GIS offers the tools to envision, analyze, process, and model spatial data.
A GIS is essentially computer software capable of storing, manipulating, dis-
playing, and analyzing geographically referenced data. Two-dimensional spa-
tial data can be represented in vector or raster form. Vectors are points, lines,
or polygons representing real-world features (i.e. sampling locations, roads, or
land use, respectively) associated with individual data attributes. In contrast,
rasters are continuous matrices of equally sized, rectangular cells, each usually
containing a single value corresponding to the landscape feature they represent
(e.g. 1 is deciduous forest, 2 is open water, 3 is urban). Adjacent cells of similar
value represent homogenous spatial objects, such as a forest or a wetland. Raster

resolution (or grain) refers to the actual dimension of the cells: high resolution
is analogous to cells representing a small area on the ground (e.g. 1–100 m2)
and captures a high degree of spatial complexity of the landscape, whereas low
resolution refers to cells corresponding to large areas on the ground (e.g. 1 km2)
and provides a much more generalized representation of the landscape. Rasters
do not reflect the exact boundaries of a spatial object in the way the vectors
do (Figure 19.3), but their continuous nature is more conducive to the model-
ing of surfaces. At the same time, mathematical operations with rasters are less
computationally intensive than with vectors. Nonetheless, using vector data is
(a) (c)
(b) (d)
400 200 0 400 m
Fig 19.3 Spatial depiction of a set of ponds in a forested landscape using:

(a) orthoimagery (1 m resolution aerial photo; the ponds correspond to the irregular
shapes); (b) vector data (polygons); (c) raster data (cell size 10 m 10 m); and
(d) raster data (cell size 30 m 30 m). Some spatial information is lost when
increasing grain from 10 m 10 m (c) to 30 m 30 m (d).
necessary when topological relations are important, such as with analysis of

networks.
GIS applications to amphibian research range widely. Minimally GIS can be
used for simple assessment of the potential use of a region for study purposes. For
example, a perfectly legitimate use of GIS is for storing, managing, and visualiz-
ing information on access, types, and distribution of habitats, their ownership,
and location. GIS can also be used for more complicated analyses and mod-
eling. For example, GIS tools and data can permit habitat-selection analyses
(for example from telemetry studies), predict habitat suitability for amphib-
ians in the face of climate change (amphibians of Europe; Araújo et al. 2006),
and modeling invasive species’ distributions (e.g. the global spread of bullfrogs,
Rana catesbeiana (Ficetola et al. 2007) and the invasion of cane toads, Bufo mari-
nus, in Australia (Urban et al. 2007)).
In recent years both raster and vector data have become readily available
through Internet-based, primarily governmental data repositories. One of
the most comprehensive and up-to-date spatial data portals we are aware of
is Collins Software web-based Resource Management System (www.
collinssoftware.com/freegis_by_region.htm). The recent increased availabil-
ity of current spatial data is epitomized by online applications like Google
Earth or Microsoft VirtualEarth; although these applications cannot be used
for analysis purposes, they nevertheless represent GIS applications that capture
a wealth of information contained, such as high-resolution imagery, digital
elevation models, and transportation networks, as well as place names. Such
tools are increasingly valuable aids in visualizing and rapidly assessing the land-
scape; for example for reconnaissance of field sites. Notably, what once was a
special class of data—geospatial data—is increasingly being placed into the
realm of generic information management by the proliferation of GPS applica-
tions (in cell phones and cars, for example), and spatial Internet applications.
Thus, GIS may not persist as a stand-alone conceptual domain for long as
conventional database management increases its capacity to integrate spatially
explicit information.
In terms of software, a roster of GIS software bewildering to the uninitiated
is available to manipulate spatial data (Table 19.1). In organizing the choices, a
useful distinction can be made between commercial and free software. Primary
in the commercial realm are the ESRI suite (ArcGIS and its predecessors
ArcView and ArcInfo), IDRISI, MapInfo, or Manifold. In the realm of free and
open-source applications, primary are GRASS GIS, MapServer, DIVA-GIS,
ArcExplorer, JUMP, and Quantum GIS (for a list of all available free and open-
source GIS software visit freegis.org and opensourcegis.org).
Table 19.1 Popular GIS software packages.

GIS package Developer URL Capabilities
Commercial ArcGIS ESRI www.esri.com Complete GIS for

software1 spatial analysis and
modeling
IDRISI Clark Labs, www.clarklabs.org/ Powerful raster-
Clark University based GIS and
image processing
MapInfo PB MapInfo www.mapinfo.com/ Full range of spatial
Corporation analysis and
modeling
Manifold CDA www.manifold.net/ Database and
International mapping
functionality
Free and GRASS GIS Open Source grass.itc.it/ Complete GIS
open-source Geospatial similar to
software Foundation commercial
software
MapServer University of mapserver.gis.umn. Spatially enabled
Minnesota edu/ Internet
applications
DIVA-GIS R.J. Hijmans et al. www.diva-gis.org/ Raster-based
climate modeling
JUMP GIS The JUMP www.jump-project. Java-based
Project org/ vector GIS
Quantum GNU Project qgis.org/ Mostly a viewer
GIS with editing
capabilities
ArcExplorer ESRI www.esri.com/ Viewer with display,
software/ query, and
arcexplorer/ data-retrieval
index1.html applications
1
Cost of the commercial GIS packages varies greatly depending on the type of license purchased
(single versus multiple, educational versus business) as well as the array of features included.
Working with spatial data usually requires high computation requirements;

generally speaking, the larger the extent of the area analyzed, the greater the neces-
sity for a faster processor, larger random-access memory (RAM; usually 1–2 GB),
and increased hard-disk storage space. However, recent advances in computer
hardware have made once crippling hardware constraints barely noticeable on
even portable computers (for typical applications).
19.2 Applications of spatial data for

amphibian conservation
19.2.1 Multiscale predictors of species
occurrence and abundance
Many amphibian studies have reported that landscape-scale features influence
both amphibian population and metapopulation dynamics (Sjögren-Gulve
1994; Pope et al. 2000). Studies of amphibian species richness, abundance, and
occurrence invariably have relied on measures of landscape composition at dif-
ferent spatial scales surrounding central breeding pools. The most commonly
applied approach uses GIS to extract variables from land cover or habitat raster
data sets (for example, percent of forest, open water, wetland) within circular
neighborhoods at distances between 30 m and 10 km from the pool edge. These
“mini-landscapes” are usually concentric discs (i.e. overlapping neighborhoods
0–100 m, 0–200 m, . . . , 0–1000 m) or rings (i.e. non-overlapping neighbor-
hoods 0–100 m, 100–200 m, . . . , 900–1000 m). Such were used in many stud-
ies predicting amphibian occurrence and abundance (Vos and Stumpel 1995;
Hecnar and M’Closkey 1998; Joly et al. 2001; Pellet et al. 2004; Gibbs et al.
2005; Herrmann et al. 2005; Gagné and Fahrig 2007). Others have focused
on potential barriers to movement such as highways or large rivers (Zanini
et al. 2008) or specific habitat “resistances” (Ray et al. 2002) to delineate the
neighborhoods used to extract the landscape composition variables, thus limit-
ing the analysis to potential usable area only. Landscape-scale variables (e.g.
density of roads and streams, distance to nearest wetland, number of occupied
ponds, pond–forest adjacency) within the extracted neighborhoods along with
site-specific variables (i.e. water pH, conductivity, wetland area, hydroperiod,
presence of predators, etc.) were also used in predicting amphibian occurrence
and abundance. Further, landscape and pond-scale variables have been used as
variables to build predictive models for amphibian species richness or occur-
rence using ordinary linear models, generalized linear models (GLMs), general-
ized additive models (GAMs), or other modeling techniques (see Guisan and
Zimmermann 2000 for a thorough review of predictive habitat distribution
models used in ecological studies).
So-called extraction neighborhoods should reflect the dispersal ecology
of the amphibian species of concern. Relying on the assumption that most
amphibians have limited dispersal abilities, many studies extend their extrac-
tion of variables up to 1–1.5 km radii. However, during the design phase of the
study, one must pay attention to the distance between the sampling points. If
they are located too close to one another, they can be spatially correlated and
the assumption of independence of observations is violated (i.e. the abundance

of a species at one site is dependent on the abundance at a neighboring site);
pseudo-replication then occurs (Guisan and Zimmerman 2000). At the same
time, overlapping of the neighborhoods built around the sampling sites can also
lead to lack of independence among extracted variables and pseudo-replication
by re-measuring the same land unit over and over again (Figure 19.4). A further
problem is inter-variable correlation (i.e. forest cover being a direct inverse lin-
ear function of open land). The primary consequence is that pseudo-replication
biases model estimates and leads to erroneous predictions. Identification of
appropriately spaced sampling locations and non-overlapping neighborhoods
during the design phase of the study can be used as a cautionary method for
avoiding issues related to non-independence. However, if the sampling situ-
ation does not allow such flexibility, statistical techniques such as GLM or
GAM that are more relaxed to the assumption of independence among vari-
ables, and techniques for assessing and quantifying spatial autocorrelation have
to be used (see section 19.3). Practical guidelines for addressing the issue of spa-
tial autocorrelation for designing of field surveys and data analysis applicable to
amphibian research are provided by Legendre et al. (2002).
19.2.2 Landscape thresholds

The relation between habitat fragmentation and the population response by
amphibians is rarely linear. In other words, population persistence in a forest-
dependent amphibian does not simply increase directly with more forest cover.
(a) (b)
2 1 0 2 km 2 1 0 2 km
Roads Ponds 1.5 km radius neighborhood
Fig 19.4 Circular neighborhoods used for extracting landscape scale variables
around selected ponds: (a) non-overlapping (independent variables) and (b) highly
overlapping (leading to non-independence of the predictors and pseudo-replication).
Often, small changes in the spatial pattern of resource distribution can prod-
uce sudden changes in the ecological response (Figure 19.5). These transitions
are termed critical landscape thresholds (With and Crist 1995). Such critical
thresholds have been identified in the response of amphibians to the habitat
100
(a) 30 m
80
60
40 100
m = 0.58
20 (d) 500 m
b = 11.45 80
r 2 = 0.54
0 60
Spotted salamander occurrence (%)
100
40
(b) 100 m
80 m = 1.13
20 b = –9.04
60 r 2 = 0.77
0
40
100
Forest (%) m b r2
(e) 1000 m
20 0–30 2.28 –28.04 0.86 80
30–100 0.30 39.76 0.28
0 60
100
40
(c) 300 m m = 1.25
80
20 b = –12.99
r 2 = 0.88
60
0
40
10 10
20 20
30 30
40 40
50 50
60 60
70 70
80 80
0
–9
0–
–
–
–
–
–
–
–
Forest (%) m b r2
20 Upland forest (%)
0–30 2.97 –33.41 0.96
30–100 0.55 30.86 0.47
0
10 1 0
2 0 20
30 3 0
40 0
50 50
60 60
70 70
80 80
90 90
00
–4
0–
–
–
–
–
–
–
–
–1
Upland forest (%)
Fig 19.5 Occurrence of spotted salamanders (Ambystoma maculatum) plotted

against the percentage of upland forest at five different spatial scales, measured
from the edge of suitable breeding ponds. Where there is one line per graph, no
significant threshold was identified. Figures with two regression lines indicate the
location of a critical threshold. In panel (c), the lines represent the break identified
a priori by visual inspection; there were three other statistically significant break
points (reprinted from Homan et al. 2004, with permission from the Ecological
Society of America).
alteration by human activities (Gibbs 1998a; Homan et al. 2004; Denoël and
Ficetola 2007). Landscape-composition thresholds (e.g. the percentage of a
land-use type within a certain neighborhood below which the long-term viabil-
ity of the population is compromised), as well as landscape configuration (e.g.
patch isolation as function of the distance between suitable breeding and for-
aging habitat) have been documented (Denoël and Ficetola 2007).
19.2.3 Connectivity/isolation in amphibian populations

Amphibian populations respond not only to the landscape composition but also
to landscape configuration. In other words, what matters to a forest-dependent
amphibian is not just how much forest is present. The number, size, and arrange-
ment of the forest patches that comprise that amount may be more important.
Of particular interest for amphibian conservation is assessment of the isola-
tion of occupied patches (ponds in the case of pond-breeding amphibians) and
measures of connectivity between patches (Tischendorf and Fahrig 2000). For
amphibians, isolation can be simply defined as the lack of a permanent surface-
water connection between ponds (Gibbons 2003). But biological and func-
tional connectivity may occur in the absence of a permanent water connection if
terrestrial corridors suitable for animal movement exist. Isolation and connect-
ivity measures are highly relevant for the metapopulation level, but are difficult
to approximate because they depend on species-specific dispersal abilities and
physiology (Joly et al. 2003).
Generally speaking, straight-line (or Euclidian) distances are not suited for
approximating patch isolation or connectivity. Depending on its composition
and configuration, the habitat matrix may oppose particular resistance to ani-
mal movements so some landscapes are more permeable than others to these
movements, even if the straight distances are the same. By extension, a well-
watered stream bed 1 km in length may represent a vastly more functional link-
age between two populations than a dry agricultural field representing just a few
meters of separation between the same two ponds.
The significance of appropriately measured isolation/connectivity metrics
has been recognized as an integral part of the efforts for managing amphibian
populations (Semlitsch 2000). All connectivity measures rely on the assumption
that the organisms use the habitat matrix for dispersal and the resistance opposed
for movements is a function of quantifiable landscape features. Because differ-
ent species perceive the environment uniquely and also differ in intrinsic vagility
(as a function of leg length, crypsis, and desiccation resistance, for example), the
spatial configuration and composition of the landscape affect their movements
in non-intuitive ways. Consequently, researchers have developed a multitude of
indices for patch isolation and connectivity that are species- and context-specific.
Typically, connectivity measures have not considered the habitat matrix between
breeding sites and simply used “nearest neighbor” measures (Euclidian distances)
(Knutson et al. 1999; Pope et al. 2000), but studies that account for habitat matrix
complexity have started to emerge in recent years (Stevens et al. 2005; Compton
et al. 2007; Figure 19.6).
In metapopulation dynamics theory, connectivity measures are critical for
understanding and describing extinction and recolonization patterns and
processes. Hanski et al. (1994) developed a connectivity index that was used
for understanding amphibian metapopulation processes. This index requires
specific knowledge of site occupancy and patch carrying capacity as well as a
detailed map of the available patches:
n
CONNECT ∑ e
dij
Aj
i =1
where dij is the distance between patch i and patch j, and Aj is the carrying cap-
acity of site j. This so-called Hanski index has provided a useful measure of pond
connectivity for amphibian metapopulations (Schmidt and Pellet 2005).
(a) (b) (c)
Fig 19.6 Compton et al. (2007) used a modified kernel estimator (commonly
used in home-range studies) associated with a friction surface to model vernal pool
connectivity for four ambystomatid salamanders at three scales in Massachusetts,
USA. Their approach is different from a simple cost-distance calculation because
they integrated the probability of pools exchanging individuals based on empirical
salamander dispersal data (using the kernel estimator, h 339.6 m) with a
resistance map (expert-based habitat specific values). The figure depicts examples
of the resistant-kernel estimator at three scales in a landscape with a focal pool
(star), five neighboring pools (circles), and two roads: (a) local scale, showing
connectivity to upland habitat from the focal pool; (b) neighborhood scale,
showing the probability of the focal pool receiving dispersing animals from each
neighboring pool; and (c) regional scale, with dark outline indicating pools that are
interconnected by a specified level of dispersal. Darker shading indicates greater
connectivity at each scale (reprinted with permission from Blackwell Publishing).
Many complementary indices for quantifying habitat fragmentation, habi-

tat loss, connectivity, or other landscape characteristics have been developed.
McGarigal et al. (2002) developed a computer software program (FRAGSTATS)
capable of computing a multitude of landscape metrics for categorical maps,
such as land cover or habitat maps. The software and documentation are avail-
able for free at www.umass.edu/landeco/research/fragstats/fragstats.html.
19.2.4 Landscape permeability

Landscape permeability refers to the resistance imposed by a habitat to move-
ment. Permeability is usually associated to the costs incurred by an individual
while using a particular type of habitat. At a population or metapopulation level
dispersing individuals are essential for providing new colonists and maintaining
typical levels of gene flow (Marsh and Trenham 2001). Any additional dispersal
cost induced by a less permeable habitat matrix reduces the probability of pond
recolonization in the case of human-induced or naturally occurring extinction.
Because most amphibians use terrestrial habitat for migration and dispersal,
any change in the configuration and composition of the landscape surrounding
the breeding pools potentially increases the costs for individual- and population-
scale processes. At the level of the individual amphibian, low permeability can
be reflected in longer exposure to inhospitable conditions. For example, low
permeability may be associated with higher mortality risk due to desiccation
in a clear-cut versus a closed forest, higher predation risk, or increased traveling
time (and hence energy expended) associated with avoidance of inhospitable
habitats.
Incorporating landscape permeability in the study of amphibian dispersal has
been attempted using two different approaches: field experiments (Rothermel
and Semlitsch 2002; Mazerolle and Desrochers 2005) and GIS modeling (Ray
et al. 2002; Joly et al. 2003; Compton et al. 2007; Figure 19.6). Field experi-
ments have explicitly measured the resistance opposed by low permeability
habitat (agricultural land, barren land) relative to high permeability habitat
(forest, wetland) to amphibian movement. Indices such as distance traveled,
rate of movement, survival, probability of homing successfully, or habitat choice
for movement were used for assessing habitat permeability empirically. This
is valuable information that has the potential to greatly improve attempts at
modeling amphibian dispersal. However, there are drawbacks of field experi-
ments pertaining to the scale of the experiment, species-specific behaviour and
vagility, and logistics. Measuring the permeability within 50 m dispersal arrays
(Rothermel and Semlitsch 2002) or 70 m distance from a pond (Mazerolle and
Desrochers 2005) is restrictive and does not reflect the extensive movements
that some amphibians perform. The methods for obtaining the data (i.e. pitfall
traps) may impede continuous movements and consequently not reflect habitat
permeability best. Dissimilarities in vagility, habitat selection, and physiology
between species, as well as age-specific differences, demand that such experi-
ments be conducted across multiple species. Also, such experiments are labor-
intensive and time-consuming, and the results cannot easily be translated to
another geographical locality. So-called friction or resistance maps provide a
potential GIS-based solution to modeling permeability realistically. The pro-
cess includes assigning land-use or land-cover classes various permeability values
based on expert opinion or empirical data, and using a GIS to create a least cost
surface expressing connectivity between populations or breeding sites. Ray et al.
(2002) used a land-use map to estimate potential migration zones for common
toads (Bufo bufo) and alpine newts (Triturus alpestris) in Switzerland. Similarly,
Joly et al. (2003) estimated biological connectivity of common toad (B. bufo)
populations in the Rhone floodplain, France, while accounting for potential
road mortality. Compton et al. (2007) used a statewide land-cover data set and a
combination of expert-based and empirical permeability data to model connect-
ivity for ambystomatid salamanders at three different scales in Massachusetts,
USA (Figure 19.6).
19.2.5 Landscape genetics

Individual amphibians move genetic material around the landscape as they dis-
perse to new areas and integrate themselves into new populations. Dispersal is
the “glue” that causes adjacent populations to coalesce genetically when they
would otherwise evolve in isolation, and steadily differentiate from one another
through random events (drift) or natural selection and adaptation to their local
environments. By disrupting amphibian movement, habitat fragmentation can
alter gene flow and hence determine how genetic variation is distributed among
populations in a landscape. Traditionally, population genetics has focused on
estimating the extent to which genetic diversity was distributed within versus
among populations; the spatially explicit forces shaping these patterns were
largely ignored. In response, the field of landscape genetics arose recently as
a means of relating genetic discontinuities to landscape features (Manel et al.
2003; Storfer et al. 2007).
Landscape genetics capitalizes on an emergence of spatial data and spatial
statistical techniques along with the capacity to assay genetic variation cheaply
and quickly, and with great resolution among of many populations (e.g. fine-
scale spatial structure of spotted salamanders, Ambystoma maculatum; Zamudio
and Wieczorek 2007). Advances in our ability to accumulate vast amounts of
fine-scale genetic data are largely permitted by the development of polymer-

ase chain reaction (PCR) methods that readily generate sufficient quantities
of DNA from field samples for detailed analysis of patterns of genetic variation
among individuals and populations (Chapter 22). Another attribute of land-
scape genetics is that genetic data are in some cases permitted to self-aggregate
through statistical patterning, thus no longer requiring geneticists to arbitrarily
define populations for sampling (Manel et al. 2003). GIS tools underpin the
entire process of linking genetic data to spatial environmental data that essen-
tially represents landscape genetics.
GIS-based landscape genetics studies for evaluating the functional connect-
ivity of the landscape have been successfully applied in the case of the redback
salamander (Plethodon cinereus) in Connecticut (Gibbs 1998b) and Virginia
(Cabe et al. 2007), natterjack toad (Bufo calamita) in southern Belgium (Stevens
et al. 2006), tiger salamander (Ambystoma tigrinum) in the western USA (Spear
et al. 2005), dwarf squeaker (Schoutedenella xenodactyloides) in Taita Hills,
Kenya (Measey et al. 2007), and spotted salamander (Ambystoma maculatum)
in central New York (Zamudio and Wieczorek 2007). The procedure generally
involves sampling distinct sites or populations and using high-resolution gen-
etic markers (mainly microsatellite loci) to analyze population genetic structure
(Chapter 22). Measures of genetic similarity between adjacent populations are
then linked in a GIS to spatial variables such as topographical distance between
sites/populations, elevation, and road/stream crossings to identify the envir-
onmental correlates of genetic similarity (and hence gene flow) between popu-
lations. Including habitat permeability (described above) has improved our
understanding, for example, when friction-based distances outperform straight
(Euclidian) distances in explaining the genetic differentiation (Stevens et al.
2006). Gibbs and Reed (2007) provide a review of genetics and landscape con-
nectivity for vernal pool-breeding amphibians.
19.3 Spatial statistics

Tobler’s First Law of geography is that “everything is related to everything else,
but near things are more related than distant things” (Tobler 1970). In other
words, data collected at any location in the landscape will have a greater simi-
larity to, or influence on, those locations within its immediate neighborhood.
Consequently, data tied to any point on the landscape is not independent but
rather spatially autocorrelated with the identity of its neighbors. Spatial auto-
correlation may lead to biased estimations of parameters and it needs to be
accounted for in the modeling process or spatial statistical techniques need to
be used. Legendre and Fortin (1989), Perry et al. (2002), and Dormann et al.
(2007) review the statistical tools for testing for the presence of spatial auto-
correlation in biological data and provide guidance for using the proper tech-
niques. Practical guidelines for addressing the issue of spatial autocorrelation
for designing of field surveys and data analysis are also provided by Legendre
et al. (2002).
Common means for analyzing spatial patterns of the data include indices
for global spatial autocorrelation (Moran’s I, Geary’s c), indices for spatial clus-
tering of group-level data (Getis–Ord Local G), and spatial autocorrelation
tests (Mantel and partial Mantel tests). Calculation of these indices and tests
can be conducted either using spatial statistics software (GSLIB, Gstat, GS,
VARIOWIN) or as packages implemented in a general statistical framework
with software (R, SAS, SYSTAT). Some GIS packages, such as ArcGIS (ESRI,
Redlands, CA, USA) and IDRISI (Clark Labs, Worcester, MA, USA), also con-
tain geostatistical tools to make these adjustments.
New modeling techniques that account for spatial autocorrelation in the
data, such as autologistic regression (Augustin et al. 1998) and geographic
weighted regression (Fotheringham et al. 2002) have been developed. For
example, Knapp et al. (2003) used autologistic regression to model probability
of pond occupancy in yellow-legged frogs (Rana muscosa) by incorporating an
autocovariate term describing the degree of isolation of ponds. As a result, the
autocovariate isolation term was the most important predictor of occupancy
in yellow-legged frogs. A similar approach was taken by Davidson (2004) for
explaining the relation between amphibian decline and historical pesticide use
in California. Dormann (2007) brings forth further evidence on the import-
ance of accounting for spatial autocorrelation: when modeling the distribution
of organisms including amphibians, reptiles, mammals, birds, plants, insects,
or mites, spatial autocorrelation biased the coefficient estimates, under- or over-
estimating by approximately 25%. Incorporating spatial autocorrelation also
improved model fit.
19.4 Limitations and future directions

One primary limitation of GIS applications to amphibian biology and conser-
vation is the chronic lag between need and availability of data with which to
populate a GIS. Landscapes are changing at rates faster than is feasible to update
of spatial information. Typically, there is a time lag of years from the acquisition
of raw satellite data to the availability of synthesized data products that support
landscape analyses. At a global scale, high-resolution spatial information may
not be available or updated for large continental areas for many years. Most
importantly, particularly data-deficient regions, such as tropical latitudes, with
the least amphibian research attention, are nevertheless of critical importance
for amphibian conservation.
Lack of a stable “taxonomy” for ecosystem classification also plagues GIS-
based amphibian research. Researchers seeking the latest spatial information
available typically analyze remotely sensed imagery (such as from Landsat,
SPOT, or IKONOS satellite sensors) and implement their own classification
of habitats or land use, often designed to suit the focus organism. This serves
a given study well but reduces the comparability among different studies with
respect to habitat use, availability, and selection by the same organism.
Inconsistent data resolution also creates problems. For example, habitat
changes critical for amphibian population persistence might occur at a finer
scale than satellite imagery can capture. Aerial photos may, however, be suffi-
cient for the task. Whatever the case, misinterpretation due to an inability to
resolve such patterns in some cases but not others may occur because of vari-
ation in coarseness of the spatial data available across studies. Not only spatial,
but also temporal and informational resolution add to the problems of GIS use.
The time lag between the publishing date of spatial data and the study period
imposes obvious restrictions on the ability to model habitat relationships.
“GIS-philia” can also blind some biologists to the inadequacy of their efforts.
GIS represents are thrilling mix of technologies for many amphibian researchers,
and resulting maps often have a visually stunning effect that can distract from
the maps’ limitations. Moreover, no matter how accurate and up-to-date the
spatial data, the animal–habitat correlations that are at the core of GIS analysis
for many amphibian studies are always only part of the puzzle. Sophisticated
analysis and modeling of spatial data require empirical data for model param-
eterization and calibration. Such data are also required for model validation, a
critical step increasingly neglected in many modeling studies. Gathering suffi-
cient field data to confront each model and thereby assess its validity should be
regarded as a requirement for publication of any modeling studies or dissemin-
ation of any guidelines extending from such models. This said, model validation
can be equally or more expensive than the cost of developing the original model,
a cost most researchers (and their funders) are rarely willing to cover.
Landscape thresholds have the potential to become a powerful tool in man-
aging amphibian populations because they offer specific management objectives
(e.g. “retain more than 70% forest within a 2 km radius pond neighborhood
to maintain a healthy population”). However, threshold research is still in its
infancy (Huggett 2005), and the temporal and spatial behaviour of thresholds
remains largely unstudied. Whether or not it is possible to translate thresholds

identified within a landscape at one spatio-temporal scale to another needs fur-
ther testing.
So much of what we can interpret from landscape-level data for amphibians
depends on the understanding of amphibian movement ecology, a topic that
remains poorly elaborated (Semlitsch 2008). We are only beginning to under-
stand the processes distinguishing migration from dispersal, let alone under-
standing how and why amphibians move about the landscape, particularly
juvenile forms, which typically represent the bulk of moving animals given the
short-life spans that characterize amphibian demographics. There is a clear need
for integrating basic amphibian ecology into modeling, especially more physio-
logical data explaining movement preferences, before we can predict with confi-
dence the implications of various landscape configurations and compositions.
In conclusion, habitat loss and degradation are now considered among the
greatest threats to amphibians worldwide (Cushman 2006). As once naturally
grading, heterogeneous landscapes are converted to fragmented mosaics with
sharp, “hard” boundaries that impede movements, landscape ecology and GIS
techniques are contributing significantly to understanding amphibian popula-
tion dynamics, including the extinction and colonization processes so critical
to sustaining amphibian populations. This understanding is slowly being par-
layed into useful guidelines for managing amphibian populations effectively
at the landscape scales at which most populations operate over the long term.
Our ability to master GIS software to harness increasingly rich, abundant, and
well-resolved spatial data will determine how well we realize the potentials of
the field of landscape ecology and thereby inform us about broad-scale spatial
processes that play such an important role in amphibian population persistence
everywhere.
19.5 References
Araújo, M. B., Thuiller, W., and Pearson, R. G. (2006). Climate warming and the decline
of amphibians and reptiles in Europe. Journal of Biogeography, 33, 1712–28.
Augustin, N. H., Mugglestone, M. A., and Buckland, S. T. (1998). The role of simulation
in spatially correlated data. Environmetrics, 9, 175–96.
Cabe, P. R., Page, R. B., Hanlon, T. J., Aldrich, M. E., Connors, L., and Marsh, D. M.
(2007). Fine-scale population differentiation and gene flow in a terrestrial salamander
(Plethodon cinereus) living in continuous habitat. Heredity, 98, 53–60.
Compton, B. W., McGarigal, K., Cushman, S. A., and Gamble, L. R. (2007). A resistant-
kernel model of connectivity for amphibians that breed in vernal pools. Conservation
Biology, 21, 788–99.
Cushman, S. A. (2006). Effects of habitat loss and fragmentation on amphibians: a review

and prospectus. Biological Conservation, 128, 231–40.
Davidson, C. (2004). Declining downwind: amphibian population declines in California
and historical pesticide use. Ecological Applications, 14, 1892–1902.
Denoël, M. and Ficetola, G. F. (2007). Landscape-level thresholds and newt conserva-
tion. Ecological Applications, 17, 302–9.
Dormann, C. F. (2007). Effects of incorporating spatial autocorrelation into the analysis
of species distribution data. Global Ecology and Biogeography, 16, 129–38.
Dormann, C. F., McPherson, J. M., Araujo, M. B., Bivand, R., Bolliger, J., Carl, G.,
Davies, R. G., Hirzel, A., Jetz, W., Kissling, W. D. (2007). Methods to account for
spatial autocorrelation in the analysis of species distributional data: a review. Ecography,
30, 609–28.
Ficetola, G. F., Thuiller, W., and Miaud, C. (2007). Prediction and validation of the
potential global distribution of a problematic alien invasive species - the American bull-
frog. Diversity and Distributions, 13, 476–85.
Fotheringham, A. S., Brunsdon, C., and Charlton, M. (2002). Geographically Weighted
Regression—the Analysis of Spatially Varying Relationships. Wiley, Chichester.
Gagné, S. and Fahrig, L. (2007). Effect of landscape context on anuran communities
in breeding ponds in the National Capital Region, Canada. Landscape Ecology, 22,
205–15.
Gamble, L. R., McGarigal, K., and Compton, B. W. (2007). Fidelity and dispersal in
the pond-breeding amphibian, Ambystoma opacum: Implications for spatio-temporal
population dynamics and conservation. Biological Conservation, 139, 247–57.
Gibbons, J. W. (2003). Terrestrial habitat: a vital component for herpetofauna of isolated
wetlands. Wetlands, 23, 630–5.
Gibbs, J. P. (1998a). Distribution of woodland amphibians along a forest fragmentation
gradient. Landscape Ecology, 13, 263–8.
Gibbs, J. P. (1998b). Genetic structure of redback salamander Plethodon cinereus popula-
tions in continuous and fragmented forests. Biological Conservation, 86, 77–81.
Gibbs, J. P. and Reed, J. M. (2007). Population and genetic linkages of vernal pool-
associated amphibians. In A. J. K. Calhoun and P. G. DeMaynadier (eds), Science and
Conservation of Vernal Pools, pp. 149–68. CRC Press, Boca Raton, FL.
Gibbs, J. P., Whiteleather, K. K., and Schueler, F. W. (2005). Changes in frog and toad
populations over 30 years in New York State. Ecological Applications, 15, 1148–57.
Gill, D. E. (1978a). Effective population size and interdemic migration rates in a metap-
opulation of the red-spotted newt, Notophthalmus viridescens (Rafinesque). Evolution,
32, 839–49.
Gill, D. E. (1978b). The metapopulation ecology of the red-spotted newt, Notophthalmus
viridescens (Rafinesque). Ecological Monographs, 48, 145–66.
Guisan, A. and Zimmermann, N. E. (2000). Predictive habitat distribution models in
ecology. Ecological Modeling, 135, 147–86.
Hanski, I., Kuussaari, M., and Nieminen, M. (1994). Metapopulation structure and
migration in the butterfly Melitaea cinxia. Ecology, 75, 747–62.
Hecnar, S. J. and M’Closkey, R. T. (1998). Species richness patterns of amphibians in
southwestern Ontario ponds. Journal of Biogeography, 25, 763–72.
Hels, T. (2002). Population dynamics in a Danish metapopulation of spadefoot toads

Pelobates fuscus. Ecography, 25, 303–13.
Herrmann, H. L., Babbitt, K. J., Baber, M. J., and Congalton, R. G. (2005). Effects of
landscape characteristics on amphibian distribution in a forest-dominated landscape.
Homan, R. N., Windmiller, B. S., and Reed, J. M. (2004). Critical thresholds associated
with habitat loss for two vernal pool-breeding amphibians. Ecological Applications, 14,
1547–53.
Huggett, A. J. (2005). The concept and utility of ‘ecological thresholds’ in biodiversity
conservation. Biological Conservation, 124, 301–10.
Joly, P., Miaud, C., Lehmann, A., and Grolet, O. (2001). Habitat matrix effects on pond
occupancy in newts. Conservation Biology, 15, 239–48.
Joly, P., Morand, C., and Cohas, A. (2003). Habitat fragmentation and amphibian con-
servation: building a tool for assessing landscape matrix connectivity. Comptes Rendus
Biologies, 326, 132–9.
Knapp, R. A., Matthews, K. R., Preisler, H. K., and Jellison, R. (2003). Developing prob-
abilistic models to predict amphibian site occupancy in a patchy landscape. Ecological
Knutson, M. G., Sauer, J. R., Olsen, D. A., Mossman, M. J., Hemesath, L. M., and
Lannoo, M. J. (1999). Effects of landscape composition and wetland fragmentation
on frog and toad abundance and species richness in Iowa and Wisconsin, U.S.A.
Legendre, P. and Fortin, M.-J. (1989). Spatial pattern and ecological analysis. Plant
Ecology, 80, 107–38.
Legendre, P., Dale, M. R. T., Fortin, M.-J., Gurevitch, J., Hohn, M., and Myers, D.
(2002). The consequences of spatial structure for the design and analysis of ecological
field surveys. Ecography, 25, 601–15.
Lemckert, F. (2004). Variations in anuran movements and habitat use: implications for
conservation. Applied Herpetology, 1, 165–81.
Löfvenhaft, K., Runborg, S., and Sjögren-Gulve, P. (2004). Biotope patterns and amphib-
ian distribution as assessment tools in urban landscape planning. Landscape and Urban
Planning, 68, 403–27.
Manel, S., Schwartz, M. K., Luikart, G., and Taberlet, P. (2003). Landscape genetics:
combining landscape ecology and population genetics. Trends in Ecology and Evolution,
18, 189–97.
Marsh, D. M. and Trenham, P. C. (2001). Metapopulation dynamics and amphibian
conservation. Conservation Biology, 15, 40–9.
Mazerolle, M. J. and Desrochers, A. (2005). Landscape resistance to frog movements.
Canadian Journal of Zoology, 83, 455–64.
McGarigal, K., Cushman, S. A., Neel, M. C., and Ene, E. (2002). FRAGSTATS: Spatial
pattern analysis program for categorical maps. University of Massachusetts, Amherst, MA.
Measey, G. J., Galbusera, P., Breyne, P., and Matthysen, E. (2007). Gene flow in a direct-
developing, leaf litter frog between isolated mountains in the Taita Hills, Kenya.
Conservation Genetics, 8, 1177–88.
Moore, G. E. (1975). Progress in digital integrated electronics. Electron Devices Meeting

1975 International, 21, 11–13.
Naveh, Z. and Lieberman, A. (1994). Landscape Ecology. Theory and Application, 2nd edn.
Springer-Verlag, New York.
Pellet, J., Hoehn, S., and Perrin, N. (2004). Multiscale determinants of tree frog (Hyla
arborea L.) calling ponds in western Switzerland. Biodiversity and Conservation, 13,
2227–35.
Perry, J. N., Liebhold, A. M., Rosenberg, M. S., Dungan, J., Miriti, M., Jakomulska, A.,
and Citron-Pousty, S. (2002). Illustrations and guidelines for selecting statistical methods
for quantifying spatial pattern in ecological data. Ecography, 25, 578–600.
Pope, S. E., Fahrig, L., and Merriam, H. G. (2000). Landscape complementation and
metapopulation effects on leopard frog populations. Ecology, 81, 2498–2508.
Ray, N., Lehmann, A., and Joly, P. (2002). Modeling spatial distribution of amphibian
populations: a GIS approach based on habitat matrix permeability. Biodiversity and
Richter-Boix, A., Llorente, G. A., and Montori, A. (2007). Structure and dynamics of an
amphibian metacommunity in two regions. Journal of Animal Ecology, 76, 607–18.
Rothermel, B. B. and Semlitsch, R. D. (2002). An experimental investigation of land-
scape resistance of forest versus old-field habitats to emigrating juvenile amphibians.
Schmidt, B. R. and Pellet, J. (2005). Relative importance of population processes and
habitat characteristics in determining site occupancy of two anurans. Journal of
Wildlife Management, 69, 884–93.
Semlitsch, R. D. (2000). Principles for management of aquatic-breeding amphibians.
Journal of Wildlife Management, 64, 615–31.
Semlitsch, R. D. (2008). Differentiating migration and dispersal processes for pond-
breeding amphibians. Journal of Wildlife Management, 72, 260–7.
Sjögren-Gulve, P. (1994). Distribution and extinction patterns within a northern metap-
opulation of the pool frog, Rana lessonae. Ecology, 75, 1357–67.
Smith, A. M. and Green, D. M. (2005). Dispersal and the metapopulation paradigm in
amphibian ecology and conservation: are all amphibian populations metapopulations?
Ecography, 28, 110–28.
Spear, S. F., Peterson, C. R., Matocq, M. D., and Storfer, A. (2005). Landscape genetics of
the blotched tiger salamander (Ambystoma tigrinum melanostictum). Molecular Ecology,
14, 2553–64.
Stevens, V. M., Polus, E., Wesselingh, R., Schtickzelle, N., and Baguette, M. (2005).
Quantifying functional connectivity: experimental evidence for patch-specific resist-
ance in the Natterjack toad (Bufo calamita). Landscape Ecology, 19, 829–42.
Stevens, V. M., Verkenne, C., Vandewoestijne, S., Wesselingh, R. A., and Baguette,
M. (2006). Gene flow and functional connectivity in the natterjack toad. Molecular
Ecology, 15, 2333–44.
Storfer, A., Murphy, M. A., Evans, J. S., Goldberg, C. S., Robinson, S., Spear, S. F.,
Dezzani, R., Delmelle, E., Vierling, L., and Waits, L. P. (2007). Putting the “land-
scape” in landscape genetics. Heredity, 98, 128–42.
Tischendorf, L. and Fahrig, L. (2000). How should we measure landscape connectivity?

Landscape Ecology, 15, 633–41.
Tobler, W. R. (1970). A computer movie simulating urban growth in the Detroit region.
Economic Geography, 46, 234–40.
Trenham, P. C., Koenig, W. D., Mossman, M. J., Stark, S. L., and Jagger, L. A. (2003).
Regional dynamics of wetland-breeding frogs and toads: turnover and synchrony.
Ecological Applications, 13, 1522–32.
Turner, M. G., Gardner, R. H., and O’Neill, R. V. (2001). Landscape Ecology in Theory
and Practice: Pattern and Process. Springer, New York.
Urban, M. C., Phillips, B. L., Skelly, D. K., and Shine, R. (2007). The cane toad’s
(Chaunus [Bufo] marinus) increasing ability to invade Australia is revealed by a dynam-
ically updated range model. Proceedings of the Royal Society of London Series B Biological
Sciences, 274, 1413–19.
Vos, C. C. and Stumpel, A. H. P. (1995). Comparison of habitat-isolation parameters in
relation to fragmented distribution patterns in the tree frog (Hyla arborea). Landscape
Ecology, 11, 203–14.
Werner, E. E., Yurewicz, K. L., Skelly, D. K., and Relyea, R. A. (2007). Turnover in
an amphibian metacommunity: the role of local and regional factors. Oikos, 116,
1713–25.
Wiens, J. A. (1989). Spatial scaling in ecology. Functional Ecology, 3, 385–97.
Wiens, J. A., Stenseth, N. C., Horne, B. V., and Ims, R. A. (1993). Ecological mechanisms
and landscape ecology. Oikos, 66, 369–80.
With, K. A. and Crist, T. O. (1995). Critical thresholds in species’ responses to landscape
structure. Ecology, 76, 2446–59.
Zamudio, K. R. and Wieczorek, A. M. (2007). Fine-scale spatial genetic structure and
dispersal among spotted salamander (Ambystoma maculatum) breeding populations.
Molecular Ecology, 16, 257–74.
Zanini, F., Klingemann, A, Schlaepfer, R., and Schmidt, B. R. (2008). Landscape effects
on anuran pond occupancy in an agricultural countryside: barrier-based buffers predict
distributions better than circular buffers. Canadian Journal of Zoology, 88, 692–9.
Part 6
Physiological ecology and genetics
20
Physiological ecology: field
methods and perspective
Harvey B. Lillywhite
20.1 Introduction
Both public and scientific awareness of the phenomenon of “amphibian declines”
(Blaustein and Wake 1995) have focused the importance of understanding the
physiological underpinnings of amphibian diversity, distribution, and survival
in varied and challenging environments. Contributions to understanding of
ecology have been important goals of physiologists and integrative biologists
interested in the structure, function, and behavior of amphibians (Feder and
Burggren 1992). Such studies are rich in history and possibilities considering
the biological diversity and evolutionary responses of amphibians to varied and
extensive environments. These range from caves and subterranean burrows to
high tree canopies, from forests to scrub and deserts, and from tropical lowlands
to mountain tops. For many species complex life cycles connect developmental
stages with both aquatic and terrestrial environments.
Scientific interest in the functional attributes of frogs began at least as
early as 1671 when Malpighi used frogs as a model for investigating kidney
function. This was followed by research in which amphibians were used for
investigating various aspects of physiology from neuromuscular function to
development and endocrinology. Frogs became familiar animals in college
and high-school anatomy and physiology classes. However, in both scientific
and education endeavors, the “amphibian” was simply a means to understand-
ing physiological principles. Publications by G. Kingsley Noble (1931) and
John Moore (1964) began to focus attention on the physiology of amphibians
as a means to understanding their biology. The development of physiological
ecology as an established discipline in the 1960s evoked parallel emphasis on
the biology of amphibians specifically in contexts of adaptation to environ-

ment (e.g. Ray 1958; Gordon et al. 1961; Hutchison 1961; Brattstrom 1963;
Thorson 1964; Warburg 1965; Whitford and Hutchison 1965; Bentley 1966;
Lee 1968; Lillywhite 1970).
There are more than 6400 species of extant amphibians, represented by
three orders (Anura, the frogs; Caudata, the salamanders; and Gymnophiona,
the caecilians) and member species that are morphologically and physiolo-
gically different from other aquatic or terrestrial vertebrates. Field methods
have been centrally important in helping us to understand how amphibians
live and reproduce in varied habitats while meeting the challenges of dehy-
dration, thermal extremes, energy balance, metabolic requirements, and life-
history transitions from aquatic to terrestrial environments. Much remains
to be learned about these important vertebrates, especially with respect to
phylogenetic variation and challenges of environmental change. Important
advances will be tied to key technological innovations related to miniaturiza-
tion, remote sensing, and isotopic, genetic, and chemical analysis.
20.1.1 Important features of amphibians

Certain features of amphibians are characteristic and generally set them apart
from most other tetrapods. These include:
• lack of a cleidoic egg, with early developmental stages limited to aquatic or
moist environments;
• integument with comparatively few layers of keratin in the stratum cor-
neum, and usually high permeability to water and respiratory gases;
• a complex life cycle in many species;
• strict ectothermy, with little evidence for incipient endothermy related to
large mass, activity, or other muscle activation.
All of these features help to defi ne what is “amphibian” and are centrally
important to the ecology and conservation of amphibian diversity. The
dynamic structure and function of skin are central to forcing functions that
influence the abundance and distribution of individuals or populations.
Attributes of skin are important not only with respect to temperature regu-
lation, water relations, and gas exchange, but are also relevant to current
evaluations of amphibian declines in the context of disease susceptibility
and transmission, impacts of xenobiotics, and sensitivity of responsiveness
to climate change. Increasingly, these will become the subjects of future field
and laboratory research by herpetologists who are interested in amphibian
diversity and conservation.
20 Physiological ecology | 365
20.2 Heat exchange, body temperature, and

thermoregulation
Research related to the ecophysiology of amphibians has tended to focus on
energetics, temperature, and water balance, but traditional single-focused stud-
ies have evolved to include increasing integration with other facets of biology.
Present understanding of the important role that temperature plays in the lives
of amphibians began largely with observations of behavior, which in turn led
to inferences and hypotheses of significance that could be tested by measure-
ment and experiments, both in the laboratory and in the field. Amphibians are
not as well studied as are fishes and reptiles. However, the data available should
be viewed by researchers in context of a voluminous, phylogenetically broader
literature on thermal biology. This rich literature stems partly from the ease of
temperature measurement. Knowledge of the pervasive importance of tempera-
ture as a principal ecological factor that influences almost all biochemical and
physiological processes, including behavior, is an important starting point for
nearly all field studies of amphibian biology.
Anyone interested in the physiological ecology of amphibians must begin
with appreciation for three things. First, the skin of amphibians couples them
to their physical environment by the exchanges of energy, water, and respiratory
gases. One must understand the physics of heat and water transfer between an
amphibian and its environment to understand the thermal relations and ecology
of these organisms. The implications of the interaction between heat and mass
transfer for the environmental physiology of amphibians have been discussed by
Tracy (1976) and Spotila (Spotila 1972; Spotila et al. 1992). Second, phenotypic
responses to temperature can be plastic, such that attributes of morphology,
physiology, and behavior are subject to reversible adjustments to temperature
(Rome et al. 1992). Thus, the role of acclimatization is essential to understand-
ing the thermal relationships of amphibians. Third, amphibians are a very
diverse group of animals (Duellman and Trueb 1986), and one must appreciate
that attributes and responses to variable microhabitats cannot be expected to
be uniform across species. One should approach studies of amphibians with an
open mind and without a narrow bias or expectation of stereotypic responses.
20.2.1 Thermal acclimation of physiological function

Some amphibians may undergo temperature changes of 30–35°C in relation to
daily or seasonal changes, respectively (Rome et al. 1992). Temperature influ-
ences all chemical, physical, and physiological processes, and it is recommended
that persons conducting field studies of amphibians be very familiar with what is
known concerning the principles and patterns of temperature effects on amphib-

ians. In particular, temperature affects energetics, muscle performance, behav-
ioral activities, digestion, water exchange, reproduction, immune function, and
susceptibility to disease and predation. To variable extents, amphibians can
adjust by means of acclimation the sensitivity and responses of processes to tem-
perature (Rome et al. 1992; Hutchison and Dupré 1992). The extent of devel-
opmental plasticity and adult phenotypic adjustments induced by environment,
in addition to genetic adaptation, are both important to amelioration of stressful
changes in environmental conditions, and hence distribution, population struc-
ture, and conservation.
20.2.2 Early methods in field studies of

amphibian thermal relations
Insight and understanding of the thermal ecology of amphibians arose initially
from observations of behavioral activities such as basking (adult anurans; e.g.
Brattstrom 1963; Lillywhite 1970) and thermal aggregations (anuran tadpoles;
e.g. Brattstrom 1962), in addition to measurements of body temperatures of
animals in the field. Early summaries of detailed temperatures of free-living
amphibians and their environments were published by Fitch (1956) and
Brattstrom (1963). Published data on random measurements of amphibian
body temperatures typically show a unimodal distribution, but interpretation
of such data requires considerable caution unless other detailed information
is available. Importantly, conclusions about thermoregulatory abilities require
information about microenvironmental conditions that are available to animals
and can be limited by time of day, season, and other factors. Moreover, because
of the high evaporative rates characteristic of many amphibians, requirements
for hydroregulation can compete with those for thermoregulation. And, as
pointed out by Tracy (1976), the relative stability of body temperature need not
represent or imply thermoregulation. Conclusions from the earlier literature
were that some amphibians exercised control over body temperature, but active
thermoregulation was not as well developed in amphibians generally as in other
ectotherms such as lizards.
Interest in the body temperatures of field-active ecothermic animals, stimu-
lated by the classic work of Cowles and Bogert (1944), created a need for simple
means of measuring body temperatures that were not cumbersome. Hence, the
Schultheis cloacal quick-reading thermometer was developed largely for field
use and has been used extensively by biologists for more than 50 years. The
Schultheis thermometer is a comparatively small (18 cm long) mercury therm-
ometer with a slim, reduced bulb that registers changes of temperature quickly.
The thermometer is calibrated in 0.2°C divisions within the range of 0–50°C.

It was manufactured originally by E.W. Schultheis and Sons of Brooklyn, New
York, USA. The thermometer is now available from Miller and Weber of Queens,
New York. In spite of easy breakage, this device made possible the rich data that
now exist on field temperatures of amphibians and reptiles. With due caution
against breakage, the thermometer has also been used to record temperatures
in the microhabitats of animals, but the upper limit of measurement imposes a
serious constraint on recording substrate temperatures in hot environments.
Cowles and Bogert (1944) used thermocouples to measure body temperatures
of desert reptiles in outdoor enclosures, but the equipment was cumbersome.
Subsequent development of small and easily portable instruments that reduced
the size of batteries and electronic components of thermistor and thermocouple
thermometers opened the way to a broader use of electronic thermometry in
the field (Table 20.1). In particular, multiple probes enable one to record tem-
peratures from a variety of microenvironment locations simultaneously without
the serial measurements required of hand-held thermometers. Moreover, the
quality of measurements have been further improved by rapid response time,
accuracy, miniaturization, and readability of sensing probes constructed of ther-
mistor or thermocouple sensors. These and infrared thermometers can also be
used to measure surface temperatures of amphibians (Rowley and Alford 2007).
Infrared probes are non-invasive and do not require manipulation of animals.
However, accurate ones are expensive and require placing the sensor close to the
animal, which can pose difficulties, especially for small species.
20.2.3 Precautions for temperature measurements

To ensure that spot measurements of body temperatures are meaningful, the
investigator must have knowledge of the microenvironment, behavior, and pref-
erably prior immediate history of the animal. This also necessitates knowledge
and understanding of the biophysics of heat exchange as it applies to amphibians
(Tracy 1976; Spotila et al. 1992). Even so, there are precautions that must be
taken to validate the accuracy and precision of such body-temperature measure-
ments.
First and foremost, the investigator needs to guard against transfer of heat
between the animal and the investigator’s hand, and between the animal and
substrate or object on which it is placed during measurement, especially if the
animal is small. Additionally, escape movements and attendant stress can alter
the body temperature, and the skin may cool quickly if the animal is moved
into a convective environment during measurement. Manipulation of smaller
amphibians that causes only a few seconds delay can alter body temperatures
Table 20.1 Transducers and monitoring and logging devices that are useful for deployment in physiological field research with amphibians.
Device Useful properties
Thermocouples Electronic measurement of temperatures with fine wire probes and sensors; available with sensors <1 mm diameter. Battery
power is available for field portable thermocouple thermometers. See also infrared thermometers.
Thermistors Similar uses and advantages as thermocouples, but with non-linear thermal characteristics that can be unstable over time.
Infrared thermometers Non-contact temperature measurement device that detects emitted infrared energy and converts this into a temperature
reading. Critical considerations include field of view (target size and distance), spectral response, response time, signal
processing, portability, hand-held options, and temperature range. Most infrared thermometers cover environmental and
body temperatures that might be encountered in field studies of amphibians. Infrared thermocouples are small, low-cost
infrared sensors that are self-powered and produce an output that mimics a thermocouple sensor. Infrared imagery may have
varied roles in research with amphibians.
Transmitters Smallest and lightest, 0.35–0.52 g total encapsulated transmitter with antenna and battery, for tracking movement. Crystal
controlled two-stage design, pulsed by a multivibrator. Optional design available with pulse rate determined by temperature.
Several configurations available intended for glue-on applications, but could be modified for implantation depending on
ingenuity of the researcher. Larger units (0.39–3.8 g) can be supplied with helical or subcutaneous antenna. Medical grade
epoxy or wax can be used to protect implanted units.
Passive integrated Many types of implantable transponders are available for individual identification of animals from an implanted chip;
transponders (PITs) location of animals over short range (1–10 m) by use of a diode tag and harmonic radar; or transmission of a temperature
signal by passive acquisition of energy from an external electromagnetic field. No battery is required, and the small size
permits implantation in the body cavity or just beneath the skin. However, usefulness for identification and temperature
measurement is limited to very short distances (generally a few centimeters).
Dataloggers Available with multiple channels for logging multiple inputs in the field, or as mini- or micro-units that can be used in multiple
locations. Measurements include temperature, humidity, pressure, salinity, depth, and voltage. Microloggers can be used to
generate thermal or humidity maps of the environment. Underwater dataloggers measure and record water temperatures to
depths >305 m. Programmable with capability to record maximum/minimum or averaging. Sampling intervals 0.5 s–9 h. Up
to 5-year battery life. Numerous units available from multiple sources with advantages of low-cost, real-time operation,
programmable start time, reusable, and miniature size. Some units may be subject to calibration drift and should be
recalibrated at intervals or before and after use.
iButtons™ Measurement of temperature and/or humidity, recorded from built in sensors into internal memory with user-defined time
interval. Small, light-weight with stainless steel casing. A durable stainless steel package is resistant to dirt, dust, contaminants,
moisture, and shock. Reusable and relatively inexpensive, with lithium battery operation for at least 10 years without
maintenance. Units should be checked for calibration drift and should be recalibrated at intervals or before and after use.
Evaporimeters Various products are available to measure area-specific transepidermal evaporative water loss and related variables. Many of
these instruments are cumbersome, but some are portable with battery operation for field applications. The VapoMeter ®,
manufactured by Delfin Technologies, Kuopio, Finland, is a hand-held easily portable device developed for use with human
skin and can also be used with small animals. Various other products known as dynamic diffusion porometers, leaf porometers,
and leaf hygrometers are sold for use with plants and can measure conductance or resistance directly. Soil hygrometers and
ceramic plates are used to measure water potential of soils.
Note: various products are available from multiple vendors, which can be accessed via the Internet. Further specifications are available online, providing details that
are too numerous to be listed here.
1.5
Temperature increase (°C)

1.0
3.6 g
0.5
8.6 g
0
5 10 15 20 25 30
Manipulation time (s)
Fig 20.1 Increases in temperature of agar replicas of frogs (N 10) following

manipulation wherein two sizes of models were held in one hand during a given
time lapse to simulate the expected change of body temperature in small anurans if
they were handled similarly. Temperatures of the agar models were measured with
thermocouples inserted into their centers and ranged initially from 22 to 23C.
Redrawn from Navas and Araujo (2000) with permission.
significantly (Figure 20.1) (Navas and Araujo 2000). Ideally, the measurement
should be taken rapidly (within a matter of seconds) upon capture of the animal
while it is held loosely but firmly by means of an insulating material such as dry
cloth, cotton, or glove. Heat transfer can also be minimized if anurans are held
by a hind limb while resting upon the insulation. Typically, investigators have
inserted thermal sensors into the hindgut via entry at the cloaca, with the result
that the core temperature is represented by the rear mass of the body. I have
found it easier and quicker to insert the probe into the stomach via the mouth.
If the tip of the sensor or thermometer is applied gently to the mouth, anurans
usually open the mouth briefly when the probe can be inserted. This method
typically measures the core temperature at a more central location of the body
where the mass and thermal inertia is greatest. Insertion of the probe also is
quicker compared with the time used in locating and probing the much smaller
orifice of the cloaca. Finally, cloacal insertion often causes release of urine from
the urinary bladder, which in turn increases evaporative cooling near the site of
measurement.
Tolerance of broad thermal regimes may be characteristic of many species of
amphibians which might be subjects of investigation. Whatever instruments
of temperature measurement are utilized, they should be capable of measur-

ing temperatures within the thermal range potentially encountered, including
microhabitat temperatures. In consideration of reproduction, development,
digestion, and activity, temperature extremes are important environmental fac-
tors and should be incorporated in ecological studies. With appropriate mod-
ifications, standard thermometers as well as electronic thermometers can be
used to measure dry- and wet-bulb air temperatures, blackbody temperature
(as an index of solar and thermal radiation), substrate, subterranean, and water
temperatures. Whatever the method that is chosen for determining body tem-
peratures of amphibians, quantitative indices have been described for evaluat-
ing the “effectiveness” of temperature regulation in these or other field-active
ectotherms (Hertz et al. 1993).
20.2.4 Telemetry
Radiotelemetry offers many advantages over the use of thermometers. Body tem-
peratures can be logged or recorded continuously or at chosen intervals without
disturbance to the animal that is being monitored. The first use of radioteleme-
try in field studies of amphibian body temperature was that of Lillywhite (1970)
who used temperature-sensitive transmitters to record internal body tempera-
tures of bullfrogs at shorelines and shallow water of ponds. Currently, a variety
of transmitter designs, sizes, and other features are available commercially, but
the options are generally not useful for studies involving smaller amphibian
species (
5 g). Where miniaturization is a requirement, dataloggers and
iButtons™ have proven to be useful (Table 20.1). Micro-dataloggers can be
used effectively to continuously monitor temperatures from both animals and
their microhabitats. They can generate a thermal map of the environment, use-
ful for understanding thermal selection behavior and constraints on physiology,
and they can monitor temperatures of real animals or physical models in simi-
lar contexts. Available temperatures can have significant effects on amphibian
activity, metabolism, muscle performance, and developmental rates. In extreme
environments the available microhabitats can quickly reach lethal temperatures,
and species must evolve strategies for avoiding thermal incapacitation or death.
For short-term measurements the unit can be simply fed to the animal (or
physically lodged in the stomach), but longer-term studies require surgical
implantation of the unit within the body cavity. Care must be taken to avoid
infection and to allow adequate recovery from the surgery before an animal is
released into the environment. Surgical implantation also avoids loss of the trans-
mitter, interference with feeding, and postprandial temperature elevation that
are possibly associated with units that are placed within the gut. Implantation
procedures should incorporate due caution to avoid cutting the intestine and
accidentally implanting units within the gut from which they are subsequently
passed. The user must consider trade-offs between range, life of operation, and
size, where the latter is determined largely by the mass of the batteries. It is rec-
ommended that transmitters or other implanted devices not exceed 3–5% of the
animal’s body mass. The use of passive integrated transponders (PITs) allows
gathering of temperature signals from even small amphibians, but the range is
very short (Table 20.1).
20.3 Water relations

Adequate levels of body water are essential to living organisms, and the mainten-
ance of adequate hydration is especially challenging for amphibians (Warburg
1965; Shoemaker 1988; Tracy et al. 1993; Lillywhite and Mittal 1999; Bartelt
2000; Lillywhite and Navas 2006). Many species are subject to rapid dehydra-
tion and require water for the completion of complex life cycles; therefore, these
vertebrates are especially sensitive to environmental changes that impact the
amount and distribution of free water in time and in space.
20.3.1 Cutaneous water exchange: terrestrial environments

One of the more uniquely important features of amphibian morphology is the
highly permeable integument that characterizes numerous species. The features
of skin that are relevant to this discussion include a stratum corneum with only
one or two layers of keratinized cells and the absence of lamellar bodies, lipo-
genic organelles which in amniotes give rise to a highly structured “permeability
barrier” for water (Lillywhite 2006). Lacking these protective features, many
species of amphibians lose water across the skin at rates similar to, or approach-
ing, that of a free water surface. This is the principal underpinning of the well-
known fact that dehydration is potentially a serious problem for many amphibian
species living in terrestrial environments. The skin may offer little resistance to
evaporative water loss, and consequently many species survive only by means of
evolved behaviors and geographic distributions that allow them to remain near
water or in relatively moist microenvironments (Child et al. 2008).
Early studies pioneered by Overton (1904), Adolph (1933), Thorson (1955),
and others suggested that “typical” anurans evaporate water at rates similar to
that of a free water surface. This initial impression, now perpetuated within a
considerable literature, was based largely on studies of ranid and bufonid anurans.
Authors of various publications and textbooks still tend to suggest or state that
“most” amphibians have little or no skin resistance to cutaneous evaporative
water loss. However, research during the past 40 years has shown that not all
amphibians evaporate water in this manner (reviews in Toledo and Jared 1993;
Lillywhite 2006).
Since the early 1970s a substantial number of anuran species have been
found to occupy very arid habitats and to withstand xeric conditions due to
plastic responses that modulate skin resistance (Shoemaker 1988; Navas et al.
2004). Arboreal species, in particular, exhibit specializations that enable them
to withstand drying conditions including exposure to full sun and low ambient
humidity. Key among these features are uricotelism and low rates of evapor-
ation attributable to superficial lipids that are secreted from specialized cuta-
neous glands and wiped over the outer skin surfaces of the body. The evolution
of uricotelism in xerophilic frogs only makes sense in the context of abilities of
animals to reduce their rates of cutaneous water loss simultaneously (Shoemaker
et al. 1972). Thus, a key adaptation is the ability of anurans to mitigate losses
of body water by means of creating an external lipid barrier to transepidermal
water loss (TEWL) (Lillywhite and Mittal 1999; Lillywhite 2006). Species with
a skin resistance to evaporative water loss exceeding about 10 s/cm have been
referred to as “waterproof” following the early publications of Loveridge (1970),
Shoemaker et al. (1972), and Blaylock et al. (1976). This designation, however,
is not literally accurate.
An alternative strategy for reducing losses of body water is the production of
“cocoons” involving multiple layers of shed epidermis when fossorial species are
subjected to dehydrating conditions while buried in soil (Lillywhite 2006). In
the case of both secreted cocoons and the supraepidermal lipids that are secreted
and wiped by arboreal frogs, subsequent function of the barrier depends on
immobility of the animal for otherwise the structural integrity of the barrier is
disturbed.
Understanding amphibian water balance in these animals requires careful
observations in the field combined with appropriate mechanistic investigations
of structure and function in the laboratory. Most physiological ecologists work-
ing with amphibians appreciate that only a small fraction of extant amphibian
species—clearly less than 2%—have actually been studied in these or related
contexts. Understanding the adaptive variation of skin resistance and its plasticity
is central to the role of water economy in limiting the distribution of amphibians
and their behavioral activities in time and in space. Increasingly, these studies will
become more integrative. Most animals require access to fresh water or a moist
substrate to replenish evaporative losses of body water, and specializations of ven-
tral skin in pelvic and abdominal regions facilitate this process behaviorally and
physiologically (Nagai et al. 1999). Water permeability generally varies directly
with gaseous permeability (Lillywhite and Mittal 1999), and the skin exchanges
respiratory gases and ions as well as water. Some amphibians rely exclusively on
the skin for respiratory gas exchange. Most studies have focused on these aspects
singly, but more integrative approaches are necessary to understand the mul-
tiple processes important to growth, distribution, and persistence of populations
(Lillywhite et al. 1973; Bartelt 2000; Child et al. 2008).
20.3.2 Cutaneous water exchange: aquatic environments

Numerous amphibian species are aquatic or spend considerable time in water
including aquatic larval life stages. Because the skin of amphibians is quite per-
meable in many species, the net movement of water across the skin reverses
direction when animals move between land and water. The skin surfaces of
submerged amphibians exchange dissolved gases, water, and ions, but the regu-
lation of these exchanges is complex owing to the numerous sites at which such
transfers can occur. Periodic losses of ions and water occur in urination, and
periodic exchanges of gases occur whenever aquatic amphibians surface to ven-
tilate lungs with air. The skin is not passive in these contexts, for the permeabil-
ity and exchange processes can be altered by changes in blood flow, respiratory
properties of blood, hormonal influence, extent of mucus covering the external
surfaces, behavioral disruption of external boundary layer, ionic transporters,
density of aquaporins, and other factors (Boutilier et al. 1992). Understanding
the physiological ecology of aquatic amphibians will continue to rely on integra-
tion of field and laboratory studies, but advances will depend to a large degree
on creative and diligent application of new methodologies and techniques in
field methods. Such methods and integration will also characterize studies of
estivation and hibernation (Pinder et al. 1992). It will be important to exam-
ine phylogenetically diverse taxa and diverse environmental conditions, as well
as the many aspects of physiology that change during development and might
differ between larval stages and adults (Ultsch et al. 1999). Finally, the aquatic
environment of the eggs of amphibians is essential to understanding the distri-
bution and persistence of species. Discussion of some of the interesting prob-
lems connecting physiology and ecology of eggs can be found in Seymour and
Bradford (1995) and in Kam et al. (1998).
20.4 Measuring water exchange

The water exchanges of amphibians in the field can be measured directly by
weighing frogs to quantify mass changes that include both cutaneous and pul-
monary components. The difference between mass and weight is negligible, and
such measurements can be easily converted to volume flux if changes are meas-
ured over time. If such data are part of experiments, animals are usually fasted
and measured with the bladder empty (‘standard’ mass: Ruibal 1962) so that
defecation and urination do not influence such measurements. Due to the low
rate of metabolism and highly permeable skin characteristic of most amphib-
ians, changes in mass over relatively short intervals (without ingestion or defe-
cation) are assumed to result from exchanges of water insofar as the component
due to metabolic losses of carbon will be negligibly small. Such measurements
can be used to estimate daily water turnover rates and water budgets for amphib-
ians. Due to the generally high water turnover rates of amphibians, isotopic
labeling procedures such as the use of tritiated water (Nagy 1975) cannot be
used without error. Urine flow rates have been used to estimate water turnover
in the aquatic Xenopus laevis (Henderson et al. 1972).
Regional measurements of cutaneous evaporative water loss can be measured
by use of a VapoMeter®, which is a portable, hand-held evaporimeter manufac-
tured by Delfin Technologies, Kuopio, Finland (Table 20.1). These measure-
ments all disturb the animal, but can be carried out in the field. To date, most
information about the evaporative properties of amphibian skin is derived from
laboratory studies (Toledo and Jared 1993; Lillywhite 2006).
20.5 Energetics
Chemical energy is important to amphibians, as the ability to maintain energy
balance is crucial to both survival and reproduction. Energy budgets can be
approached from perspectives of mass balance and are most meaningful when
viewed over broad time scales such as annual cycles. Feeding rates are related
to temperature, water balance, and seasonal and hormonal influences (e.g.
growth hormone). Energy assimilation efficiencies can vary greatly, but most
values average around 80–90%. Rates of metabolism (aerobic energy expend-
iture) are commonly estimated by manometry (pressure/volume changes) or
respirometry (concentration changes of O2 or CO2), but cannot be measured
in free-ranging amphibians by means of doubly labeled water technique due to
characteristically high rates of water turnover (Nagy 1975). Therefore, energy
data incorporated into amphibian energy budgets are based on laboratory esti-
mates of metabolic energy expenditure coupled with time allocations for dif-
ferent activities (e.g. Jorgensen 1988). During periods of restricted resources,
metabolic depression facilitates survival of amphibians by lowering levels of
energy expenditure below normal resting values (generally 20–30% of stand-
ard metabolic rate; Guppy and Withers 1999). Brumation, dormancy, or
inactivity are common responses of amphibians to cold or dehydrating condi-

tions, especially in desert environments.
20.6 Modeling amphibian–environment interactions

Tracking variations in body temperature, energy expenditure, and water bal-
ance is critical to understanding the ecology of amphibians and their responses
to environmental change. Many studies of ectotherms have emphasized
thermoregulation and heat exchange, but we now appreciate that hydroregula-
tion is equally important in amphibians (Tracy et al. 1993; Spotila et al. 1992).
Moreover, the permeable skin of most amphibians influences body temperature,
which is coupled to exchange of energy and water. Temperature and water bal-
ance interactively influence behavior, activity cycles, and habitat use. Such bio-
physical interactions potentially contribute to amphibian population declines
linked to microclimate directly (Davidson et al. 2001) or to epidemic disease
mediated by fungal parasites (Pounds et al. 1999, 2006).
20.6.1 Mathematical models

The transfer of heat and water between an amphibian and its environment occurs
by means of several processes and can be described by energy and water budget
equations based on physical principles. It is beyond the scope of this chapter to
review detailed descriptions of heat energy and water budgets, or the processes
involved, and the reader is referred to the reviews of Tracy and Spotila cited
above, and the references therein. Further discussion of modeling with respect
to amphibians can also be found in Chapter 21. Note that energy balance can
be modeled at thermal equilibrium or in dynamic states of so-called “transient
energy balance.” Construction of a steady-state climate space describes the tol-
erable conditions and limits of thermal environment within which equilibrium
body temperature is compatible with the physiological well-being of an animal
(Spotila et al. 1992). Climate spaces can illuminate the biophysical constraints
of the environment and are used to define the boundaries of microclimate in
which an animal must remain to function and survive.
The status of heat exchange at any given moment determines the body tem-
perature of an amphibian, which in many circumstances tends to be relatively
stable (Tracy 1976). Aquatic amphibians are usually close to the temperature
of surrounding water due to its high heat capacity and relatively rapid conduc-
tion to or from the animal. Heat exchange of fossorial amphibians is predom-
inantly by conduction, and body temperatures of animals such as burrowing
toads or caecilians are usually in equilibrium with surrounding soil. Terrestrial
amphibians are sensitive to radiation, convection, and conductive exchange

with the substrate. At high temperatures and low humidities, water evaporates
at high rates and the skin as well as body is subject to rapid dehydration. Hence,
the activities of amphibians often depend importantly on water storage and
whether losses of body water exceed the possible uptake of water from soil, wet
surfaces, ponds, or other sources. In relation to studies of population and com-
munity ecology, both thermal and hydric environments can be considered to
represent ecological resources (Magnuson et al. 1979).
The significance of mathematical models is threefold. First, they illuminate
the sensitivities of various processes in context of their relative importance in
determining the magnitude of heat and mass exchange and resulting equilib-
rium body temperatures. Second, they can be used to generate hypotheses about
amphibian interactions with their environment. Third, they have predictive
value to enhance understanding of behaviors, energetics related to population
growth, distributional limitations, and responses to climate change. On the
other hand, modeling efforts have limitations and can perform no better than
the assumptions and data that provide the inputs to the model. As example,
substrate temperature can vary greatly between shaded and sunlit areas, provid-
ing thermoregulatory opportunities for amphibians that might shuttle between
these two patches of microenvironment (Lillywhite 1970; Lillywhite et al.
1973). However, thermoregulation might appear to be much more constrained
if one does not consider substrate temperature as one of the parameters in heat-
balance models (Tracy 1976).
20.6.2 Physical models

Physical models are useful for simulating the thermal properties, and sometimes
evaporative properties, of an animal in steady state. Like mathematical models,
physical models integrate the physical attributes of a simulated animal with
heat exchange processes related to the microenvironment. Thus, models can be
placed at multiple locations and indicate the range of possible steady-state tem-
peratures available to an animal. When compared with actual body tempera-
tures of free-ranging animals, the models enable researchers to form conclusions
related to thermoregulatory behavior, selection of retreat sites, timing of activity,
and the range of possible environments occupied. Unlike mathematical models,
physical models can indicate the dynamics of heat-exchange processes in local
environments in real time.
Various objects have been used to represent amphibians in different thermal
environments. One difficulty in the use of model amphibians is deciding if the
animal should be represented as “wet-skinned” and therefore evaporates water
as would the species represented. Various investigators have used models of frogs
cast as agar (from alginate molds; Navas and Araujo 2000), plaster (O’Connor
1989), sponges (Hasegawa et al. 2005), and copper casts or tubes covered with
water-saturated cotton or cloth (Bradford 1984; Bartelt and Peterson 2005).
Care should be taken to ensure that the reflectance of the cloth or model surface
is similar to that of the animals being compared, and that evaporation rates are
not compromised by depletion of water associated with the model. Bartelt and
Peterson (2005) have developed a physical model with a datalogger and water
reservoir, which maintains a wet cloth sleeve over a copper model (Figure 20.2).
To date, physical models have represented either wet- or dry-skinned animals.
Agar replicas of anurans have been used as null models in studies of amphibians
having near-zero resistance to evaporative water loss, and these models have
been shown to exhibit rates of water evaporation and temperatures identical
to those of living frogs which they are intended to mimic (Navas and Araujo
2000). The use of models for investigating amphibians having an intermediate
resistance, or which vary the resistance of skin by wiping of lipids, still present
special challenges.
Copper tube
with cloth sleeve
EWL datalogger Te datalogger
Soil surface
(to battery pack)
Cloth sleeve and wick

Electronic
interface
250 ml water reservoir
Strain gauge
Plastic container
Fig 20.2 A physical model used to simulate the thermal and evaporative properties
of toads. The model records operative (equilibrial) temperature (Te) and evaporative
water loss (EWL), both tested for accuracy, limitations, and applicability. The model
assumes there is no physiological control over EWL and is designed specifically for
western toads, Anaxyrus ( Bufo) boreas, but can be used or modified to measure Te
and EWL in other species of amphibians. Modified after Bartelt and Peterson (2005)
with permission.
To understand the requirements for effective conservation or management

of amphibian populations and communities, biologists need to understand
how habitat structure and conditions of microenvironment affect thermal and
hydric states of amphibians simultaneously. Recently, Bartelt (2000) compared
operative temperatures and evaporative water losses of physical models with
actual body temperatures and evaporative rates of toads determined by their
natural behaviors in different habitats. The results suggest that habitats allowing
these animals to conserve body water as well as to achieve warm body tempera-
tures provide the more favorable conditions for the population. Future studies
that integrate physiology with environmental interactions will be increasingly
important to understanding the habitat requirements and conservation of
amphibian populations and communities.
In other studies, thermal equilibria and evaporative properties of skin have
been investigated using anurans themselves, either as live frogs with movement
restricted by caging (Tracy 1975) or as dead or paralyzed frogs that are placed in
locations of choice (Lillywhite et al. 1997, 1998). In these situations, care must
be taken to avoid excess dehydration of animals or their exposed skin surfaces,
for the evaporative properties of skin and the health of the animal changes if
the skin itself is allowed to dehydrate (Lillywhite 1975). Moreover, the use of
muscle-relaxing agents in field studies may not be allowed by animal-care-and-
use committees, although there is no harm or excessive stress to the animal if the
experiments are designed properly and executed carefully with due caution.
20.7 Other issues and future research directions

The subjects of temperature, water exchange, and energetics will continue to
be centrally important subjects of future investigations into the physiological
ecology of amphibians, but these will become increasingly integrated with other
aspects of biology especially in the context of conservation and biodiversity con-
cerns. Space does not permit detailed commentary on these other subjects, but
here I will mention some directions that will likely be part of future advances in
understanding amphibian biology based in field methods.
With respect to technology, we may expect to see increasing miniaturiza-
tion of telemetry systems and creative employment of their usage in a variety
of field contexts. Largely due to size limitations, there has been very little use
of datalogger packages that can record physiological variables related to blood
chemistry, gas exchange, cardiovascular function, muscle activity, and other
parameters in addition to temperature and location. These will be particularly
useful in contexts of understanding brumation (sensu Mayhew 1965), aquatic
life, and survival in extreme or changing environments (Lillywhite and Navas

2006). Use might also be made of osmotic pumps, which provide accurate and
reliable delivery of drugs or hormones in chronic infusion studies. Some newer
pumps are suitable for implantation in mice or small rats and could be adopted
for uses in amphibians related to nutrition, physiology, immunology, pharma-
cokinetic analysis, and other subjects of interest. Compact osmotic pumps can
be used in conjunction with telemetry devices or catheters, and the versatility in
drug delivery allows for behavioral and physiological analysis without disrupt-
ing study parameters.
The decline of amphibian populations in many parts of the globe beckons
increasing attention to physiological studies related to threats such as chemical
contaminants, emerging infectious diseases, overexploitation, habitat destruc-
tion, exotic species introductions, and climate change (Lips et al. 2008). Future
biophysical studies related to climate change should increase attention to climate-
linked hypotheses concerning reproduction, invasive disease outbreaks and their
spread across landscapes, transmission processes, and amphibian population
declines. Population dynamics and distribution of amphibians might be linked
with features of habitat that are subject to temporal variation and behavioral
choices related to physical interactions and physiological resources. Microbial
communities associated with amphibian skin might also respond to aspects
of biophysical exchange, pesticides, or other contaminants across gradients of
habitats or in relation to habitat disturbance and climate change. Recently, anti-
fungal skin bacteria have been implicated to form mutualistic relationships with
amphibian species by assisting protection from pathogenic fungi (Lauer et al.
2007). These may well turn out to be very complex areas of research, and future
investigations that might benefit from integration of physiology with amphib-
ian ecology will likely involve teams of trained researchers working in different
but complementary disciplines.
Stable isotopes are being used increasingly in ecological studies as indicators
of water and nutrient metabolism, trophic positions of consumers in food webs,
and potential energy and mass flow through ecological communities. Stable
isotopes can be measured from very small tissue samples and provide a measure
of energy and nutrients assimilated through trophic interactions by an organism
(Lajtha and Michener 1994). The many applications and uses of stable isotope
analysis have increased explosively during recent years, and their continued use
in ecological studies will undoubtedly increase. Stable isotope techniques have
not been widely used in studies of amphibians, but the methodology has much
to offer in the way of advancing ecological knowledge of this important group
of vertebrates.
The use of drugs and chemicals that are introduced into animals or the envir-
onment by investigators must be considered as a potential problem in future
herpetological research in context of unwanted and unknown or inadequately
studied effects on amphibians. Chemicals used as anesthetics, sedatives, chem-
ical markers such as dyes, hormones, and various pharmaceutical agents all have
potential cumulative, long-term effects on populations if used improperly or
without knowledge of potential problems. As an example, the anesthetic tric-
aine methanesulfonate, popularly known as MS-222, has been used widely as an
anesthetic in field studies of amphibians. A review of the literature indicates that
MS-222 was used for decades before field biologists understood its mechanisms
of action, and such uses potentially impact individual animals or populations by
means of increasing stress, impairing sensory perception, and interfering with
abilities of researchers to quantify parasite loads or diagnose bacterial infection
(Byram and Nickerson 2009). During studies of declining populations of hell-
benders (Cryptobranchus alleganiensis), thousands of these animals have been
exposed to MS-222, many of them multiple times. As a consequence, Byram
and Nickerson (2009) recommend that usage of MS-222 should be avoided
when possible. When the anesthetic is used, it should be properly buffered, and
the treated animals should be allowed adequate time and facilities for recovery.
Field biologists need to recognize that the use of this or other anesthetics for
immobilizing animals for marking, etc., can alter the animal’s physiology and
behavior as well as that of associated parasites and microbes. The authors also
suggest that herpetologists and field biologists work as a community to develop
humane and informed field anesthesia protocols. Such a recommendation would
apply also to use of any other chemical in field research to ensure that researchers
do not negatively impact the declining populations they might study.
Finally, it is important to comment briefly on the topics of reproduction and
development. The evolutionary success of amphibians ultimately depends on
their capabilities to reproduce, grow, and in many cases transition between dif-
ferent environments associated with metamorphosis. Amphibian growth and
reproduction is characterized by diversity of reproductive modes, plasticity, and
interplay between internal physiological cycles and external conditions of envir-
onment. The identification of factors in the environment that influence growth
and development will continue to be important, especially with respect to cli-
mate, environmental chemistry, and the quality of food and water. Methods for
measurement of egg properties, body mass, composition, and nutritional and
hormonal status in the field (e.g. Licht et al. 1983) will be important to valid-
ate conclusions based strictly in laboratory studies, which means that access,
timing, sampling effort, and other aspects of field investigation will need to
be planned carefully by investigators. The important interplay between field

and laboratory investigation will require careful integration as the role of endo-
crine system and other physiological controls are still not well understood
with respect to reproduction and development in many species of amphibians.
Insofar as most amphibians undergo “aquatic” embryonic and/or larval devel-
opment before transitioning to more terrestrial (hence aerial) environments,
survival depends on adaptation of structure and function—especially feeding,
locomotory, osmoregulatory, cardiorespiratory, and sensory systems—in both
environments during the lifetime of individuals (Burggren and Just 1992).
As with the many other topics related to physiological ecology of amphib-
ians, aspects of reproduction and development are complex, diverse, and chal-
lenging with respect to the varied and intense selective pressures facing extant
species. Integration of approaches to questions, methods, and data analysis will
be an important theme in current and future studies having sufficient quality
and novelty to advance understanding. Knowledge of the physiological ecology
of amphibians has much to offer persons who are interested in the conserva-
tion and management of amphibian populations. To quote a postscript from
Burggren and Just (1992): “Until we understand all the physiological responses
to the environment in all developmental stages, we cannot hope to understand
why a particular species thrives or disappears from the biosphere.”
20.8 References
Adolph, E. F. (1933). Exchanges of water in the frog. Biological Reviews, 8, 224–40.
Bartelt, P. E. (2000). A Biophysical Analysis of Habitat Selection in Western Toads (Bufo
boreas) in Southeastern Idaho. PhD dissertation, Idaho State University, Pocatello, ID.
Bartelt, P. E. and Peterson, C. R. (2005). Physically modeling operative temperatures and
evaporation rates in amphibians. Journal of Thermal Biology, 30, 93–102.
Bentley, P. J. (1966). Adaptations of amphibian to arid environments. Science, 152,
619–23.
Blaustein, A. R. and Wake, D. B. (1995). The puzzle of declining amphibian populations.
Scientific American, 272, 52–7.
Blaylock, L. A., Ruibal, R., and Plat-Aloia, K. (1976). Skin structure and wiping behavior
of phyllomedusine frogs. Copeia, 1976, 283–95.
Boutilier, R. G., Stiffler, D. F., and Toewes, D. P. (1992). Exchange of respiratory gases, ions,
and water in amphibious and aquatic amphibians. In M. E. Feder and W. W. Burggren
(eds), Environmental Physiology of the Amphibians, pp. 81–124. University of Chicago
Press, Chicago, IL.
Bradford, D. F. (1984). Temperature modulation in a high-elevation amphibian, Rana
muscosa. Copeia, 1984, 966–76.
Brattstrom, B. H. (1962). Thermal control of aggregation behavior in tadpoles.
Brattstrom, B. H. (1963). A preliminary review of the thermal requirements of amphib-

ians. Ecology, 44, 238–55.
Burggren, W. W. and Just, J. J. (1992). Developmental changes in physiological systems.
In M. E. Feder and W. W. Burggren (eds), Environmental Physiology of the Amphibians,
Byram, J. K. and Nickerson, M. A. (2009). The use of tricaine (MS-222) in amphibian
conservation. Reptile and Amphibian Conservation Corps, Occasional Papers in Reptile
and Amphibian Conservation no .1, 1–15.
Child, T., Phillips, B. L., and Shine, R. (2008). Abiotic and biotic influences on the dis-
persal behaviour of metamorph cane toads (Bufo marinus) in tropical Australia. Journal
of Experimental Zoology, 309A, 215–24.
Cowles, R. B. and Bogert, C. M. (1944). A preliminary study of the thermal requirements
of desert reptiles. Bulletin of the American Museum of Natural History, 83, 261–96.
Davidson, C. H., Shaffer, B., and Jennings, M. R. (2001). Declines of the California red-
legged frog: climate, UV-B, habitat, and pesticide hypotheses. Ecological Applications,
11, 464–79.
Duellman, W. E. and Trueb, L. (1986). Biology of Amphibians. McGraw-Hill Book Co.,
New York.
Feder, M. E. and Burggren, W. W. (eds) (1992). Environmental Physiology of the
Amphibians. University of Chicago Press, Chicago, IL.
Fitch, H. S. (1956). Temperature responses in free-living amphibians and reptiles of
northeastern Kansas. University of Kansas Publications of the Museum of Natural History,
8, 417–76.
Gordon, M. S., Schmidt-Nielsen, K., and Kelly, H. M. (1961). Osmotic regulation in the
crab-eating frog (Rana cancrivora). Journal of Experimental Biology, 38, 659–78.
Guppy, M. and Withers, P. C. (1999). Metabolic depression in animals: physiological
perspectives and biochemical generalizations. Biological Review, 74, 1–40.
Hasegawa, M., Suzuki, Y., and Wada, S. (2005). Design and performance of a wet sponge
model for amphibian thermal biology. Current Herpetology, 24, 27–32.
Henderson, I. W., Edwards, B. R., Garland, H. O. and Chester Jones, I. (1972). Renal
function in two toads, Xenopus laevis and Bufo marinus. General and Comparative
Endocrinology Supplement 3, 350–9.
Hertz, P. E., Huey, R. B., and Stevenson, R. D. (1993). Evaluating temperature regu-
lation by field-active ectotherms: the fallacy of the inappropriate question. American
Naturalist, 142, 796–818.
Hutchison, V. H. (1961). Critical thermal maxima in salamanders. Physiological Zoology,
34, 92–125.
Hutchison, V. H. and Dupré, R. K. (1992). Thermoregulation. In M. E. Feder and
W. W. Burggren (eds), Environmental Physiology of the Amphibians, pp. 206–49.
Jorgensen, C. B. (1988). Metabolic costs of growth and maintenance in the toad, Bufo
bufo. Journal of Experimental Biology, 138, 319–31.
Kam, Y.-C., Yen, C.-F., and Hsu, J.-L. (1998). Water balance, growth, development, and
survival of arboreal frog eggs (Chirixalus eiffingeri, Rhacophoridae): importance of egg
distribution in bamboo stumps. Physiological Zoology, 71, 534–42.
Lajtha, K. and Michener, R. H. (eds) (1994). Stable Isotopes in Ecology and Environmental
Science. Blackwell Scientific Publications, Oxford.
Lauer, A., Simon, M. A., Banning, J. L., Andre, E., Duncan, K., and Harris, R. N. (2007).
Common cutaneous bacterial from the eastern red-backed salamander can inhibit
pathogenic fungi. Copeia, 2007, 630–40.
Lee. A. K. (1968). Water economy of the burrowing frog, Heleiporus eyrei (Grey). Copeia,
1968, 741–5.
Licht, P., McCreery, B. R., Barnes, R., and Pang, R. (1983). Seasonal and stress related
changes in plasma gonadotropins, sex steroids, and corticosterone in the bullfrog, Rana
catesbeiana. General and Comparative Endocrinology, 50, 124–45.
Lillywhite, H. B. (1970). Behavioral temperature regulation in the bullfrog. Copeia,
1970, 158–68.
Lillywhite, H. B. (1975). Physiological correlates of basking in amphibians. Comparative
Biochemistry and Physiology, 52A, 323–30.
Lillywhite, H. B. (2006). Water relations of tetrapod integument. Journal of Experimental
Biology, 209, 202–26.
Lillywhite, H. B. and Mittal, A. K. (1999). Amphibian skin and the aquatic-terrestrial
transition: constraints and compromises. In A. K. Mittal, F. B. Eddy, and J. S. Datta
Munshi (eds), Water/Air Transitions in Biology, pp. 131–44. Oxford and IBH
Publishing, New Delhi.
Lillywhite, H. B. and Navas, C. A. (2006). Animals, energy, and water in extreme envi-
ronments: perspectives from Ithala 2004. Physiological and Biochemical Zoology, 79,
265–73.
Lillywhite, H. B., Licht, P., and Chelgren, P. (1973). The role of behavioral thermoregula-
tion in the growth energetics of the toad, Bufo boreas. Ecology, 54, 375–83.
Lillywhite, H. B., Mittal, A. K., Garg, T. K., and Agrawal, N. (1997). Wiping behav-
ior and its ecophysiological significance in the Indian tree frog Polypedates maculatus.
Copeia, 1997, 88–100.
Lillywhite, H. B., Mittal, A. K., Garg, T. K., and Das, I. (1998). Basking behavior,
sweating and thermal ecology of the Indian tree frog, Polypedates maculatus. Journal of
Lips, K. R., Diffendorfer, J., Mendelson, III, J. R., and Sears, M. W. (2008). Riding the
wave: Reconciling the roles of disease and climate change in amphibian declines. PLOS
Biology, 6, 441–54.
Loveridge, J. P. (1970). Observations on nitrogenous excretion and water relationships of
Chiromantis xerampelina (Amphibia, Anura). Arnoldia (Rhodesia), 5, 1–6.
Magnuson, J. J., Crowder, L. B., and Medvick, P. A. (1979). Temperature as an ecological
resource. American Zoologist, 19, 331–43.
Mayhew, W. (1965). Hibernation in the horned lizard, Phrynosoma m’calli. Comparative
Biochemistry and Physiology, 16, 103–19.
Moore, J. A. (1964). Physiology of the Amphibia. Academic Press, New York.
Nagai, T., Koyama, H., Hoff, K. V. S., and Hillyard, S. D. (1999). Desert toads discrim-
inate salt taste with chemosensory function of the ventral skin. Journal of Comparative
Neurology, 408, 125–36.
Nagy, K. A. (1975). Water and energy budgets of free-living animals: measurement using
isotopically labeled water. In N. F. Hadley (ed.), Environmental Physiology of Desert
Organisms, pp. 227–45. Dowden, Hutchinson and Ross, Stroudsburg, PA.
Navas, C. A. and Araujo, C. (2000). The use of agar models to study amphibian thermal
ecology. Journal of Herpetology, 32, 330–4.
Navas, C. A., Antoniazzi, M. M., and Jared, C. (2004). A preliminary assessment of
anuran physiological and morphological adaptation to the Caatinga, a Brazilian
semi-arid environment. In S. Morris and A. Vosloo (eds), International Congress
Series, vol. 1275, Animals and Environments, pp. 298–305. Proceedings of the Third
International Conference of Comparative Physiology and Biochemistry. Elsevier,
Cambridge.
Noble, G. K. (1931). The Biology of the Amphibia. McGraw-Hill, New York.
O’Connor, M. P. (1989). Thermoregulation in Anuran Amphibians: Physiology, Biophysics,
and Ecology. PhD dissertation, Colorado State University, Fort Collins, CO.
Overton, E. (1904). Neununddreissig Thesen über die Wasserökonomie der Amphibien
and osmotischen Eigenschaften der Amphibienhaut. Verhandlungen der Physikalisch-
medizineschen Gesselschaft zu Würzburg, 36, 277–95.
Pinder, A. W., Storey, K. B., and Ultsch, G. R. (1992). Estivation and hibernation. In
M. E. Feder and W. W. Burggren (eds), Environmental Physiology of the Amphibians,
Pounds, J. A., Fogden, M. P. L., and Campbell, J. H. (1999). Biological response to cli-
mate change on a tropical mountain. Nature, 398, 611–15.
Pounds, J. A., Bustamante, M. R., Coloma, L. A., Consuegra, J. A., Fogden, M. P.L.,
Foster, P. N., La Marca, E., Masters, K. L., Merino-Viteri, A., Puschendorf, R. et al.
(2006). Widespread amphibian extinctions from epidemic disease driven by global
warming. Nature, 439, 161–7.
Ray, C. (1958). Vital limits and rates of desiccation in salamanders. Ecology, 39,
75–83.
Rome, L. C., Stevens, E. D., and John-Alder, H. B. (1992). The influence of tem-
perature and thermal acclimation on physiological function. In M. E. Feder and
W. W. Burggren (eds), Environmental Physiology of the Amphibians, pp. 183–205.
Rowley, J. J. and Alford, A. R. (2007). Non-contact infrared thermometers can accurately
measure amphibian body temperatures. Herpetological Review, 38, 308–11.
Ruibal, R. (1962). The adaptive value of bladder water in the toad, Bufo cognatus.
Physiological Zoology, 35, 218–23.
Seymour, R. S. and Bradford, D. F. (1995). Respiration of amphibian eggs. Physiological
Zoology, 68, 1–25.
Shoemaker, V. H. (1988). Physiological ecology of amphibians in arid environments.
Journal of Arid Environments, 14, 145–53.
Shoemaker, V., Balding, D., Ruibal, R., and McClanahan, Jr, L. (1972). Uricotelism and
low evaporative water loss in a South American frog. Science, 175, 1018–20.
Spotila, J. R. (1972). The role of temperature and water in the ecology and distribution of
some lungless salamanders. Ecological Monographs, 42, 95–125.
Spotila, J. R., O’Connor, M. P., and Bakken, G. S. (1992). Biophysics of heat and mass
transfer. In M. E. Feder and W. W. Burggren (eds), Environmental Physiology of the
Amphibians, pp. 59–80. University of Chicago Press, Chicago, IL.
Thorson, T. B. (1955). The relationship of water economy to terrestrialism in amphibians.
Ecology, 36, 100–16.
Thorson, T. B. (1964). The partitioning of body water in Amphibia. Physiological Zoology,
37, 395–9.
Toledo, R. C. and Jared, C. (1993). Cutaneous adaptations to water balance in amphib-
ians. Comparative Biochemistry and Physiology, 105A, 593–608.
Tracy, C. R. (1975). Water and energy relations of terrestrial amphibians: insights
from mechanistic modeling. In D. M. Gates and R. M. Schmerl (eds), Perspectives in
Biophysical Ecology, pp. 325–46. Springer-Verlag, New York.
Tracy, C. R. (1976). A model of the dynamic exchange of water and energy between a ter-
restrial amphibian and its environment. Ecological Monographs, 46, 293–326.
Tracy, C. R., Christian, K. A., O’connor M. P., and Tracy, C. R. (1993). Behavioral
thermoregulation by Bufo americanus: the importance of the hydric environment.
Ultsch, G. R., Bradford, D. F., and Freda, J. (1999). Physiology: coping with the environ-
ment. In R. W. McDiarmid and R. Altig (eds), Tadpoles: The Biology of Anuran Larvae,
Warburg, M. R. (1965). Studies on the water economy of some Australian frogs. Australian
Whitford, W. G. and Hutchison, V. H. (1965). Gas exchange in salamanders. Physiological
Zoology, 38, 228–42.
21
Models in field studies of
temperature and moisture
Jodi J. L. Rowley and Ross A. Alford
21.1 Introduction
21.1.1 The importance of understanding the temperature
and moisture environments of amphibians
In ectothermic organisms, variation in body temperature directly affects fac-
tors such as rates of energy acquisition, growth, and reproduction (Shoemaker
and McClanahan 1975; McClanahan 1978). Temperature also exerts a strong
influence on ecological interactions such as predator–prey and host–pathogen
interactions, and changes in temperature can completely reverse the outcome
of such interactions at both the individual and population levels (Elliot et al.
2002; Woodhams et al. 2003). Although the body temperature of terrestrial
ectotherms is broadly correlated with environmental temperature, the actual
body temperatures of many species can differ considerably from macroenviron-
mental temperatures due to species-specific behavior, physiology, morphology,
and microenvironment use. As a result, ectotherms exposed to identical mac-
roenvironmental conditions can experience very different body temperatures
(Kennedy 1997).
The physiology and ecology of amphibians, and their thermal relations,
are also strongly influenced by moisture. The degree of evaporative water loss
(EWL) varies considerably within and among species (Shoemaker and Nagy
1977; Wygoda 1984; Buttemer et al. 1996; Young et al. 2005), and individuals
of many species are able to adjust their rates of water loss over relatively short
periods of time (Withers et al. 1982; Wygoda 1989a; Withers 1995; Tracy et al.
2008). The moisture and thermal environments, and interactions between them,
can constrain the performance of amphibians (Snyder and Hammerson 1993;
Tracy et al. 1993). However, the actual temperature and humidity experienced
by an amphibian can differ dramatically within a range of available macroen-
vironmental conditions, depending on microenvironment use (Schwarzkopf
and Alford 1996; Seebacher and Alford 2002; Rowley and Alford 2007a). To
understand how the environment is experienced by amphibians in the field,
both the thermal and moisture environments must be characterized as they are
experienced by amphibians.
Understanding the thermal and water relations of amphibians in the field
is important not only for understanding their biology, but also for conserva-
tion. Recent declines in amphibian populations around the world have been
attributed in part to the epidemic disease amphibian chytridiomycosis (Lips
et al. 2006). Changes in thermoregulatory opportunities available to rainforest
frogs may have contributed to their widespread declines in association with the
disease (Pounds et al. 2006). Laboratory experiments have shown that elevated
body temperatures, as can be produced by basking, can cure amphibians of the
disease (Woodhams et al. 2003).
Global warming is also likely to have a large impact on many species, with
macroenvironmental modeling suggesting that the distributions of many
ectothermic species will be dramatically altered (Thomas et al. 2004). This
approach is limited by the fact that the present distributions of many species
may not reflect their fundamental climatic requirements (Parmesan et al.
2005). Understanding those requirements may aid in refining such predictions
(Kennedy 1997; Parmesan et al. 2005).
There are two complementary approaches to collecting field data on tem-
perature and water relations. The first is to sample living animals. This can
involve single measurements from individuals (e.g. Snyder and Hammerson
1993; Navas 1996), or intensive monitoring via mark–recapture or tracking
(e.g. Schwarzkopf and Alford 1996; Seebacher and Alford 2002). Sampling
living animals typically results in a relatively small number of measurements
per unit of investigator effort. In addition, it often requires attaching tags to
animals or repeatedly handling them, both of which may alter normal behavior.
The second approach is to place physical models that mimic the thermal and
hydric properties of animals in the field. Because the thermal and hydric states
of many models can be monitored simultaneously, a large number of microenv-
ironments can be sampled intensively. The best approach is probably a com-
bination of techniques. Living animals provide information on behavior and
microhabitat use, which is used to inform the design of studies using models and
to compare with their results. Models are used to collect large volumes of data
21 Models in temperature and moisture studies | 389
that can provide an integrated picture of how the environment should affect
animals across space and time. Sampling of living animals is covered elsewhere
in this volume; here we focus on the collection and interpretation of data using
physical models.
21.1.2 Physical models of amphibians

Physical thermal models mimic the thermal properties of a particular ani-
mal in steady state, thereby producing an approximation of body tempera-
ture, which is often referred to as the operative environmental temperature,
although the original definition of operative environmental temperature
requires that it be measured using a model with no thermal inertia (Bakken
and Gates 1975). Relatively simple physical models made of substances such
as hollow or water-filled copper casts or even tubes can be used to mimic the
thermal properties of many reptiles (reviewed by Dzialowski 2005), and their
use has led to a detailed understanding of many aspects of reptilian thermal
biology (Dzialowski 2005).
Although such models are suitable for species with low and constant rates
of EWL, the higher and often variable rates of EWL that occur in many
amphibians complicate the design of physical models for these animals and
the interpretation of data collected using them. Many amphibians have cuta-
neous resistances to EWL of zero or near zero, although it is increasingly clear
that many anurans, particularly arboreal frogs, can have substantially higher
levels of resistance (Wygoda 1984; Buttemer et al. 1996; Young et al. 2005).
This is caused by mucus or lipid secretions (Shoemaker and McClanahan 1975;
McClanahan 1978; Christian and Parry 1997) or possibly by physical barriers
of dermal iridophores (Schmuck et al. 1988; Kobelt and Linsenmair 1992). In
such species, cutaneous resistance to EWL may at times be more than 100 times
greater than that of ‘typical’ ranid and bufonid frogs (Buttemer 1990; Buttemer
and Thomas 2003), with the skin of some frogs being as resistant to water loss
as that of terrestrial reptiles (McClanahan and Shoemaker 1987; Shoemaker
et al. 1987). In many species, levels of skin resistance are variable, depending on
the current physiological state and behavior of the individual (Toledo and Jared
1993; Lillywhite 2006; Tracy et al. 2008). The difficulty of accurately incorpor-
ating the effects of EWL has hindered the use of physical models in amphibian
thermal biology (Bartelt and Peterson 2005). In species with variable resistance
to EWL, measurements of thermal and water relations need to encompass the
range of states possible for a given set of environmental conditions. In this chap-
ter, we describe the use of models to measure this range of states directly.
21.2 Field models for investigating

temperature and moisture
21.2.1 Models with zero resistance to EWL
To allow EWL to occur, physical models of amphibians for use in the field
have been constructed as hollow copper tubes covered in wet fabric (Bradford
1984; Bartelt and Peterson 2005), periodically or continuously re-wetted plaster
(O’Connor and Tracy 1987; Tracy et al. 2007), dead or immobilized amphibians
(Wygoda 1989a, 1989b; Seebacher and Alford 2002), wet sponges (Hasegawa
et al. 2005), or agar casts of animals (Schwarzkopf and Alford 1996; Navas
and Araujo 2000). Most of these models lose water as free water surfaces, and
can therefore only be used to model the body temperatures of amphibians with
zero cutaneous resistance to EWL. The rates of EWL of sponge models can be
adjusted by controlling the moisture content of the sponges (Hasegawa et al.
2005), which may make them preferable for species with constant skin resist-
ances greater than zero. Modified agar models (Navas et al. 2002) may also be
used to mimic constant higher skin resistance. Some types of models can also
provide data on rates of evaporative water loss, and therefore of the potential for
water stress, in the locations where they are deployed (Schwarzkopf and Alford
1996; Bartelt and Peterson 2005; Tracy et al. 2007). However, many species
have variable skin resistance to EWL (Toledo and Jared 1993; Lillywhite 2006;
Tracy et al. 2008). Models with zero resistance will significantly underestimate
body temperature for such amphibians at many times, while models with fixed
resistances above zero will overestimate them. No single model can therefore be
used alone to accurately characterize the ranges of body temperatures available
to many amphibians in the field.
21.2.2 A system of models that allows for variable EWL

21.2.2.1 Rationale
Instead of a single body temperature for any combination of environmental
temperature, relative humidity, wind speed, and solar irradiance, the body tem-
peratures available to an amphibian with variable cutaneous resistance, or one
with unknown resistance, can be characterized as falling within an envelope
the upper and lower boundaries of which are formed by the temperatures that
would be experienced by animals with zero and very high skin resistance.
To define the upper and lower boundaries of body temperatures available
to amphibians occupying any microenvironment, it is possible to use pairs
of models that are identical except that one has zero resistance to evaporative
water loss (a permeable model) and the other (an impermeable model) has
either near-complete resistance or resistance equivalent to the highest known
for the species of interest, if good data on this exist. Although we developed and
have implemented this technique using agar models, it could be adapted to use
with models of other types. If the upper boundary of skin resistance for a spe-
cies is known, the technique we suggest could be modified by designing models
similar to those we use, but with resistance that matches the species of interest.
One way to do this has been suggested by Hasegawa et al. (2005); manipulat-
ing the water content of sponge models can vary their rates of EWL. Another
technique would be to apply an impermeable coating in patches distributed
over the body or concentrated in particular regions with lower EWL, and cali-
brate the resulting models against living animals (Navas et al. 2002).
A pair of permeable and impermeable models should always have equilib-
rium temperatures that indicate the lower and upper limits, respectively, of body
temperature available to a frog at thermal equilibrium in any location. Placing
pairs of models in a wide range of activity and retreat sites known to be used by
a species will provide data that set boundaries on the body temperatures that
can be experienced by that species in the habitat being sampled. If models with
zero and near-complete resistance to EWL are used, this is true even if the range
of skin resistance exhibited by the species is unknown, since the models span
the possible extremes. Using pairs of models ensures that the measured range
encompasses the range available to any individual of a species with variable skin
resistance, regardless of its present state.
21.2.2.2 Model design and construction

Models of frogs in the water-conserving posture are constructed of 3% agar.
Most frogs remain in the water-conserving posture during daylight hours, when
temperature relationships are of most interest. Agar loses water as a free surface
(equivalent to a cutaneous resistance of 0; Spotila and Berman 1976). The agar
models are produced in latex molds, which themselves are made from plaster
casts produced in alginate impressions taken from anesthetized or preserved
frogs. To prevent water exchange with the substrate, the ventral surfaces of all
models are coated with an impermeable plastic (PLASTI DIP®, clear, PLASTI
DIP International, Blaine, MN, USA), which is intended for coating tool han-
dles and other surfaces. In this condition, models should simulate the thermal
performance of frogs with zero resistance to EWL in their dorsal skin (Spotila
and Berman 1976), and with their ventral surfaces in contact with a dry, imper-
meable substrate. Leaving the ventral surface uncoated would allow water
exchange with the substrate, but it is unlikely that this would occur in a manner
similar to that experienced by living frogs due to the highly dynamic role of frog
ventral skin in water exchange. Coating the ventral surface simulates frogs in a
known, realistic situation, and eliminates what would otherwise be an uncon-
trolled source of variation in measurements. To create impermeable models, the
remainder of the surface of half of the models is also coated with PLASTI DIP.
These models match the permeable models in every respect, including spectral
absorbence, which spectrophotometric measurements have shown is not meas-
urably affected by clear PLASTI DIP between 330 and 800 nm. Coated and
uncoated models also cannot be distinguished in infrared images taken through
a B W 094 infrared-pass filter, indicating that their reflectance does not dif-
fer substantially in the range from 800 to 1000 nm. Small thermal datalog-
gers (Thermochron iButtons, Dallas Semicondictor, Dallas, TX, USA; diameter
15 mm, height 6 mm) are embedded in each model to allow regular, repeated
measurements of the models’ core temperature.
21.2.2.3 Example of model construction and

validation in the laboratory
Physical models (Figure 21.1) were made from 3% agar, in the shape of Litoria
caerulea in the water-conserving posture. L. caerulea has a variable and moder-
ate cutaneous resistance to EWL (maximum approx 10 s/cm; Buttemer 1990;
Christian and Parry 1997).
The snout–vent length and weight of the models were approximately 75 mm and
44 g. The models were coloured green using four drops per 100 ml of green food
colouring, made by mixing yellow, blue, and rose pink food colouring (Queen Fine
Foods, Alderley, Queensland, Australia) at a ratio of 10:4:1. This produces spectral
absorbence that falls within the range we have measured for L. caerulea between
330 and 800 nm, and qualitatively similar reflectance in the infrared between 800
and approximately 1000 nm as revealed by simultaneous photographs of models
and living frogs taken through a B W 094 infrared-pass filter.
The iButton embedded in each model was programmed to record the
temperature every 30 min. Each datalogger was tied into a knot in a piece of
Agar Datalogger
Twine
Side view Top view
Fig. 21.1 Details of physical models.

twine, which was pegged at each end of the mold as the agar set, so that the
datalogger was suspended in the centre of the model (Figure 21.1). The free
twine remaining at each end of the models could be used to attach them to
surfaces either by using tape over the twine or by tying the twine around fea-
tures such as branches and roots. We set up three opaque, white, plastic con-
tainers (60 cm 40 cm 40 cm), each with a small water bowl in the center
and a metal fly-screen lid. The containers were housed in a constant tempera-
ture room, which maintained ambient temperature between 19.5 and 21.5°C.
Relative humidity fluctuated between 64 and 96% (mean 74%). A 150 W heat
lamp was provided at one end of each container and was illuminated between
0930 and 2130 h to create a temperature gradient simulating the normal range
of temperatures available to this species. We ran four temporal replicates of the
experiment, creating a total of 12 sets of measurements of frog and model tem-
peratures for comparison. Each replicate ran for 3 days.
At the start of each replicate, models were placed in each container in pairs,
with one permeable and one impermeable model. The pairs were located so
they spanned the range of thermal and light environments available in the con-
tainers. Twelve adult L. caerulea (11 males and one female) were captured near
Townsville, Queensland, Australia. They ranged from 74.1 to 91.8 mm snout–
vent length and 26 to 65 g body mass. Prior to experiments, they were main-
tained in smaller containers at ambient temperature in the constant-temperature
room in which the experiments were carried out. Each frog was used in a single
run of the experiment.
We recorded the body temperatures of frogs five times per day (0900, 1100,
1300, 1500, and 1700 h) over the 3 days of each temporal replicate, producing 15
measurements for each individual. The first time (0900 h) was chosen because
at that time the frogs had not had access to any source of extra heat for almost
12 h, and their body temperatures should have been similar to those that would
be measured during nocturnal readings in the field. Each temporal replicate was
set up at least 60 min before the first temperature reading was taken, allowing
models and frogs to reach a thermal steady state. At each reading, cloacal tem-
peratures were measured using a small, chromel-alumel K-type thermocouple
(diameter ≈1 mm) with the tip coated in plastic, attached to a digital therm-
ometer (Small Pocket Thermometer model 90000, Industrial Automation,
Joondalup, Western Australia). During temperature measurement, each frog
was held by a single leg only. When temperatures were taken, the posture and
location of the individual was noted; this information was later used to deter-
mine whether individuals had changed location, and to select the pair of models
that should represent the upper and lower boundaries of the thermal envelope
available to each frog at each measurement. All dataloggers and the thermo-
couple were calibrated before the experiment.
All frog body temperatures fell within the broad envelope defined by the
maximum and minimum temperatures of all impermeable and permeable models,
respectively (Figure 21.2a). Temperatures of frogs fell within the narrower enve-
lope defined by the temperatures measured in the same time interval in the
permeable and impermeable models nearest to the frog on 144 of the 180 (80%)
(a) 45
40
Temperature (°C)
35
30
25
20
15
0 400 800 1200 1600 2000 2400

Time of day (h)
(b)
40
Range of nearest models (°C)
35
30
25
20
15
20 22 24 26 28 30 32
Cloacal temperature (°C)
Fig. 21.2 (a) Range of temperatures available to Litoria caerulea in laboratory

thermal gradients, as indicated by maximum and minimum temperatures of
impermeable and permeable agar models (solid lines) and cloacal temperatures
measured for individual frogs (points). (b) Lines indicating the range of temperatures
between the permeable and impermeable models closest to each individual L.
caerulea at each time its cloacal temperature was taken in laboratory thermal
gradients. The line of equality is included to ease visualization of where cloacal
temperature fell within each pair of model temperatures.
occasions on which we measured frog temperatures (Figure 21.2b). As expected,

permeable models were cooler than frog body temperatures, while imperme-
able models were warmer than frog body temperatures. For 111 of our observa-
tions, frogs had changed their location since the last observation, or were being
observed for the first time. On 28 of the 36 (77.8%) occasions when frog cloacal
temperatures were outside the envelope defined by the nearest models, frogs
had changed their location since the time of the last observation or were being
observed for the first time, and therefore might not have equilibrated thermally.
The association between changing location and being outside the thermal enve-
lope was statistically significant (Fisher’s Exact Test, two-tailed, P 0.034). No
cloacal temperature was more than 2.6°C below or 1.2°C above the boundaries
of the envelope as defined by the nearest pair of models, and all were well within
the extremes defined by all models.
21.2.2.4 Field validation of use of models to

characterize available temperatures
We tracked 27 Litoria lesueuri using radiotelemetry and harmonic radar in the
warm/wet season (19–21 March 2005) and the cool/dry season (5–13 August
2005) in Frenchman Creek, near Babinda, Queensland, Australia (17°20S,
145°55E; 20–100 m above sea level; Rowley and Alford 2007a). Weather
conditions during both tracking periods varied considerably with respect to
ambient temperature, humidity, solar radiation, wind speed, and rainfall. Body
temperatures were recorded once during the day (between 1200 and 1600 h)
and once at night (between 1900 and 0700 h) by holding a Raytek ST80 Pro-
Plus Non-contact thermometer approximately 5 cm away from the frog and
aiming at the lower dorsal area, between the thigh and point of transmitter
attachment. We have demonstrated that the temperatures we measure in this
manner accurately reflect the skin and cloacal temperatures of frogs (Rowley
and Alford 2007b). During the tracking period, we placed pairs of permeable
and impermeable models, constructed as previously described, in sites previ-
ously used as diurnal retreat sites by the tracked L. lesueuri (on bare ground, leaf
litter, in vegetation, and on gravel, in exposed and sheltered positions; Rowley
and Alford 2007c; Figure 21.3). Fifteen pairs of models were deployed during
the warm/wet season, and eight pairs during the cool/dry season. The embed-
ded iButtons recorded temperatures of the models every 30 min.
Models were left in the field and weighed every 24 h until at least one per-
meable model had lost 50% of its original weight. Although these models were
reduced in size, they were within the range of body sizes of L. lesueuri, retained
their basic shape, and were still 94% water, and thus continued to lose water at a
Fig. 21.3 Two of pairs of models placed in retreat sites used by frogs in the field
very similar rate per unit surface area. We would not recommend allowing models
to fall below 50% of their original mass, as they begin to lose shape, and their
proportional water content, and thus equivalent resistance to EWL, increases at
an accelerating rate. Model size affects rates of heating and cooling, and could
possibly alter equilibrium temperatures by altering the ratio of surface area to
volume. However, our field data (Figure 21.4) suggest that any effects of changes
as permeable models dehydrated on equilibrium temperature were very small.
The distributions of temperature difference between the permeable and imper-
meable models were similar for all levels of desiccation, and the distribution in
the 50–60% desiccated class was more similar to that in the 90–100% class
than it was to those in the 60–70% and 70–80% classes, probably because of
weather conditions at the time of measurement rather than any effect of drying
on model temperature. The outlying extreme differences shown in Figure 21.4
25 N = 77 139 334 728 2502 2144
Impermeable – permeable (°C) 20
15
10
50–60 60–70 70–80 80–90 90–100 100–110

Permeable model: percent of original weight
Fig. 21.4 Differences recorded between permeable and impermeable members

of pairs of agar models of Litoria lesueuri placed in frog diurnal retreat sites in
rainforest for up to 8 days near Babinda, Queensland, Australia. Box plots illustrate
the median (heavy line within boxes), upper and lower quartile boundaries (box
boundaries), range (lines), and outlying (circles) and extreme (asterisks) data points
for differences, grouped by the weight of the permeable model as a percentage of
its original weight. A small number of points at which the temperatures of permeable
models exceeded those of impermeable models because only the permeable model
was in sunlight have been omitted.
usually occurred at the same time each day and reflected occasions on which
only the impermeable model was in the sun.
To determine the accuracy with which the models delineated the envelope
of body temperatures available to frogs, we compared each measured frog body
temperature with the minimum and maximum temperatures attained by models
at the nearest 30 min mark (
15 min time difference). We removed 16 frog
body temperatures taken when frogs were found in microhabitats in which
we had not placed models (i.e. under logs, on roads, in mown pastures, or in
exposed sites high in terrestrial vegetation). Of the 145 field body temperatures
remaining (30 taken in the warm/wet season and 115 in the cool/dry season),
121 (83.4%) were within the limits defined by the hottest impermeable model
and the coldest permeable model, and 19 of the remaining 24 were within
±0.5°C of those limits. Only one body temperature was more than 0.9°C out-
side them (1.5°C above; Figure 21.5). Two-thirds of the temperatures that fell
outside the limits were taken during nocturnal activity; in most cases these
frogs were warmer than the models (Figure 21.5), which were all in diurnal
retreat sites. This suggests that the nocturnal activity sites of frogs are warmer
30
Model temperature range (°C)

27
24
21
Night
Day
18
15 Cool/dry season Warm/wet season
15 18 21 24 27 30
Frog body temperature (°C)
Fig. 21.5 Lines indicating the range of temperatures delineated by the maximum
and minimum temperatures of all permeable and impermeable models at rainforest
field site near Babinda, Queensland, Australia at the times body temperatures were
measured for tracked Litoria lesueuri during the cool/dry and warm/wet seasons.
Models were placed at known diurnal retreat sites of L. lesueuri. The line of equality
is included to ease visualization of where cloacal temperature fell within each pair of
model temperatures.
at night than their diurnal retreat sites are, and that models should be placed
in nocturnal activity sites to more accurately delineate the environment expe-
rienced by frogs.
The physical models produced a relatively accurate outline of the temperature
range into which the body temperatures of thermally equilibrated frogs fell, and
the ways in which the model temperature envelope differed from the frog data
were informative in themselves. This should hold true when this technique is
used for any species of amphibian.
Our field validation is more rigorous than most that have been undertaken
with reptiles (Dzialowski 2005). Many reptile studies evaluate the range of
available temperatures by deliberately placing models to cover the maximum
possible range of environmental temperatures (Dzialowski 2005). Had we done
this, all of our field temperatures would have fallen within the range measured
by models, as our laboratory temperatures did. This is not the main objective
of our technique, which is intended to use models as a means of gaining more
detailed information than is otherwise possible on the range of body tempera-
tures a species is likely to experience, given its known behavior and microhabitat
use. Measured body temperatures that fall outside the range experienced by
models are not a failure of the models. Rather, they indicate that the species
being studied uses a broader range of microenvironments than was provided by
the locations in which models were placed.
21.2.2.5 The importance of matching the spectral

properties of models to living animals
There has been substantial debate in the literature on the degree to which the
spectral properties of models should match those of the species being modeled
(reviewed by Dzialowski 2005). The color and therefore absorbence of many
amphibians can alter when basking, or with thermal stress (Shoemaker et al.
1987; Snyder and Hammerson 1993; Withers 1995). For example, the total
reflectance of Litoria rubella can range between approximately 19 and 32%
(Withers 1995). This variation among and within individuals means that
even a particular individual of many species cannot be precisely mimicked
by a static model. We performed a short-term experiment to examine how
color affected the thermal behavior of permeable and impermeable models
of L. caerulea. We used green agar models with reflectance spectra that fell
within the range of living individuals, constructed as described in section
21.2.2.3, and also constructed agar models with no added coloring and with
3 g/100 ml of carbon black added. Three replicate models of each type were
placed, interspersed by color, in an open area during late summer, and their
temperatures were measured over 24 h. Weather was largely clear with scat-
tered clouds and no rain during the trial. We found that the relatively large
difference in absorbance between green and clear agar models had only small
effects on the daytime temperatures of impermeable models (Figure 21.6);
green impermeable models were only slightly, and never significantly, warmer
than colorless models. Even extreme differences in color had little effect on
midday temperatures of permeable models (Figure 21.6). Although it is clearly
best to produce models that approximate the overall absorbance of the species
being examined, our results suggest that thermal envelopes estimated using
models are likely to be reasonable approximations of the range of body tem-
peratures available to amphibians, as long as obviously extreme colors such as
carbon black are avoided.
21.2.2.6 Models as indicators of relative moisture

availability of microenvironments
Although physical models have often been used to study rates of evaporative
water loss under laboratory conditions, very few field studies have used them to
55 Black impermeable
Temperature (°C) ± 95% CI
Green impermeable
45
Colorless impermeable
35
25
All permeable
15
10 12 14 16 18 20 22 0 2 4 6 8 10
Time
Fig. 21.6 Temperatures at hourly intervals for permeable and impermeable agar
models of three colors exposed outdoors to ambient summer conditions. Vertical
bars indicate 95% confidence intervals (CI) for the three replicate models of each
color and permeability, lines connect the means for each color and type to aid in
visualization of patterns. Because all permeable models performed almost identically
while well hydrated, we have not attempted to make it possible to distinguish their
lines, although the black models were slightly hotter. 95% Confidence interval bars
were omitted for the last hour because steep slopes of curves made them unreadable.
All were of magnitudes similar to those at hours 10–12 on the previous day.
assess the relative degree of water stress in a range of habitats or microhabitats,

or across seasons. O’Connor and Tracy (1987) compared the water loss of wet-
ted plaster models in different microhabitats to estimate water loss that would
be experienced by Rana pipiens. Schwarzkopf and Alford (1996) placed agar
models in retreat sites used by Bufo marinus to characterize the relative degree
of water stress experienced by this introduced species in the Australian tropics.
Their data strongly supported the hypothesis that seasonal shifts in retreat site
use by B. marinus are related to water conservation. Seebacher and Alford (2002)
obtained similar results using formalin-fixed Bufo marinus as models. Navas
et al. (2002) compared water loss in frogs and in agar models under simulated
field conditions, and Tracy et al. (2007) performed short-term comparisons of
temperature and water loss between plaster models of B. marinus and living
toads. It appears that no other published studies have used models extensively to
examine the hydric conditions available to amphibians in the field.
Wet season
1.1
1.0
Permeable model weight as 0.9
proportion of initial weight 0.8
0.7
0.6
0.5
Dry season
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0 2 4 6 8
Days since model placement
Fig. 21.7 Relative water loss by permeable agar models placed in diurnal retreat
sites of Litoria lesueuri in rainforest near Babinda, Queensland, Australia. There was
substantial cloud cover and rainfall during the winter, nominally dry season, sampling
period (top), while the summer, nominally wet season, sampling period (bottom)
was unusually dry. In both seasons, some retreat sites provided very low rates of
evaporative water loss while some led to higher rates; in the summer, one of the
models reached 50% water loss on day 4, causing data collection to cease.
Data collected using models to examine the envelope of thermal conditions

available to amphibians can also be used to delineate an envelope of relative degrees
of water stress. When agar models are used, the simplest approach is to weigh
models daily to document cumulative change in water content. If automatically
rehydrated models of other types, such as those developed by Bartelt and Peterson
(2005) or Tracy et al. (2007) are used, cumulative water lost from reservoirs can be
recorded manually or monitored automatically. Figure 21.7 presents data collected
from the models of L. lesueuri placed in the field near Babinda, Queensland, as
described in section 21.2.2.4. These data demonstrate that although the rainforest
where the data were collected experiences annual wet and dry seasons, short-term
weather patterns can result in greater opportunity for water stress in the wet sea-
son than in the dry season; this is probably compounded by the fact that the wet
season occurs in the summer months, in which diurnal temperatures can be up
to 10°C higher than in the winter dry season. They also indicate that the diurnal
retreat sites that were monitored presented a wide range of potential levels of water
stress. This is likely to affect the frogs’ behavior and metabolism, and is also likely
to affect their interactions with the amphibian chytrid fungus, which may repro-
duce more slowly on hosts with relatively dry skin surfaces (Pounds et al. 2006).
21.3 Summary and future developments

Models are useful for delineating the range of body temperatures and degrees of
moisture stress available to amphibians in the field. Because numerous models
can be placed in known retreat and activity sites, and temperatures and water
loss can be recorded in models at frequent intervals, they can rapidly provide
a detailed picture of the envelope of temperatures and moisture available to a
species, without requiring that individuals be handled or otherwise disturbed.
If a species is known to have zero resistance to EWL, permeable models should
be sufficient to characterize its thermal and hydric environment. However, for
species that have higher and usually variable levels of resistance to EWL, pairs
of models should be employed to delineate the boundaries of the thermal enve-
lope available.
Pairs of models cannot predict the exact body temperatures of specific indi-
vidual frogs. Rates of EWL can fluctuate considerably over even short periods
in individual frogs (McClanahan 1978; Withers et al. 1982; Wygoda 1989a;
Withers 1995; Barbeau and Lillywhite 2005; Tracy et al. 2008). Although many
frogs alter EWL with respect to temperature or humidity in a moderately pre-
dictable manner (Withers et al. 1982; Wygoda 1989a; Buttemer and Thomas
2003), in others, EWL fluctuates unpredictably (Barbeau and Lillywhite 2005).
Because rates of EWL are so variable, measuring the limits of the envelope of
body temperatures available to a species in a given microenvironment is likely to
be more ecologically relevant to population-level questions for most species than
precise measurements for particular individuals can be.
The physical models we have described are easy to construct in large numbers,
and are also relatively inexpensive to produce, features that are highly desirable,
since to truly characterize the range of conditions available to animals in the
field it is necessary to sample a large number of locations (Dzialowski 2005).
Our models are also compact and self-contained, and can therefore be placed
in the exact shelter sites used by the study species without altering their struc-
ture, which might prove difficult with models of other designs (e.g. Bartelt and
Peterson 2005; Tracy et al. 2007).
Using data obtained from models in the field in conjunction with the detailed
knowledge of temporal patterns of microhabitat use made possible by advances
in tracking technology, it will be possible to develop pictures of temporal pat-
terns of body temperature experienced by frogs that are as detailed as those
that have been available for some time for many reptiles. Variation in relative
rates of water loss among sites and through time can also be determined. The
knowledge that can be gained using models is likely to be increasingly useful
in understanding the effects of climate change on amphibians. The majority

of predictive models forecast a near-term global temperature rise of 1.5–4.5°C
(Thomas et al. 2004), and these estimates have been used in predicting species
responses (Thomas et al. 2004) despite their questionable relevance to micro-
environmental conditions or body temperature (Kennedy 1997). Detailed
species-specific data, as can be obtained using models, will therefore be essen-
tial for accurately predicting responses and designing management plans. This
approach is also likely to be of use in developing more sophisticated models of
energy use and growth within individuals, and thus of population dynamics,
and in understanding the factors affecting host-pathogen interactions in nature
(e.g. Woodhams et al. 2003).
21.4 References
Bakken, G. S. and Gates, D. M. (1975). Heat-transfer analysis of animals: some implica-
tions for field ecology, physiology, and evolution. In D. M. Gates and R. B. Schmerl
(eds), Perspectives in Biophysical Ecology, pp. 255–90. Springer, New York.
Barbeau, T. R. and Lillywhite, H. B. (2005). Body wiping behaviors associated with
cutaneous lipids in hylid tree frogs of Florida. Journal of Experimental Biology, 208,
2147–56.
Bartelt, P. E. and Peterson, C. R. (2005). Physically modeling operative temperatures and
evaporation rates in amphibians. Journal of Thermal Biology, 30, 93–102.
Bradford, D. F. (1984). Temperature modulation in a high-elevation amphibian, Rana
muscosa. Copeia, 1984, 966–76.
Buttemer, W. A. (1990). Effect of temperature on evaporative water loss of the Australian
tree frogs Litoria caerulea and Litoria chloris. Physiological Zoology, 63, 1043–57.
Buttemer, W. A. and Thomas, C. (2003). Influence of temperature on evaporative water
loss and cutaneous resistance to water vapour diffusion in the orange-thighed frog
(Litoria xanthomera). Australian Journal of Zoology, 51, 111–18.
Buttemer, W. A., van der Wielen, M., Dain, S., and Christy, M. (1996). Cutaneous
properties of the Green and Golden Bell Frog Litoria aurea. Australian Zoologist, 30,
134–8.
Christian, K. A. and Parry, D. (1997). Reduced rates of water loss and chemical properties
of skin secretions of the frogs Litoria caerulea and Cyclorana australis. Australian Journal
of Zoology, 45, 13–20.
Dzialowski, E. M. (2005). Use of operative temperature and standard operative tempera-
ture models in thermal biology. Journal of Thermal Biology, 30, 317–34.
Elliot, S. L., Blanford, S., and Thomas, M. B. (2002). Host-pathogen interactions in
a varying environment: temperature, behavioural fever and fitness. Proceedings of the
Royal Society of London Series B Biological Sciences, 269, 1599–1607.
Hasegawa, M., Suzuki, Y., and Wada, S. (2005). Design and performance of a wet sponge
model for amphibian thermal biology. Current Herpetology, 24, 27–32.
Kennedy, A.D. (1997). Bridging the gap between general circulation model (GCM) out-
put and biological microenvironments. International Journal of Biometeorology, 40,
119–22.
Kobelt, F. and Linsenmair, K. E. (1992). Adaptations of the reed frog Hyperolius viridifla-
vus (Amphibia: Anura: Hyperoliidae) to its arid environment. Journal of Comparative
Physiology B, 162, 314–26.
Lillywhite, H. B. (2006). Water relations of tetrapod integument. Journal of Experimental
Biology, 209, 202–26.
Lips, K. R., Brem, F., Brenes, R., Reeve, J. D., Alford, R. A., Voyles, J., Carey, C., Livo, L.,
Pessier, A. P., and Collins, J. P. (2006). Emerging infectious disease and the loss of bio-
diversity in a Neotropical amphibian community. Proceedings of the National Academy
of Sciences USA, 103, 3165–70.
McClanahan, L. L. (1978). Skin lipids, water loss, and energy metabolism in a South
American tree frog (Phyllomedusa sauvagei). Physiological Zoology, 51, 179–87.
McClanahan, L. L. and Shoemaker, V. H. (1987). Behavior and thermal relations of the
arboreal frog Phyllomedusa sauvagei. National Geographic Research, 3, 11–21.
Navas, C. A. (1996). Implications of microhabitat selection and patterns of activity on the
thermal ecology of high elevation neotropical anurans. Oecologia, 108, 617–26.
Navas, C. A. and Araujo, C. (2000). The use of agar models to study amphibian thermal
ecology. Journal of Herpetology, 34, 330–4.
Navas, C. A., Jared, C., and Antoniazzi, M. M. (2002). Water economy in the casque-
headed tree-frog Corythomantis greeningi (Hylidae): role of behaviour, skin, and skull
skin co-ossification. Journal of Zoology, 257, 525–32.
O’Connor, M. P. and Tracy, C. R. (1987). Thermal and hydric relations of leopard frogs
in the field. American Zoologist, 27, 118A.
Parmesan, C., Gaines, S., Gonzalez, L., Kaufman, D. M., Kingsolver, J., Peterson, A. T.,
and Sagarin, R. (2005). Empirical perspectives on species borders: from traditional
biogeography to global change. Oikos, 108, 58–75.
Pounds, J.A. , Bustamante, M. R., Coloma, L. A., Consuegra, J. A., Fogden, M. P. L.,
Foster, P. N., La Marca, E., Masters, K. L., Merino-Viteri, A., Puschendorf, R. et al.
(2006). Widespread amphibian extinctions from epidemic disease driven by global
warming. Nature, 439, 161–7.
Rowley, J. J. L. and Alford, R. A. (2007a). Movement patterns and habitat use of rainfor-
est stream frogs in northern Queensland, Australia: implications for extinction vulner-
ability. Wildlife Research, 34, 371–8.
Rowley, J. J. L. and Alford, R. A. (2007b). Non-contact infrared thermometers can accur-
ately measure amphibian body temperatures. Herpetological Review, 38, 308–11.
Rowley, J. J. L. and Alford, R. A. (2007c). Behaviour of Australian rainforest stream
frogs may affect the transmission of chytridiomycosis. Diseases of Aquatic Organisms,
77, 1–9.
Schmuck, R., Kobelt, F., and Linsenmair, K. E. (1988). Adaptations of the reed frog
Hyperolius viridiflavus (Amphibia, Anura, Hyperoliidae) to its arid environment:
V. Iridophores and nitrogen metabolism. Journal of Comparative Physiology B, 158,
537–46.
Schwarzkopf, L. and Alford, R. A. (1996). Desiccation and shelter-site use in a trop-

ical amphibian: comparing toads with physical models. Functional Ecology, 10,
193–200.
Seebacher, F. and Alford, R. A. (2002). Shelter microhabitats determine body temperature
and dehydration rates of a terrestrial amphibian (Bufo marinus). Journal of Herpetology,
36, 69–75.
Shoemaker, V. H. and McClanahan, L. L. (1975). Evaporative water loss, nitrogen excre-
tion and osmoregulation in phyllomedusine frogs. Journal of Comparative Physiology,
100, 331–45.
Shoemaker, V. H. and Nagy, K. A. (1977). Osmoregulation in amphibians and reptiles.
Annual Review of Physiology, 39, 449–71.
Shoemaker, V. H., McClanahan, L. L., Withers, P. C., Hillman, S. S., and Drewes, R. C.
(1987). Thermoregulatory response to heat in the waterproof frogs Phyllomedusa and
Chiromantis. Physiological Zoology, 60, 365–72.
Snyder, G. K. and Hammerson, G. A. (1993). Interrelationships between water econ-
omy and thermoregulation in the canyon tree frog Hyla arenicolor. Journal of Thermal
Biology, 25, 321–9.
Spotila, J. R. and Berman, E. N. (1976). Determination of skin resistance and the role of
the skin in controlling water loss in amphibians and reptiles. Comparative Biochemistry
and Physiology, 55A, 407–11.
Thomas, C. D., Cameron, A., Green, R. E., Bakkenes, M., Beaumont, L. J.,
Collingham, Y. C., Erasmus, B. F. N., de Siquera, M. F., Grainger, A., Hannah, L. et al.
(2004). Extinction risk from climate change. Nature, 427, 145–8.
Toledo, R. C. and Jared, C. (1993). Cutaneous adaptations to water balance in amphib-
ians. Comparative Biochemistry and Physiology, 105A, 593–608.
Tracy, C. R., Christian, K. A., O’Connor, M. P., and Tracy, C. R. (1993). Behavioral
thermoregulation by Bufo americanus: the importance of the hydric environment.
Tracy, C. R., Betts, G., Tracy, C. R., and Christian, K. A. (2007). Plaster models to
measure operative temperatures and evaporative water loss of amphibians. Journal of
Tracy, C. R., Christian, K. A., Betts, G., and Tracy, C. R. (2008). Body temperature
and resistance to evaporative water loss in tropical Australian frogs. Comparative
Biochemistry and Physiology, 150A, 102–8.
Withers, P. C. (1995). Evaporative water loss and colour change in the Australian desert
tree frog Litoria rubella (Amphibia: Hylidae). Records of the Western Australian Museum,
17, 277–81.
Withers, P. C., Hillman, S. S., Drewes, R. C., and Sokol, O. M. (1982). Water loss and
nitrogen excretion in Sharp-nosed Reed frogs (Hyperolius nasutus: Anura, Hyperoliidae).
Journal of Experimental Biology, 97, 335–43.
Woodhams, D. C., Alford, R. A., and Marantelli, G. (2003). Emerging disease of amphib-
ians cured by elevated body temperature. Diseases of Aquatic Organisms, 55, 65–7.
Wygoda, M. L. (1984). Low cutaneous evaporative water loss in arboreal frogs. Physiological
Zoology, 57, 329–37.
Wygoda, M. L. (1989a). Body temperature in arboreal and nonarboreal anuran amphib-

ians: differential effects of a small change in water vapour density. Journal of Thermal
Biology, 14, 239–42.
Wygoda, M. L. (1989b). A comparative study of heating rates in arboreal and nonarboreal
frogs. Journal of Herpetology, 23, 141–5.
Young, J. E., Christian, K. A., Donnellan, S., and Tracy, C. R. (2005). Comparative
analysis of cutaneous evaporative water loss in frogs demonstrates correlation with eco-
logical habits. Physiological and Biochemical Zoology, 78, 847–56.
22
Genetics in field ecology and conservation
Trevor J. C. Beebee
22.1 Background: the importance of

genetics in ecology and conservation
The genetic analysis of wild populations has become increasingly popular with
the advent of techniques based on large-scale and non-destructive sampling.
Nevertheless, using molecular methods remains one of the most expensive and
time-consuming aspects of ecological research. It is therefore important to con-
sider carefully when it is useful to acquire genetic data. Certainly there are such
circumstances for amphibian researchers (Beebee 2005). Areas in which genet-
ics can be invaluable include the following.
1) Identification of cryptic species or life stages. There are many examples
among the amphibia of difficulties with identification based on morph-
ology alone. This can be true for adults, but is a more common problem
with larvae. Molecular genetic markers permit species identification in
virtually all situations.
2) Measuring genetic diversity, and detecting inbreeding and genetic load
effects. Small and isolated populations are vulnerable to the accumula-
tion of mildly deleterious alleles. These decrease the fitness of individuals,
and eventually the viability of the population, contributing to a poten-
tial “extinction vortex” in combination with other factors such as demo-
graphic and environmental stochasticity. Genetic methods can provide
useful indicators of this problem, potentially leading to genetic restor-
ation. They can also estimate the “effective size” of a population (roughly
the number of successful breeders, and usually smaller than the census
size), which is what matters in risk assessment.
3) Identification of barriers to individual movement. Amphibians are, in gen-
eral, poor dispersers but long-term population viability may often depend
on occasional movements of animals between distant breeding sites. Such

connectivity maintains a large effective population size over wide areas,
but can be difficult to demonstrate when individual movements are rare.
Genetic techniques can track and quantify such movements indirectly.
4) Defining populations. Knowing the extent and size of individual popula-
tions is important for demographic studies, and for conservation manage-
ment. Assemblages in ponds have commonly been used as surrogates for
“populations” of pond-breeding amphibians but this is often too simplistic.
Many species exist as metapopulations with individuals moving regularly
between neighbouring ponds. Genetic analysis can refine population size
and boundary information.
5) Historical issues. Genetic data can be used to reveal the sources of alien
introductions, and to show whether a population of uncertain origin is
recently introduced, perhaps warranting eradication, or is truly native and
thus meriting conservation.
6) Behavior and sexual selection. High-resolution molecular markers now
permit the determination of individual DNA profiles. This means that
issues such as parentage, and thus individual reproductive success, can be
determined and correlated with features such as body size and condition.
In the following sections, I describe how genetic data can be obtained from wild
amphibian populations, and how they can be analysed to address the questions
outlined above. More detailed accounts can be found in Plötner et al. (2007)
and Vences and Wake (2007). The overwhelming majority of genetic studies
on wildlife populations thus far have used “neutral” markers; that is, changes
in DNA sequences that are invisible to natural selection. These types of marker
perform well for addressing most of the issues listed above, although there is a
serious question about their application to inbreeding depression and genetic
load. In the final section I consider likely future developments, including the
identification of genetic markers under the influence of selection for investigat-
ing population viability.
22.2 Molecular methods for investigating

amphibian populations
22.2.1 Laboratory facilities
There is a substantial amount of core equipment necessary for any genetic study.
This includes: weighing balances; pipettes; plasticware including disposable
pipette tips, and microcentrifuge and polymerase chain reaction (PCR) tubes;
22 Genetics in field ecology and conservation | 409
fridges and freezers; water baths; microcentrifuges; vortex mixers; a vacuum

desiccator; PCR machines; gel electrophoresis rigs; and either an automated
DNA sequencer or high-voltage electrophoresis equipment to run sequencing
gels, together with X-ray development or phosphorimager facilities. Also very
useful are ultraviolet (UV) spectrophotometers for quantifying DNA yields,
UV photographic equipment for identifying DNA bands on gels, microwave
ovens, and autoclaves (although a pressure cooker will often suffice!). For more
sophisticated work, such as the generation of DNA libraries for isolating micro-
satellite loci, microbiology facilities are also necessary, including sterile hoods
and incubators. These facilities are often not readily available to ecologists, so
there is as strong case for collaborations with suitably equipped laboratories
when contemplating genetic studies. All laboratory work should be carried out
using high standards of cleanliness, and especially with the use of disposable
gloves (non-latex when handling live animals) for all procedures.
22.2.2 Sampling
All the genetic methods described below are based on PCR amplification of very
small amounts of DNA. Thus tiny amounts of tissue suffice for analysis, and
sacrifice of individuals is not necessary. There are several ways of obtaining tis-
sue samples for DNA extraction. Small (2–3 mm2) segments of larval tailfins or
tail tips can be excised with a sharp scalpel; terminal digits can be removed with
sharp scissors or buccal swabs can be taken from immature or adult amphibians
(Pidancier et al. 2003). None of these procedures are likely to cause significant
mortality. It is important to ensure that sampling is representative of a popula-
tion. Taking all the samples from one aggregation of tadpoles, for example, may
bias in favour of a few sibships.
Each tissue sample should be preserved immediately in a small volume (≈1 ml)
of at least 70% ethanol, in sealed tubes, and returned to the laboratory. DNA is
stable in alcohol for many months at field or room temperature. Buccal swabs
(e.g. from Epicentre Biotechnologies, Madison, WI, USA) are an exception, and
should be treated as per the manufacturer’s instructions. With future statistical
analysis in mind, it is desirable to sample at least 20 individuals per population.
22.2.3 DNA extraction

Several methods are available for rapid extraction of DNA from multiple small
tissue samples. This initial step is important, because the PCR requires high-
quality DNA with minimal contamination. Buccal swabs come with extraction
solutions and method details. Four other techniques are outlined below, fol-
lowed by a summary of their relative merits.
22.2.3.1 Proteinase K/phenol/chloroform method

Tissue samples are digested overnight at high temperature ( 50°C) in the
presence of the enzyme proteinase K, conditions that leave DNA intact. The
mixture is then extracted with organic solvents (phenol and chloroform) that
precipitate residual proteins and other cell debris, leaving DNA in solution above
the organic layer after a brief centrifugation. The DNA is then concentrated by
precipitation using ethanol, and obtained as a pellet following centrifugation.
After drying, the DNA is redissolved in less than 50 μl of sterile distilled water
and stored (always at 20°C).
22.2.3.2 Kit-based DNA extraction

Many companies (e.g. Qiagen, Crawley, West Sussex, UK) produce kits designed
for efficient DNA extraction from multiple samples. These kits include all the
buffer solutions required, and mini-columns with a silica-based plug that select-
ively binds DNA. Detailed protocols are provided, and DNA is recovered from
the columns in a final elution step.
22.2.3.3 Chelex method

Chelex-100 is an ion-exchange resin marketed by several chemical companies.
It forms a suspension in distilled water, and when tissue is added and incubated
overnight (at 50°C) DNA is released into solution. A final boiling step is
necessary to complete the process, followed by a brief centrifugation. The super-
natant above the resin pellet is stored as a source of DNA (Walsh et al. 1991).
22.2.3.4 FTA cards

A recent development is the use of FTA cards (Whatman, Maidstone, Kent,
UK), a substrate onto which tissue extracts (including buccal swabs) can be
blotted directly and which can then store DNA without degradation for months
or years. Small sections of the card are removed using a punch, and after brief
washes can be used directly in PCR amplifications. Several hundred amplifica-
tions can be made from half an FTA card, and subsequent genotyping is very
reliable (Livia et al. 2006).
22.2.3.5 Summary of DNA-extraction methods

The first two DNA extraction methods discussed above generate high-quality
native (double-stranded) DNA suitable for virtually all types of genetic ana-
lysis. However, care must be taken with the use of the organic solvents, which
are potentially toxic. Kits are more expensive than the phenol/chloroform
method, and tend to produce lower yields. However, extraction time is shorter
with a kit and there is less risk of contamination with substances that can
interfere with subsequent PCR amplifications. Some amphibians have high
concentrations of tissue pigments, and these are not always removed by the
phenol/chloroform approach. The Chelex method is attractively cheap and
simple, and contamination problems are rare. However, Chelex extraction pro-
duces denatured (single-stranded), not native, DNA. This makes it unsuitable
for techniques such as amplified fragment length polymorphism (AFLP; see
below) that absolutely require native DNA. FTA cards may also be unsuitable
for AFLP analysis, where DNA needs to be digested with restriction enzymes
prior to PCR amplification.
22.2.4 Quantifying DNA recoveries

It is important to estimate DNA recoveries from the tissue samples to ensure
that yields are both sufficient and consistent. This is best done using a UV spec-
trophotometer because DNA absorbs light strongly at 260 nm. Native DNA has
an absorbance of 1.0 at 50 μg/ml, whereas denatured DNA and oligonucleotides
at this concentration have an absorbance of about 1.25.
22.2.5 The basis of PCR analysis

The PCR is fundamental to the molecular genetic methods relevant to ecol-
ogy and conservation. It is based on the exponential increase of a particu-
lar DNA sequence that can be obtained in just a few hours using a series of
denaturation, priming, and polymerization cycles in a programmable labora-
tory machine. In principle, a single DNA molecule can give rise to 1 billion
copies (more than enough for any analysis) in just 30 cycles of amplification.
In each cycle, a denaturation step separates the two DNA strands, at over
90°C; a priming step allows specially designed oligonucleotide primers, typic-
ally 18–25 nucleotides long and complementary in base sequence to flanking
sections of the piece of DNA, to anneal to the separated strands, usually at
50–60°C; and a polymerization step uses a heat-stable (Taq) DNA polymer-
ase and nucleotide substrates to synthesise new strands of DNA between the
priming sites, starting at the inward-facing ends of each primer, at 70–72°C.
The ability to amplify such tiny amounts of starting material permits non-
destructive sampling, but also renders the process vulnerable to traces of DNA
contamination from unexpected sources (such as shed human skin cells), a
problem that must always be tested for by using blanks with no added DNA.
PCR machines can process up to 100 samples simultaneously, and are thus
well suited to population studies.
22.2.6 Choice of analytical methods

Given a good batch of DNA samples, what are the options for genetic analysis?
A lot depends on the problem under consideration but there are nevertheless
only a small number of common marker types from which to choose (Beebee
and Rowe 2008).
22.2.6.1 Mitochondrial DNA (mtDNA)

Mitochondria contain multiple copies of a single, circular DNA chromosome.
The structure of this chromosome is highly conserved in vertebrates, and all
have the same set of 13 protein-encoding genes, together with genes coding
for ribosomal and transfer RNAs, and a control region (or CR) or D-loop. The
control region is concerned with the regulation of RNA and DNA synthesis.
mtDNA has a mutation rate of around tenfold higher than nuclear protein-coding
genes, with the result that genetic diversity accrues relatively rapidly. It does
not undergo meiosis, so sequence changes can be tracked over time as distinct
lineages unscrambled by recombination. Inheritance is purely maternal because
although mtDNA occurs in sperm, it does not survive fertilization. Finally,
because most cells have many mitochondria, each with many copies of mtDNA,
there are typically hundreds or thousands of mtDNA molecules per diploid cell
compared with just two copies of each nuclear chromosome. This means that
mtDNA is relatively easy to detect in old or partly degraded tissue samples such
as museum specimens, despite constituting less than 1% of a cell’s total DNA.
In a typical mtDNA analysis, oligonucleotide primers complementary to parts
of a mtDNA gene are used in PCRs to generate partial gene fragments. These
are then purified (commercial kits are available to clean up PCR products) and
sequenced. Sequencing can be done in-house or sent to one of the many com-
panies that provide this service at low cost. The sequences are then compared
for polymorphisms among many different individuals, each of which will carry
only a single mtDNA sequence, referred to as its mtDNA haplotype.
mtDNA is particularly useful for phylogeographic studies, tracing the ori-
gin of populations over periods of up to a few million years. mtDNA has, for
example, provided a mass of evidence concerning postglacial colonization routes
of amphibians in the northern hemisphere over the past 20 000 years. However,
it is usually less powerful than the nuclear DNA markers now in widespread
use for fine-scale studies of population structure. Although high by compari-
son with nuclear protein-coding genes, the mtDNA mutation rate is often still
too low to generate enough diversity for inter-population analysis. Exceptions
are places where species have persisted for a long time, such as glacial refugia in
southern Europe, or in the tropics. A particular advantage in the use of mtDNA
is the existence of conserved (or universal) primers for PCR reactions that func-
tion with many different species, particularly for the cytochrome b (cytb) gene.
This means that study of a new species can be initiated with minimal lead-in
time for the development of genetic markers. A disadvantage is that due to its
lack of recombination mtDNA can only be considered as a single locus despite
the multiplicity of genes that it bears. Inferences about population structure
from a single locus, particularly one that relies solely on maternal inheritance,
are questionable because they may not accurately reflect events across the gen-
ome as a whole.
22.2.6.2 Random amplification of polymorphic DNA (RAPD) and

amplified fragment length polymorphism (AFLP) analyses
Both of these techniques aim to analyse up to several hundred nuclear DNA loci
simultaneously (a locus is any specific sequence of DNA, whether a functional
gene or not). Both RAPD and AFLP are “anonymous” because the genome is
sampled randomly with no knowledge of the sequences that ultimately turn
up. Both methods have the advantage that no DNA sequence information is
required from the species under study before starting the investigation. The
techniques use oligonucleotide primers that can be purchased cheaply from bio-
technology companies.
22.2.6.2.1 RAPD analysis

Each PCR reaction includes a single, random-sequence primer just 10 nucle-
otides in length. Any 10-nucleotide sequence should occur by chance, on aver-
age, once every million base pairs along a DNA molecule. RAPD PCR produces
a product when two such sequences happen to occur in facing orientation and
relatively close to each other. These PCR products are then visualized as bands
on agarose gels following electrophoresis, after staining with ethidium bromide
(caution is required, as this is a mutagenic chemical), followed by photography
using a UV light source. More sophisticated analytical procedures are also pos-
sible (see AFLP section below). A primer may generate anything from no bands
to more than 20, and most of them will be the same for every individual, and
thus uninformative. It is normal practice to screen initially with many differ-
ent primers (maybe 50 or more) to identify those generating a useful number
of polymorphisms (i.e. those that produce bands in some individuals but not
others).
Despite the potentially tedious initial screening of multiple primers, this is the
simplest and cheapest type of DNA-based genetic analysis. However, for consist-
ent results it is vital to use the same reagents (such as the same manufacturer’s Taq
polymerase), the same PCR machine, and especially the same amount of DNA
(≈25 ng) in all analyses. Because of reproducibility problems, some high-profile
journals such as Molecular Ecology usually reject population studies based solely
on RAPD analyses. However, RAPDs remain an excellent approach to interspe-
cific problems such as the identification of cryptic species and life stages.
22.2.6.2.2 AFLP analysis

In this method, DNA samples are first digested with two restriction endo-
nucleases that cut at different, specific sequences (Figure 22.1). Short, double-
stranded DNA linkers of known sequence are then ligated onto the ends of the
DNA fragments, and after that the DNA is amplified by PCR using primers
partly complementary to the linkers, but with one to three random nucleotide
additions at the internal ends (i.e. overlapping the DNA fragments between
the linkers). This means that only a subset of the DNA fragments, those where
the sequences happen to complement the primer ends, will be amplified. Like
RAPDs, this procedure generates a large number of fragments. Polyacrylamide
gels (with higher resolving power than agarose) are usually used to separate them,
and the PCR products are labelled (e.g. with a radionuclide containing the 33P
isotope for detection by autoradiography, or fluorescently for analysis using
an automated DNA sequencer). All this is more complicated than RAPDs
and, because of the need for restriction endonuclease digestion, more DNA is
required. However, a major advantage of the technique, relative to RAPDs, is its
reproducibility. AFLPs are widely used for intraspecific population studies and
have been employed successfully with amphibians. Some investigations includ-
ing one with the alpine salamander Salamandra atra have indicated that they
may not reveal high-resolution population structure (Riberon et al. 2004), but
it is commonplace for AFLP analysis to generate several hundred polymorphic
markers, and thus be broadly representative of the genome.
22.2.6.2.3 Microsatellite analysis

Microsatellites consist of di-, tri-, or tetranucleotides repeated multiple times at a
particular locus (e.g. GTGTGTGTGT, GTCGTCGTC, GTACGTACGTAC).
Most vertebrate genomes have hundreds or thousands of microsatellite loci
spread across their nuclear chromosomes, and almost all are in non-coding (i.e.
neutral) regions. Microsatellites have even higher mutation rates than mtDNA,
and largely for this reason have become the marker of choice for most popula-
tion studies (Jehle and Arntzen 2002). Microsatellite loci are amplified by the
PCR using primers complementary to their flanking regions. Alleles differ in
the number of repeats, and thus in size, and can be separated and identified by
(a) AFLP template preparation Msel adaptor
Whole genomic DNA G

CAT
EcoRI adaptor
Restriction enzymes
+ (Msel + EcoRI) and + AATTG
DNA ligase C
(b) Restriction and ligation

Msel cut EcoRI cut
T T A A G A A T T C
T C T T A A G
A A T
Msel adaptor EcoRI adaptor
G AA T T G
CAT C
G T A A G A A T C G
C A T T C T T A A C
(c) Selective amplification (one of many primer combinations shown)
Msel primer 1
C A T T G T A
G T A A C A T C G A G A A T T G
C A T T G T A G C T C T T A A C
C G A G A A T T G
EcoRI primer 1
Fig. 22.1 Basis of AFLP analysis using the restriction enzymes MseI and EcoRI.
electrophoresis (Figure 22.2, natterjack toads) or on automated DNA sequencers.

It is often possible to “multiplex” microsatellite loci, meaning to amplify several
at a time in the same reaction tube, using primers with different fluorescent
labels.
The main problem with microsatellites is characterizing them for a new spe-
cies. To design primers it is necessary to know the flanking sequences, and par-
ticularly in the amphibia these seem to be poorly conserved even among species
M13 M13
GA/TC GA/TC
Hom
11
130
128
Hom
Fig. 22.2 Gel analysis of two microsatellite loci in 20 individuals. M13 GA/TC are
size markers, Hom show examples of homozygotes. Lanes with two band sets are
heterozygotes. Other numbers are reference samples.
in the same genus (Primmer and Merila 2002). It is nevertheless always worth
trying primers from a previously characterized close relative if these are avail-
able (Molecular Ecology Resources and Conservation Genetics publish details for
dozens of new species every year), but often it is necessary to start from scratch.
This means creating a genomic library enriched for microsatellite sequences by
selective hybridization, screening clones for microsatellites, sequencing positive
clones, and then designing primers from the flanking sequence information
(Zane et al. 2002). All of this is expensive and time-consuming. The rewards,
however, can be great. Both RAPD and AFLP analyses generate “dominant”
markers, in which heterozygotes cannot be distinguished from homozygotes.
Microsatellites, however, are “codominant” markers because, due to differing allele

sizes, homozygotes and heterozygotes are readily distinguished (Figure 22.2).
This provides very valuable extra information. Typically, seven or eight poly-
morphic microsatellite loci are sufficient for basic fine-scale analysis of popu-
lation structure, and to identify individual animals with negligible (
10 −10)
probabilities of mistakes.
22.3 Analysis of genetic data

The results of genotype analyses are usually in one of three forms: DNA
sequences (e.g. mtDNA haplotypes), binary (presence or absence of a band for
dominant RAPD or AFLP data), or allele composition (such as 120.160, the
sizes of two alleles in a microsatellite heterozygote). Many computer programs
are available as freeware for analyses of all these data types.
22.3.1 Cryptic species or life-stage identification

This usually just requires visual inspection of gel photographs following RAPD
analyses. The banding patterns (RAPD “phenotypes”) should be distinct and
reproducible for each species, and consistent across multiple widely separated
populations. There have been numerous RAPD identification successes with
amphibians (e.g. Bardsley et al. 1998; Mikulicek and Pialek 2003).
22.3.2 Genetic diversity, inbreeding, and bottlenecks

Analysis of mtDNA sequences involves establishing the extent of change between
haplotypes, as the numbers of mutational events that have led (usually) to single
nucleotide polymorphisms (SNPs). Software packages such as DNASTAR and
T-COFFEE align multiple DNA sequences and highlight SNPs. ARLEQUIN
will use the full sets of haplotypes from different population samples and quan-
tify genetic diversity in each population, in terms of both haplotype diversity and
nucleotide diversity. The two measures are not identical, because haplotypes can
vary by more than one SNP. For dominant data such as RAPDs and AFLPs, gen-
etic diversity is commonly estimated as gene diversity, defined as 1 Σ(Pi)2, where
Pi is the frequency of the ith allele, assuming a biallelic state for each locus and a
state of Hardy–Weinberg equilibrium in the population. Most commonly, how-
ever, genetic diversity is estimated from codominant data such as those provided
by microsatellites. Many programs, such as GENEPOP and FSTAT, calculate
heterozygosities, allelic richness (mean numbers of alleles per locus), and percent-
age of loci that are polymorphic in a population. It is important to test (the soft-
ware provides methods) for concordance with Hardy–Weinberg equilibrium, and
to ensure that the loci are in linkage equilibrium. Microsatellites often have tens
of alleles, and mean heterozygosity estimates across loci in large populations are
commonly more than 0.6.
Low levels of genetic diversity are possible indicators of inbreeding depression
or high genetic load. This was true in a study of natterjack toads Bufo calamita
(Rowe and Beebee 2003). However, there are important caveats on this inter-
pretation. Firstly, the estimates must be low relative to other populations, for
exactly the same suite of loci. Second, low genetic diversity at neutral loci can arise
for various reasons apart from inbreeding, such as proximity to a range edge after
postglacial colonization, as with Italian agile frogs Rana latastei (Ficoleta et al.
2007). Such populations were fully viable, whereas isolated populations with
equally low genetic diversity showed fitness reductions (Figure 22.3). It is there-
fore important to test directly for low fitness in suspect populations (for example
by measuring embryonic or larval survival and growth rates) rather than relying
solely on genetic diversity measures. However, another test can be applied to
any population with data from multiple codominant loci. When population size
declines rapidly, allelic richness is lost much faster than average heterozygosity.
BOTTLENECK quantifies this “heterozygote excess,” which is transient but
extends over several generations during and after a decline (Luikart et al. 1998).
This approach accurately identified known recent population crashes in natter-
jack toads, B. calamita (Beebee and Rowe 2001).
0.98
1.00
0.89 0.90
0.80 0.77
0.75
0.54
0.48
Hatch rate
0.38
0.50
0.25
0.00
AL CU MZ T1 T2 T3 A1 A2
Fig. 22.3 Low fitness of isolated Rana latastei populations in Italy. Three isolated
populations ( 3 km from the nearest neighbor) are shown as gray bars, and five
non-isolated populations as white bars. Standard errors are shown, based on more
than 25 egg clutches per population.
22.3.3 Identification of barriers to movement

A powerful use of genetic data is to identify barriers limiting, or corridors
facilitating, the migration of animals. There are two main ways of doing this.
Programs such as ARLEQUIN, GENEPOP, and FSTAT estimate F statistics
from any type of genetic data, especially FST, which is a measure of genetic
divergence between two populations. FST varies from 0 (no differentiation, the
two “populations” are really just one) to 1 (complete distinction). FST values
over 0.2 imply such low gene flow that random genetic drift will dominate over
migration, and the populations will become increasingly divergent over time.
More precise interpretations of migrant numbers are now shunned because they
make unverifiable assumptions (Whitlock and McCauley 1999). Nevertheless,
F statistics are still useful, especially in multiple pairwise comparisons between
populations. Isolation by distance (or IBD), a significant correlation between
geographical distance and FST, implies regular historical gene flow (i.e. migra-
tion) between populations, and unusually high FST suggests the existence of
barriers to movement. Many programs such as GENEPOP have algorithms for
assessing isolation by distance. Gene flow and isolation by distance in amphib-
ians based on F statistics have been widely reported, for example with mtDNA
in the case of the Yosemite toad Bufo canorus in California (Shaffer et al. 2000)
and with microsatellites in the Cascades frog Rana cascadae (Monsen and Blouin
2004).
A problem with FST is that it assesses movement as an average over some
unknown past time period. Assignment methods based on multilocus genetic
data give a more precise and contemporary indication of gene flow. These cluster
genotypes into populations, and then highlight genotypes which were sampled
in the “wrong” place (i.e. which genetically belonged to one population, but
were found in another). Such methods that can identify individual migrants
are computationally intensive but very powerful. STRUCTURE and BAPS are
perhaps the best known. It is even possible to detect, using BAYESASSNM,
the first-generation progeny of migrants by genotyping amphibian larvae, as
shown with crested and marbled newts Triturus cristatus and Triturus marmora-
tus (Jehle et al. 2005a). There are many published studies of amphibian popula-
tion structures, mostly based on microsatellite analyses, from all over the world.
Examples include North American wood frogs, Rana sylvatica (Newman and
Squire 2001), European crested newts (Jehle et al. 2005b), and Australian tree-
frogs, Litoria aurea (Burns et al. 2004).
Landscape genetics defines the spatial arrangement of populations using gen-
etic data and identifies both impermeable habitats (barriers) and permeable ones
(corridors) between population clusters. A study with the Columbia spotted frog
Rana luteiventris exemplifies this approach (Funk et al. 2005). Programs such
as GENELAND explicitly combine both genetic and geographical informa-
tion for landscape studies. Figure 22.4 shows an example of population differ-
entiation in the long-toed salamander Ambystoma macrodactylum, determined
using seven microsatellite loci, mediated primarily by altitude in Montana, USA
(Giordano et al. 2007). Relevance to conservation is exemplified by a study of
eastern red-backed salamander (Plethodon cinereus) populations, which were
much more fragmented in semi-urban than in pristine habitats in Canada (Noel
et al. 2007). This approach can therefore reveal the significance of urban devel-
opment as disrupters of population connectivity.
22.3.4 Defining populations and population size

Software such as STRUCTURE, BAPS, and GENELAND can test whether
genetic clustering corresponds to individual aggregations, or to groups of
them. STRUCTURE, for example, can take a full genotype data set from
multiple sampling sites and make no initial assumptions about how many
populations the data represent. It can then estimate the number of “true”
(genetically distinct) populations and list the individuals that constitute each,
as with natterjack toads in Britain (Rowe and Beebee 2007). This requires a
lot of genetic information; preferably from more than 10 highly polymorphic
loci, and more than 30 individuals per sampling site. Genetic data can also be
MEL
Rock
RM3
RM4 Meadow
P = 0.002
CG
RM1 LEL
RM2
SP Bear
BAM 410 M
SL MM 26 BOHS
RCM RCP High vs. low
altitude
PRT P = 0.01255
TBP Bull river
BLH
Wpt 35 P = 0.0004
Fig. 22.4 Population genetic structure in long-toed salamanders. Spheres are

high-altitude sites, cubes are low-altitude sites, and sizes are proportional to
intrapopulation microsatellite variation. Connection lengths are proportional to site
similarity, and the probabilities of genetic distinction between low- and high-altitude
groups, as well as the distinctiveness of two main clusters, are also shown.
used to estimate effective population size, Ne, once populations are defined.
Ne is virtually always lower than census size N, on average by a factor of 10 in
wildlife populations. Ne can be estimated most accurately by two samplings of
a population, preferably separated by an interval of several generations (so 5
years for most amphibians), followed by genotyping each sample set across
multiple loci. The larger the change in allele frequencies, the greater has been
the effect of genetic drift, and thus the smaller is the Ne. Programs such as
NeESTIMATOR and TM3 are available for these calculations. An alterna-
tive is to estimate Nb, the numbers of successfully breeding adults in a single
generation, by collecting genetic data from parents and offspring in the same
year. This is more convenient than waiting for several generations, but has
large error limits especially when the parental sample includes more than one
age cohort. Most estimates of Ne in amphibian populations have been low,
often in the tens or hundreds (Beebee and Griffiths 2005).
22.3.5 Historical issues

To show the relationships between population clusters, tree-building programs
such as PHYLIP or PAUP are required. Statistical rigour is provided by boot-
strapping, which involve random resampling of the primary data set to put
percentage confidence estimates on each tree branch. Trees are central to phylo-
geography (the application of intraspecific phylogenetics), and provide valuable
insights into population origins and histories. They can be generated in a var-
iety of ways using any type of genetic data. However, a further level of analysis
is possible with mtDNA sequences. Haplotype sequences can be connected as
networks, showing in detail their likely evolutionary histories. These in turn
can be subject to nested clade analysis to infer historical events such as range or
population expansion (although the reliability of nested clade analysis is contro-
versial), using software packages such as TCS and GEODIS. The Miocene and
Pleistocene histories of several amphibian species have been unravelled using
these techniques, including (for example) Rana arvalis in Europe (Babik et al.
2004), Bufo fowleri in North America (Smith and Green 2004), and the origins
of tropical amphibian distributions such as Dendrobates tinctorius (Noonan and
Gaucher 2006) in South America and Xenopus species in Africa (Evans et al.
2004). The tailed frog Ascaphus provides an example (Figure 22.5) of how phy-
logeography has improved understanding of Pacific northwest biogeography.
Inland and coastal populations were separated several million years ago into two
distinct species (Ascaphus montanus and Ascaphus truei), whereas internal dif-
ferences among populations in each region were probably generated in refugia
during the Pleistocene glaciations (Nielson et al. 2001).
(a) DD
Q
C
63 G
63
F
20
19 E
Palouse Range
29 B Bitterroot Range
48
6 P Clearwater Mts
61 <5 Seven-Devils Mts
J Northern Rocky Mts
55
D Salmon River Mts
Blue Mts
A
22
20
AA
CC
95 L
100
BB
Inland 100 K
66
100 64 M
East Fk. of the South Fk. o
98 N the Salmon R.
100
O
21
87 Z
97 Northern Cascade Mts
18 H
88 X
88 85 Y Olympic Mts
93 31
40 W
93
R
Coastal 99 98 S Siskiyou Mts
98 T
96 U Oregon Coast Range
100 V
0.005 substitutions/site
(b)
2-3 2-6 1-15 T
R 0 0 0 0 0 0 0 0
1-16
101 steps 1-10
S O N M
3-5
BB
23 steps
X 22 steps 1-8 0
1-18 0
3-4
3-3 2-4
0
2-11 0 0
0 0 0 W 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Z
0
B 1-4 P 1-1 3-2
0 0
1-13
2-51-9 0 J 2-1
2-10 0 2-9 1-17 2-7 0
2-7 L 0 0 0 A E F G
0 26 steps 0 D 0 1-3
U 0 0 V 2-8 0 C 0
1-19 1-11 1-12 2-3 Q 1-2
0
Y AA
0 DD
0 1-6
0 0
1-14
1-7 2-2
Coastal H
K
01-5
CC
4-1
3-1
Fig. 22.5 Relationships among tailed frog populations, based on mtDNA. (a)
Phylogenetic tree. (b) Haplotype network in nested clades.
Phylogeography has also proved useful when investigating the origins of

introduced species, such as marine toads Bufo marinus in Australia (Estoup et al.
2001). This approach also demonstrated that the pool frog Rana lessonae was a
longstanding native of Britain rather than a nineteenth century introduction as
previously believed (Zeisset and Beebee 2001).
22.3.6 Behavior and sexual selection

Work in this area requires genotype information that can identify individuals
with minimal error rates. Suites of highly polymorphic microsatellite loci are
the most commonly used to generate unique DNA fingerprints (or profiles).
It is then possible to investigate multiple paternity and individual reproduct-
ive success. Molecular markers demonstrated, for example, multiple paternity
as a result of “clutch piracy” in the European common frog Rana temporaria
(Vieites et al. 2004). In the salamander Plethodon cinereus multiple paternity
occurs even in species that seems socially monogamous (Liebgold et al. 2006).
Computer programs such as GENECLASS, KINSHIP, and CERVUS carry
out such inferences.
22.4 Future developments

The range of neutral molecular markers available for genetic studies on wild
populations has increased dramatically in recent decades. What will happen
next? Molecular ecology is expensive, and perhaps the greatest hope is that
costs will continue to decline as the necessary technology and consumables
become more widely available. Nevertheless, collaboration between ecologists,
conservation biologists, and molecular biologists will remain essential for the
foreseeable future. Although microsatellites seem likely to dominate the neutral-
marker field for some time, other possibilities are on the horizon. SNP analysis
of nuclear DNA sequences, for example, has some promise. For adequate power
this requires 100 loci or more, but high-throughput genotyping makes nuclear
SNP studies increasingly feasible (Morin et al. 2004). The main advantage is
that the mutation mechanism for SNPs is simpler than that for microsatellites,
permitting more rigorous statistical approaches to data analysis.
However, it is in the field of adaptive rather than neutral variation that we
can expect to see the most exciting new developments. It is increasingly import-
ant to assess genetic loci directly related to fitness, rather than relying on the
surrogacy of neutral loci. There are already several ways of doing this, some
of which are non-molecular. It is possible to carry out controlled crosses in
the laboratory and, with half-sib experimental designs, to relate the fitness of
offspring to parents from different source populations. The genetic and envir-
onmental contributions to critical attributes such as larval growth rates and
size at metamorphosis can be determined in this way. There are also statistical
approaches for estimating population differentiation based on such quantita-
tive traits, using Q ST as an analogue of FST. Comparisons of these two statistics
can be particularly informative. Where Q ST FST, the implication is that selec-
tion rather than random drift dominates any genetic differentiation (and vice
versa). Work along these lines has been very informative with R. temporaria
(Laugen et al. 2005)
Breeding experiments are time-consuming and, like molecular methods,
require specialized laboratory facilities. It can also be problematic to work on a
large scale, and sacrifice of animals (especially males to provide sperm) is often
necessary. There are, however, also non-destructive molecular approaches to
the study of adaptive variation. The use of highly polymorphic microsatellites
can, after analysis with programs such as KINSHIP, determine kin relations
retrospectively in samples from wild populations and thus reduce the need for
controlled crosses (although this cannot be applied to interpopulation compari-
sons). Fortunately it is now possible to identify loci involved in adaptive vari-
ation and investigate them directly. One such approach is genome scanning. In
this method, AFLPs are used initially to generate multiple polymorphic bands.
Each band (locus) is then analysed separately, in cross population comparisons,
to estimate pairwise FST values. Most loci will be neutral, but those with FST
estimates significantly greater than predicted under neutrality can be identi-
fied. These (or sequences in linkage disequilibrium with them) may be under
directional selection, between (for example) different habitat types. Once iden-
tified, diversity at these loci can be quantified to generate a “population adaptive
index”. These anonymous loci represent a valuable comparison with definitely
neutral loci and are thus a significant step forward. This method has been used
successfully with R. temporaria, tentatively identifying loci associated with
adaptation to different altitudes (Bonin et al. 2007).
Identifying loci definitely under selection is still problematic for most species,
but an exception is the major histocompatibility complex (MHC) of proteins
involved in the cellular immune response. These are the most polymorphic
protein-encoding genes known in vertebrates, and are under strong selection
to maintain resistance to pathogens. Understanding this defence mechanism
is particularly relevant in an age of global amphibian declines at least partly
mediated by chytrid fungi. PCR primers for amplification of the most variable
part of MHC class II loci (exon 2) are now available for some species such as
North American ambystomatid salamanders (Bos and DeWoody 2005), and
population studies will certainly follow. There are technical difficulties with
MHC, largely because most animals have multiple loci that can be hard to
distinguish, but this gene family has great prospects for estimating adaptive
variation in wild populations.
Genetic studies have already told us a lot about amphibians that we could not
have discovered without them. Much more is undoubtedly in store, and this will
remain an exciting and challenging research area for decades ahead.
22.5 References
Babik, W., Branicki, W., Sandera, M., Lutrinchuk, S., Borkin, L. J., Irwin, J. T., and
Rafinski, J. (2004). Mitochondrial phylogeography of the moor frog, Rana arvalis.
Bardsley, L., Smith, S., and Beebee, T. J. C. (1998). Identification of Bufo larvae by
molecular methods. Herpetological Journal, 8, 145–8.
Beebee, T. J. C. (2005). Amphibian conservation genetics. Heredity, 95, 423–7.
Beebee, T. J. C. and Rowe, G. (2001). Application of genetic bottleneck testing to the
investigation of amphibian declines: a case study with natterjack toads. Conservation
Biology, 15, 266–70.
Beebee, T. J. C. and Griffiths, R. A. (2005). The amphibian decline crisis: a watershed for
conservation biology? Biological Conservation, 125, 271–85.
Beebee, T. J. C. and Rowe, G. (2008). An Introduction to Molecular Ecology, 2nd edn.
Oxford University Press, Oxford.
Bonin, A., Nicole, F., Pompanon, F., Miaud, C., and Taberlet, P. (2007). Population
adaptive index: a new method to help measure intraspecific genetic diversity and priori-
tise populations for conservation. Conservation Biology, 21, 697–708.
Bos, D. H. and DeWoody, J. A. (2005). Molecular characterisation of major histocom-
patibility complex class II alleles in wild tiger salamanders (Ambystoma tigrinum).
Immunogenetics, 57, 775–81.
Burns, E. L., Eldridge, M. D. B., and Houlden, B. A. (2004). Microsatellite variation and
population structure in a declining Australian Hylid Litoria aurea. Molecular Ecology,
13, 1745–57.
Estoup, A., Wilson, I. J., Sullivan, C., Cornuet, J. M., and Moritz, C. (2001). Inferring
population history from microsatellite and enzyme data in serially introduced cane
toads, Bufo marinus. Genetics, 159, 1671–87.
Evans, B. J., Kelley, D. B., Tinsley, R. C., Melnick, D. J., and Cannatella, D. C. (2004). A
mitochondrial DNA phylogeny of African clawed frogs: phylogeography and implica-
tions for polyploid evolution. Molecular Phylogenetics & Evolution, 33, 197–213.
Ficoleta, G. F., Garner, T. W. J., and DeBernardi, F. (2007). Genetic diversity, but not
hatching success, is jointly affected by postglacial colonisation in the threatened frog,
Rana latastei. Molecular Ecology, 16, 1787–97.
Funk, W. C., Blouin, M. S., Corn, P. S., Maxell, B. A., Pilliod, D. S., Anrish, S., and
Allendorf, J. W. (2005). Population structure of Columbia spotted frogs (Rana luteiv-
entris) is strongly affected by the landscape. Molecular Ecology, 14, 483–96.
Giordano, A. R., Ridenhour, B. J., and Storfer, A. (2007). The influence of altitude and
topography on genetic structure in the long-toed salamander (Ambystoma macrodacty-
lum). Molecular Ecology, 16, 1625–37.
Jehle, R. and Arntzen, J. W. (2002). Microsatellite markers in amphibian conservation
genetics. Herpetological Journal, 12, 1–9.
Jehle, R., Burke, T., and Arntzen, J. W. (2005a). Delineating fine-scale genetic units in
amphibians: probing the primacy of ponds. Conservation Genetics, 6, 227–34.
Jehle, R., Wilson, G. A., Arntzen, J. W., and Burke, T. (2005b). Contemporary gene flow
and the spatio-temporal genetic structure of subdivided newt populations (Triturus
cristatus, T. marmoratus). Journal of Evolutionary Biology, 18, 619–28.
Laugen, A. T., Kruuk, L. E. B., Laurila, A., Rasanen, K., Store, J., and Merila, J. (2005).
Quantitative genetics of larval life-history traits in Rana temporaria in different envir-
onmental conditions. Genetical Research, 86, 161–70.
Liebgold, E. B., Cabe, P. R., Jaeger, R. G., and Leberg, P. L. (2006). Multiple paternity in
a salamander with socially monogamous behaviour. Molecular Ecology, 15, 4153–60.
Livia, L., Antonella, P., Hovirag, L., Mauro, N., and Panara, F. (2006). A non-destructive,
rapid, reliable and inexpensive method to sample, store and extract high-quality DNA
from fish body mucus and buccal cells. Molecular Ecology Notes, 6, 257–60.
Luikart, G., Sherwin, W. B., Steele, B. M., and Allendorf, F. W. (1998). Usefulness of
molecular markers for detecting population bottlenecks via monitoring genetic change.
Mikulicek, P. and Pialek, J. (2003). Molecular identification of three crested newt species
(Triturus cristatus superspecies) by RAPD markers. Amphibia-Reptilia, 24, 201–7.
Monsen, K. J. and Blouin, M. S. (2004). Extreme isolation by distance in a montane frog
Rana cascadae. Conservation Genetics, 5, 827–35.
Morin, P. A., Luikart, G., and Wayne, R. K. (2004). SNPs in ecology, evolution and con-
servation. Trends in Ecology and Evolution, 19, 208–16.
Newman, R. A. and Squire, T. (2001). Microsatellite variation and fine-scale population
structure in the wood frog (Rana sylvatica). Molecular Ecology, 10, 1087–1100.
Nielson, M., Lohman K., and Sullivan, J. (2001). Phylogeography of the tailed frog
(Ascaphus truei): implications for the biogeography of the Pacific Northwest. Evolution,
55, 147–60.
Noel, S., Ouellet, M., Galois, P., and Lapointe, F. J. (2007). Impact of urban fragmen-
tation on the genetic structure of the eastern red-backed salamander. Conservation
Genetics, 8, 599–606.
Noonan, B. P. and Gaucher, P. (2006). Refugial isolation and secondary contact in the
dyeing poison frog Dendrobates tinctorius. Molecular Ecology, 15, 4425–35.
Pidancier, N., Miquel, C., and Miaud, C. (2003). Buccal swabs as a non-destructive tissue
sampling method for DNA analysis in amphibians. Herpetological Journal, 13, 175–8.
Plötner, J., Kohler, F., Uzzell, T., and Beerli, P. (2007). Molecular systematics of amphib-
ians. In H. Heatwole (ed.), Amphibian Biology, vol. 7, Systematics, pp. 2672–2756.
Surrey Beatty & Sons, Chipping Norton, NSW.
Primmer, C. R. and Merila, J. (2002). A low rate of cross-species microsatellite amplifica-
tion success in Ranid frogs. Conservation Genetics, 3, 445–9.
Riberon, A., Miaud, C., Guyetant, R., and Taberlet, P. (2004). Genetic variation in an
endemic salamander, Salamandra atra, using amplified fragment length polymorph-
ism. Molecular Phylogenetics & Evolution, 31, 910–14.
Rowe, G. and Beebee, T. J. C. (2003). Population on the verge of a mutational meltdown?
Fitness costs of genetic load for an amphibian in the wild. Evolution, 57, 177–81.
Rowe, G. and Beebee, T. J. C. (2007). Defining population boundaries: use of three
Bayesian approaches with microsatellite data from British natterjack toads (Bufo calam-
ita). Molecular Ecology, 16, 785–96.
Shaffer, G., Fellers, G. M., Magee, A., and Voss, R. (2000). The genetics of amphibian
declines: population substructure and molecular differentiation in the Yosemite Toad,
Bufo canorus (Anura, Bufonidae) based on single-strand conformation polymorph-
ism analysis (SSCP) and mitochondrial DNA sequence data. Molecular Ecology, 9,
245–57.
Smith, M. A. and Green, D. M. (2004). Phylogeography of Bufo fowleri at its northern
range limit. Molecular Ecology, 13, 3723–33.
Vences, M. and Wake, D. B. (2007) Speciation, species boundaries and phylogeography of
amphibians. In H. Heatwole (ed.), Amphibian Biology, vol. 7, Systematics, pp. 2613–71.
Surrey Beatty & Sons, Chipping Norton, NSW.
Vieites, D. R., Nieto-Roman, S., Barluenga, M., Palanca, A., Vences, M., and Meyer, A.
(2004). Post-mating clutch piracy in an amphibian. Nature, 431, 305–8.
Walsh, P. S., Metzger, D. A., and Higuchi, R. (1991). Chelex R100 as a medium for sim-
ple extraction of DNA for PCR-based typing from forensic material. Biotechniques, 10,
506–13.
Whitlock, M. C. and McCauley, D. E. (1999). Indirect measures of gene flow and migra-
tion: F-ST not equal 1/(4Nm+1). Heredity, 82, 117–25.
Zane, L., Bargelloni, L., and Patarnello, T. (2002). Strategies for microsatellite isolation:
a review. Molecular Ecology, 11, 1–16.
Zeisset, I. and Beebee, T. J. C. (2001). Determination of biogeographical range: an
application of molecular phylogeography to the European pool frog Rana lessonae.
Proceedings of the Royal Society of London Series B Biological Sciences, 268, 933–8.
Part 7
Monitoring, status, and trends
23
Selection of species and sampling
areas: the importance to inference
Paul Stephen Corn
23.1 Introduction
Inductive inference, the process of drawing general conclusions from specific
observations, is fundamental to the scientific method. Platt (1964) termed
conclusions obtained through rigorous application of the scientific method
as “strong inference” and noted the following basic steps: generating alterna-
tive hypotheses; devising experiments, the results of which will exclude one
or more hypotheses; conducting the experiment to get a “clean result”; and
repeating the process with revision based on the information obtained. Every
student is exposed to these basics in introductory courses, and a consider-
able proportion of a modern graduate education in the sciences is devoted
to acquiring the analytic (statistical) skills necessary to apply the scientific
method. Not even considering the field of mathematical statistics or applied
statistics in disciplines such as social sciences, library shelves groan under the
weight of texts on applied statistics, ranging from introductory (Hayek and
Buzas 1997) to advanced (Williams et al. 2002), for conducting research in
ecology, and new works are published every year. Much effort is currently
devoted to the mechanisms of analysis and the issues involved in choosing
among statistical methods; specifically, traditional hypothesis testing versus
information-theoretic or Bayesian approaches (Hobbs and Hilborn 2006).
This chapter does not address these topics, but instead discusses some of
the issues related to selection of study sites and species necessary to obtain
a “clean result.” Regardless of the method of analysis, successful inference
relies on correctly designed data collection, meaning that the observations
represent the population of interest. As Anderson (2008, p. 7) puts it, valid
inference requires, “some type of probabilistic sampling of the well-defi ned

population.”
Experimental design and sampling are well-developed topics in applied stat-
istics. There are numerous books available (see Anderson 2001 for a sample of
representative works), and many universities offer their graduate students spe-
cific courses in these areas. Most biologists have had the requisite education and
are familiar with and employ statistically valid designs. However, good study
designs are not universal. Anderson (2008, p. 143) stated:
. . . there is no excuse for collecting data that are fundamentally flawed. Still, I see data collected
from convenience sampling where any valid inductive inference is precluded. In some cases the
population of interest is not even defined.
Convenience sampling is a broad term that describes non-random study-site

selection (Hayek 1994; Anderson 2001). Inadequate design and sampling is
often justified on logistical grounds, or, to put it more crudely, that it is unreal-
istic to satisfy the demands of a statistician, who is usually office-bound in an
ivory tower. As a field biologist and not a statistician, I sometimes find myself
sympathetic to such rationalizations. However, the solution is not to politely
thank the statistician and then go ahead and sample as originally planned (see
the first phrase in the Anderson quote above). Instead, more planning is neces-
sary. Either a means must be devised for obtaining a valid sample, or the pro-
ject’s objectives should be revised so that a population can be defined that
will allow a valid sample, or, in extreme cases, the project might need to be
abandoned.
In this chapter, I provide a brief overview of sampling design, but this is a
large, complex topic and it is impossible to do more than scratch the surface
here. For more thorough instruction and discussion, including statistics spe-
cific to different designs, consult one of the many texts that have been writ-
ten for that purpose (e.g. Cochran 1977; Thompson 1992; Hayek and Buzas
1997; Williams et al. 2002). The US Geological Survey (www.pwrc.usgs.gov/
monmanual/) and National Park Service (http://science.nature.nps.gov/im/
monitor/) also have extensive and very useful web-based resources to aid in
study design for inventory and monitoring studies. Following this introduction
to sampling, I will discuss issues related to selection of study sites and effects
of estimating abundance on inference. This discussion will use a few selected
cases to illustrate problems that can arise from either poor study design or inter-
pretation of results that cannot be supported by the design. These examples
include issues arising from both selection of study sites and collection and ana-
lysis of data from individual species, and describe some of the methods that can
be used to minimize bias.
23 Selection of species and sampling areas | 433
23.2 Sampling
In all but a tiny number of exceptional cases, it is impossible to survey every
possible habitat or catch every individual in a population. Therefore, inference
depends on statistics generated from a subset of individuals or habitats drawn
from the population. Ideally, a sample has three qualities (Williams et al. 2002):
it comprises separate individual units, and these share the same underlying dis-
tribution and are statistically independent. These criteria are more difficult to
satisfy in inventory and monitoring studies than in a carefully designed experi-
ment, but a good study design should try to come as close as possible.
Before a sample can be taken, the sampling units must be defined and be
available for selection. Sample units can be individual animals, artificially
bounded areas (quadrats), or natural habitat features, such as ponds, streams,
or drainage basins. For population studies, where the individual is the sam-
ple unit, the availability for selection is often assumed, but this assumption is
tested when the data are analyzed. For inventory or monitoring studies, site
delineation and selection has become easier with the increasing availability
of tools such as satellite imagery and Geographic Information Systems (GIS)
software. A good example is the study by Kroll et al. (2008), which took advan-
tage of detailed GIS data layers to select stream reaches for sampling. However,
detailed GIS coverage exists mainly where there is an economic use for the
data (e.g. the commercial forests surrounding the streams studied by Kroll
et al. 2008), and even when GIS data are available the amphibian habitats may
be poorly described. Temporary wetlands are notoriously underrepresented in
aquatic data layers, and studies of lentic-breeding amphibians often will have
more success in defining sampling units if areas are used instead of specific
habitats, for example, 1 km2 blocks (Johannson et al. 2006) or drainage basins
(Corn et al. 2005).
The most basic form of probabilistic sampling is a simple random sample.
All the available sample units are put into a metaphorical hat (the mechanics
of choosing a sample now usually involve a computer) and the desired num-
ber of units are selected in random order. The number of sample units should
be sufficient to estimate parameters of interest with sufficient precision, but
samples that are too large should be avoided. Determining the desired sample
size requires knowing the variability of the parameter to be estimated, the mag-
nitude of the effect to detect (e.g. a 10% difference or a 5% annual trend), and
the strength of the inference. The formula for determining sample size depends
on the sampling scheme. Hayak and Buzas (1997) and Williams et al. (2002)
describe the details for determining sample size.
If, as is usually the case, habitats are not uniform, species are patchily dis-
tributed, or the number of samples is small relative to the area of interest, sim-
ple random sampling may be inadequate to characterize the variability of the
system being studied. There are several slightly more complex designs that can
be used to achieve a more representative sample. In the case where the num-
ber of samples is relatively small, strictly random selection can result in sample
sites clumped together instead of dispersed throughout the study area. If the
study area is not uniform, this is not desirable, and a common modification is
to employ a systematic random sample (Figure 23.1), in which, after a random
start, every nth unit is selected, where the number of available units divided by n
equals the desired sample size. Williams et al. (2002) cautioned that systematic
sampling risks a biased result if the units are arranged in such a way that envir-
onmental gradients are correlated with the order of the sampling units.
0 10 20 km
Fig. 23.1 Sample locations for monitoring amphibian populations in Glacier

National Park, MT, USA. Twenty small drainage basins (shaded) composed a
systematic random sample from a frame of 235 basins (outlined). Areas without
sufficient surface water, where the elevation precluded amphibian occurrence, or
where extremely rugged terrain prevented safe access were excluded from the frame.
Inference about status of amphibians in the park is limited to the frame.
If a study area is not homogenous and differences within the area can be
described (e.g. altitudinal gradients, different types of wetlands, differences in
ability to access potential sample units) then a stratified design is usually pref-
erable to a simple random or systematic random sample. Once the strata and
sampling units within them are delineated, then simple or systematic random
samples can be drawn from each stratum. The formulas for calculating various
statistics vary among sampling designs. See Williams et al. (2002) for a concise
description of these. For long-term monitoring studies, strata should be based
on features that are not expected to change significantly during the course of the
study. For example, strata based on geological differences would be preferable to
strata based on vegetative land cover.
This section is not a comprehensive treatment of sampling design. Advanced
schemes, such as adaptive sampling (Thompson 1992; Williams et al. 2002),
are beyond the scope of this chapter, but may be necessary or desirable in many
cases. Sources listed above should be consulted when beginning the design of
any study.
Any valid sampling design must incorporate replication of sample units. A study
that compared toad populations in only two ponds could only describe the dif-
ferences in abundance between the two ponds. Tests of hypotheses of the causes
of the differences are not appropriate because there is no replication (Underwood
1998). Most inventory or monitoring studies include numerous sample units, but
care must be taken to ensure that external factors are dealt with in the design stage
(e.g. through stratification). If variation among sample units can be attributed to
something other than random processes, then the study would suffer from pseu-
doreplication (Hurlbert 1984). Large field studies are vulnerable to temporal pseu-
doreplication (sample units vary in some systematic fashion over time), because
it is seldom possible to visit all sample units simultaneously. This is a particular
problem for studies of amphibians. For example, surveys that use breeding activ-
ity are affected by the changing composition of breeding choruses over time, both
among and within species, and external factors, such as weather, that influence
behavior of individuals. Studies that focus on larval stages must deal with growth
and development and changes in abundance that may influence detection. It is
always a good practice to minimize the time span of field surveys where possible.
23.3 Study sites and consequences of

convenience sampling
If study sites or sample units are not selected using a probalistic design, the result
is a convenience sample. The motivations for convenience sampling are several,
but common reasons are to choose study sites that are most accessible, or where
the target species are most abundant. The latter case poses particular problems for
analysis of trends in abundance, because there may be a built-in bias for detecting
declining populations (Alford and Richards 1999; but see Green 2003).
The effects of convenience sampling on inference can be subtle. Results of
ecological experiments are often interpreted to demonstrate the generality of
effects or even causality, but experiments by themselves are insufficient to explain
complex ecological phenomena; such an effort requires integrating observation
and theory with experimentation (Werner 1998). Broad inference is limited
when experiments, even those that are internally well designed, are conducted
at locations that are convenient for the researcher. The results may be useful for
investigating possible mechanisms, but cannot be generalized without making
the unsupportable assumption that the study sites are an unbiased represen-
tation of the habitats in question. For example, Blaustein et al. (1994) found
that ambient ultraviolet-b (UV-B) radiation caused higher mortality of amphib-
ian embryos than when embryos were shielded from UV-B at four lakes in the
Cascade Mountains in Oregon, USA. This and subsequent research formed the
basis of a theory of how climate change, UV-B, and pathogens might play a sig-
nificant role in global amphibian declines (Kiesecker et al. 2001; Pounds 2001;
Blaustein and Kiesecker 2002). The issue of amphibian declines and UV-B has
been controversial with a large literature that I will not delve into here. Relevant
to this chapter, the sites used for the UV-B studies were not chosen with respect
to the potential for exposure to UV-B, but were a convenience sample. The pri-
mary study site turned out to have water much more transparent to UV-B trans-
mission than most amphibian breeding habitats in the Pacific Northwest (Palen
et al. 2002), severely restricting the generality of the proposed hypothesis.
Convenience sampling is almost always done without any intention to prod-
uce a biased result. The water chemistry and UV-B transmission at the study
sites in the Oregon UV-B studies were not known beforehand. These sites were
known to the researchers to have suitable amphibian populations and included
locations where long-term studies had been conducted. Similarly, Corn and
Bury’s (1989) study of the effects of logging on stream amphibians in western
Oregon did not include conscious bias in selection of streams to be sampled,
but it was none-the-less based on a convenience sample. In this study, Bruce
Bury and I identified likely sample locations on topographic maps beforehand,
but the decision to sample was made in the field after inspecting each stream.
We attempted to select typical streams and sample reaches based on our know-
ledge of the characteristics of headwater streams in the region. However, we
did not begin with a well-defined population of streams, and we did not apply
any probabilistic sampling. As stated by Hayek and Buzas (1997, p. 113), “If
the basis for inclusion in a sample is judgment, regardless of how expert, we
will not have a reproducible measure of our field study’s usefulness.” Corn and
Bury (1989) found strong differences in abundance and diversity of amphibians
between streams in logged and unlogged forest. The paper was an early influ-
ence on what has become a spate of research on stream amphibians and forest
management in the Pacific Northwest. Although the original results are largely
supported by subsequent work (Olson et al. 2007), the convenience sampling
design employed limits the scope of the conclusions and Corn and Bury (1989)
should be viewed as hypothesis-generating work rather than a definitive demon-
stration of differences between streams in managed and unmanaged forests.
Cautions about convenience sampling apply equally to the sample frame, the
pool from which the sample is drawn, as to individual study sites. Study areas,
meaning the regions containing the sample frames, are almost never chosen at
random. Study areas may be defined by a relevant management question (e.g.
what is the status of amphibians in a National Park?), or they might be contain
habitats or species of interest, yet be located conveniently near a researcher’s
institution. Probabilistic methods can be used properly to select study sites, but
if the frame is defined by convenience, then inference is restricted to the study
area and the generality of the results may be limited. Two recent studies, con-
ducted less than 200 km apart in Switzerland, illustrate this point. Pellet et al.
(2004) found that urbanization and roads had a strong negative influence on
presence of European treefrogs (Hyla arborea). Conversely, Van Buskirk (2005)
found only weak support for landscape variables (including urbanization) to
explain occurrence and abundance of treefrogs. Both studies were well designed,
and although different methods of analysis were used (logistic regression versus
information theoretic analysis) the different conclusions likely resulted from
intrinsic differences between the study areas.
The North American Amphibian Monitoring Program (NAAMP; Weir and
Mossman 2005), patterned after the North American Breeding Bird Survey
(BBS; Peterjohn et al. 1995), also suffers from a sampling frame defined by con-
venience. NAAMP conducts manual calling surveys of breeding amphibians
on prescribed road routes. Species are identified by their breeding calls, and
data are collected mainly by volunteers. Survey routes are generated through a
random process, but the goal of the program is to monitor trends in amphib-
ian populations throughout the region where routes are conducted (Weir and
Mossman 2005). Reliance on roadside observations limits the inference to
those areas accessed by roads, or requires investigators to make an additional,
untested assumption that the roadside amphibian population experiences the
same trends in abundance as those found in populations away from roads. The
biases that potentially undermine this assumption can be related to both habitat
and the observations themselves (Peterjohn et al. 1995). In Chapter 16 in this
volume Dorcas et al. discuss assumptions about auditory observations. Habitat
assumptions require that roadside wetlands reflect the same conditions found
at wetlands away from roads, and that habitat condition changes over time in
the same direction and at a similar pace alongside and away from roads. These
assumptions are more likely to be violated than satisfied. For example, Keller
and Scallan (1999) found that land cover types were similar near and away from
BBS routes in Maryland and Ohio, but that in Maryland, urbanization was
proceeding at a more rapid pace along roads. If urbanization is associated largely
with existing road networks, then roadside habitats may diverge from distant
habitats more rapidly in Maryland than in Ohio.
A more immediate and concrete difference between wetlands near and dis-
tant from roads (and also a difference between amphibians and birds) is that
roads, especially paved roads with higher volumes of traffic, are a signifi-
cant source of mortality of adult amphibians moving in and out of breeding
ponds. Studies in North America and Europe have found negative relation-
ships between traffic intensity and amphibian mortality and breeding activity
(Fahrig et al. 1995), habitat occupancy (Pellet et al. 2004), and species richness
and abundance (Eigenbrod et al. 2008). Similarly in New York, Karraker et al.
(2008) found a twofold increase in density of egg masses of two amphibian spe-
cies away from roads compared to roadside ponds. This may have been more an
indirect road effect than from direct mortality. Demographic models showed
significant negative effects on these species in roadside ponds resulting from the
application of salt for de-icing (Karraker et al. 2008). Road de-icers epitomize
a confounding variable contributing to differences that would be very difficult
to model in analyses of NAAMP data. Application of de-icers varies among
geographic regions, states, road types, and chemistry from year to year, and the
data quantifying application are likely to be extremely difficult to obtain. The
assumption that roadside habitats are equivalent to habitats away from roads is
not supported by research to date.
The road studies illustrate the point that concerns about convenience sam-
pling also apply to choosing variables for explaining patterns or trend. Data may
be readily available in GIS data layers (e.g. road density), but less available data
(e.g. traffic intensity, de-icer application) may be more important for generating
well-supported models. Anderson (2008) emphasizes that considerable effort
should be devoted to generating hypotheses before data are collected; this obvi-
ously applies to selection of the variables that make up the candidate models.
23.4 Abundance and inference

Issues of inference regarding species typically arise when deciding how to quan-
tify abundance or the related estimate of habitat occupancy. These are import-
ant and somewhat controversial topics, and other chapters deal with estimating
occupancy (Chapter 24) and abundance (Chapter 25) in detail. In this section,
I focus on how uncertainty about numbers affects inference.
Studies often use an index, either simple counts of individuals observed or
counts converted to relative abundance, to represent the abundance of a species
(Hayek and Buzas 1997), and many standard field methods (e.g. in this volume
or Heyer et al. 1994) are designed to obtain counts. However, inference about
trends over time or differences among habitats requires that there is no trend
in the detection probability (the relationship between the index and the true
abundance). Bart et al. (2004) contended that this was a reasonable assumption
for bird studies. Anderson (2001) regarded this as unlikely, and recently put it
even more strongly (Anderson 2008, p. 20): “. . . the evidence is conclusive that
they [index values] represent an amateur, unthinking approach and is not scientific”
and “. . . index values are not data, they are just numbers.” This viewpoint is
not universal. Reliable estimation of detection probability may be difficult in
many situations (Johnson 2008), and variation in detection probability, par-
ticularly among individuals, may result in unreliable estimates of abundance
(Link 2003). Nevertheless, evidence from amphibian studies tends to support
the idea that use of indices should be avoided, because estimates incorporat-
ing detection are more closely correlated to actual abundance than are counts
(Schmidt 2004).
Welsh and Droege (2001) advocated use of count data from surveys of
plethodontid salamanders to monitor forest condition and biological diversity.
Terrestrial salamanders are associated with habitat features that are often dis-
rupted by activities such as logging, and counts of salamanders in several studies
sampled by a variety of techniques show relatively consistent numbers from year
to year (Welsh and Droege 2001). Low inter-annual variation increases ability to
detect trends. However, at any given time, the majority of Plethodon in a popula-
tion are subterranean and unavailable to be captured (Bailey et al. 2004a), and
counts of terrestrial plethodontids vary considerably among years, habitats, and
sampling methods, violating many of the assumptions required for use of indi-
ces in monitoring (Hyde and Simons 2001; Dodd and Dorazio 2004). More
critically, temporary immigration between surface and subterranean habitats
varies among sampling occasions, so that the proportion of the population avail-
able to be sampled is not constant. Additionally, capture–recapture studies on
plethodontids have found low detection probabilities, often less than 0.05 (Jung
et al. 2000; Welsh and Droege 2001; Dodd and Dorazio 2004). Low capture
probabilities are not a problem if they are relatively uniform among habitats
and observers (Welsh and Droege 2001), but there is considerable reason to
doubt that this is true. Detection probabilities of plethodontids varied from
0.06 to 0.41 among years (Dodd and Dorazio 2004), and.between 0.01 and
0.58 among sampling occasions in one model evaluated by Bailey et al. (2004a).
Low detection magnifies any bias due to variation among observers, habitats, or
years. For example, an increase in detection from 0.50 to 0.55 would result in a
10% increase in numbers of animals observed, but an increase from 0.06 to 0.38
would result in 6.5 times as many observations in the second count. This could
produce a result similar to that illustrated in Schmidt (2004, figure 1), where
counts were similar among years, but capture probability varied, with the result
that counts did not reflect a large increase in actual abundance.
Concerns about use of count data apply more broadly than to just terrestrial
salamanders. Johannson et al. (2006) used uncorrected counts of common frog
(Rana temporaria) egg masses as an index to population size, and concluded
that population size declined with increasing latitude and smaller populations
had less genetic variability. Johannson et al. conceded that egg mass counts
likely underestimated true abundance of breeding females but contended it was
an unbiased index, because sampling was the same at all study sites. However,
counts often fail to detect all egg masses present in a pond for a variety of rea-
sons, such as differences in habitat complexity, weather conditions that might
affect visibility, or variation in ability among observers. Grant et al. (2005)
found detection probabilities of ranid egg masses to vary between 0.78 and 1.0.
Variation in detection introduces uncertainty into conclusions about popula-
tion size; this uncertainty is magnified if detection varies in a systematic manner
across a study area.
Count data are incorporated into many indices of species diversity (Hayek
and Buzas 1997), but calculation of these indices for amphibian assemblages
are not appropriate, unless unbiased estimates of abundance are used instead of
the raw counts. It has long been known that number of captures varies among
sampling methods, so that a diversity index that included counts of species
made using different techniques (for example, pitfall traps for one species and
time-constrained searches for another) would not be valid (Corn 1994). Hyde
and Simons (2001) demonstrated sampling efficiency varied among methods,
but also among habitat types for some species of plethodontids when the same
method was employed. Interpretations of diversity indices suffer from the same
problems as interpretations about abundance.
The uncertainties about count data, the expense of obtaining unbiased

estimates of population size, and high amount of year-to-year variation in
abundance of many amphibian species prompted Green (1997) to suggest that
tracking the changes in the occurrence of a species across the landscape—that
is, whether or not a species occupied a given patch of habitat—was a more efficient
way to monitor status and trend of amphibians. Such presence/absence (or,
more accurately, detected/not detected) data used to be considered inferior
to or less scientific than count data (MacKenzie et al. 2003). In part, this is
because surveys to detect presence are typically at a reduced level of effort than
surveys that include counts, and in most situations it is impossible to prove
absence even with great effort. False negatives, or failures to detect species that
are actually present, introduce bias that underestimates occupancy and can
lead to errors in interpretation, such as incorrectly identifying the influence
of habitat variables on occurrence (Mazerolle et al. 2005; Hossack and Corn
2007).
Occupancy estimation (MacKenzie et al. 2006; see also Chapter 24 in this
volume) has been implemented in a variety of studies of amphibians, including
effects of disturbance (e.g. Mazerolle et al. 2005; Hossack and Corn 2007), in
addition to more general monitoring efforts (e.g. Corn et al. 2005; Schmidt
2005). Occupancy analysis has been recommended in the US Geological
Survey’s Amphibian Research and Monitoring Initiative (Muths et al. 2005),
mainly through surveys for tadpoles or egg masses to indicate breeding popula-
tions. Tadpoles typically have high detection probabilities (Brown et al. 2007;
Hossack and Corn 2007), which increases the precision of estimated occupancy.
Occupancy analysis also shows promise as a means for monitoring terrestrial
salamanders. Detection probabilities for occurrence of plethodontids on small
plots were much higher than capture–recapture studies have found for detec-
tion of individual salamanders, and estimated occupancy was much less variable
among years compared to estimated abundance (Bailey et al. 2004b).
Occupancy analysis may not be the best choice in all situations. It includes
assumptions about the relationship between occurrence on the landscape and
abundance of a species (Royle and Nichols 2003), although combining occu-
pancy and count data is topic of active development (Royle et al. 2005; see also
Chapter 24). Occupancy analysis can be difficult to implement in habitats that
are not discrete, for example extensive wetlands. It also may not be possible
to obtain reliable estimates of occupancy for rare species with low detection
probabilities (Bailey et al. 2004b). A reservation about use of occupancy for
detecting trends in salamanders is that local populations must go extinct or be
recolonized for change to be observed. This may be a common occurrence for
pond-breeding amphibians (Green 1997), but it is less likely that populations of

terrestrial salamanders undergo the same dynamics.
One final note of caution: research can be well designed, but the resulting
data and analysis can still be undermined if the investigators do not pay suf-
ficient attention to the biology of the organisms they are studying. Kroll et al.
(2008) recently conducted a study on stream amphibians that employed a rigor-
ous statistically valid design for sample selection, estimated occupancy among
streams for several species, and analyzed effects on detection and occupancy
using information-theoretic methods. One of their findings was that detection
of coastal tailed frogs (Ascaphus truei) declined as the field season progressed.
This was likely a consequence of the timing of sampling, which was conducted
from 19 July to 6 October. Although in parts of the Pacific Northwest, tadpoles
of A. truei require 2 or more years to reach metamorphosis, many of the streams
examined by Kroll et al. are in a region where tadpoles metamorphose in less
than 1 year, beginning as early as late June (Bury and Adams 1999). During
the time when Kroll et al. were sampling, only adult frogs or hatchling tadpoles
(which tend to remain clustered for several weeks under rocks near the nest
site) would have been present in many streams. Both of these life stages are less
observable than older tadpoles. Reduced detection probability results in positive
bias in the occupancy estimate. Studies that employ occupancy analysis should
be designed so that detection does not vary among sample units in a systematic
manner.
23.5 Conclusions
The path to strong inference leads through good study design that incorporates
probabilistic sampling from a well-defined population. Inventory and, especially,
monitoring studies stray from this path when scientific rigor is sacrificed to logis-
tic constraints and convenience in data collection. Tension often exists between
the field biologist and the consulting statistician regarding the requirements of
good study design and the logistical realities of data collection. Having been on
the field biologist’s side of the argument, I can testify that the attitude summa-
rized by “Yes, we realize valid sample selection is important, and it would be nice,
but we have to collect data from the real world”, is fairly common. Constraints in
site selection can be incorporated into study design, such as by stratifying based
on accessibility, and the resulting analysis can test hypotheses about whether
populations that are easily accessible differ from those that are not.
The perils of convenience sampling also apply to choice of life stage to study
or explanatory variables to incorporate in a model. The easiest life stage to study
may not be the same one that is most sensitive to external factors, and variables
should not be included in a model simply because the data are available. There is
no “magic bullet” for sampling amphibians. No single technique encompasses
the variety of life histories of amphibians or the habitats in which they can be
found. Occupancy analysis provides a useful tool for avoiding the pitfalls of
using simple count data or the logistic difficulties of obtaining unbiased esti-
mates of abundance, but it is not a panacea. Ultimately, the design that allows
the strongest inference will be one that avoids convenience sampling and min-
imizes untested assumptions when the data are analyzed.
I thank Larissa Bailey, Blake Hossack, Benedikt Schmidt, and Susan Walls for
suggestions that greatly improved the manuscript.
23.7 References
Anderson, D. R. (2001). The need to get the basics right in wildlife field studies. Wildlife
Anderson, D. R. (2008). Model Based Inference in the Life Sciences: a Primer on Evidence.
Springer, New York.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004a). Estimating detection probabil-
ity parameters for plethodon salamanders using the robust capture–recapture design.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004b). Estimating site occupancy
Bart, J., Droege, S., Geissler, P., Peterjohn, B., and Ralph, C. J. (2004). Density estima-
tion in wildlife surveys. Wildlife Society Bulletin, 32, 1242–7.
Blaustein, A. R. and Kiesecker, J. M. (2002). Complexity in conservation: lessons from
the global decline of amphibian populations. Ecology Letters, 5, 597–608.
Blaustein, A. R., Hoffman, P. D., Hokit, D. G., Kiesecker, J. M., Walls, S. C., and Hays, J. B.
(1994). UV repair and resistance to solar UV-B in amphibian eggs: a link to population
declines? Proceedings of the National Academy of Sciences USA, 91, 1791–5.
Brown, G. W., Scroggie, M. P., Smith, M. J., and Steane, D. (2007). An evaluation of
methods for assessing the population status of the threatened alpine tree frog Litoria
verreauxii alpina in southeastern Australia. Copeia, 2007, 765–70.
Bury, R. B. and Adams, M. J. (1999). Variation in age at metamorphosis across a latitu-
dinal gradient for the tailed frog, Ascaphus truei. Herpetologica, 55, 283–91.
Cochran, W. G. (1977). Sampling Techniques, 3rd edn. John Wiley & Sons, New York.
Corn, P. S. (1994). Straight-line drift fences and pitfall traps. In W. R. Heyer, M. A.

Donnelly, R. W. McDiarmid, L.-A. C. Hayek, and M. S. Foster (eds), Measuring
and Monitoring Biological Diversity: Standard Methods for Amphibians, pp. 109–17.
Corn, P. S. and Bury, R. B. (1989). Logging in western Oregon: responses of headwater
habitats and stream amphibians. Forest Ecology and Management, 29, 39–57.
Corn, P. S., Hossack, B. R., Muths, E., Patla, D. A., Peterson, C. R., and Gallant, A. L.
(2005). Status of amphibians on the Continental Divide: surveys on a transect from
Montana to Colorado, USA. Alytes, 22, 85–94.
Dodd, Jr, C. K. and Dorazio, R. M. (2004). Using counts to simultaneously estimate
abundance and detection probabilities in a salamander community. Herpetologica, 60,
468–78.
Eigenbrod, F., Hecnar, S. J., and Fahrig, L. (2008). The relative effects of road traffic and
forest cover on anuran populations. Biological Conservation, 141, 35–46.
Fahrig, L., Pedlar, J. H., Pope, S. E., Taylor, P. D., and Wegner, J. F. (1995). Effect of road
traffic on amphibian density. Biological Conservation, 73, 177–82.
Grant, E. M., Jung, R. E., Nichols, J. D. and Hines, J. E. (2005). Double-observer
approach to estimating egg mass abundance of pool-breeding amphibians. Wetland
Ecology and Management, 13, 305–20.
Green, D. M. (1997). Perspectives on amphibian population declines: defining the prob-
lem and searching for answers. In D. M. Green (ed.), Amphibians in Decline: Canadian
Studies of a Global Problem. Herpetological Conservation, vol. 1, pp. 291–308. Society
for the Study of Amphibians and Reptiles, St. Louis, MO.
Green, D. M. (2003). The ecology of extinction: population fluctuation and decline in
amphibians. Biological Conservation, 111, 331–43.
Hayek, L.-A. C. (1994). Research design for quantitative amphibian studies. In W. R. Heyer,
M. A. Donnelly, R. W. McDiarmid, L.-A. C. Hayek, and M. S. Foster (eds), Measuring and
Monitoring Biological Diversity: Standard Methods for Amphibians, pp. 21–39. Smithsonian
Hayek, L.-A. C. and Buzas, M. A. (1997). Surveying Natural Populations. Columbia
University Press, New York.
(eds), (1994). Measuring and Monitoring Biological Diversity: Standard Methods for
Hobbs, N. T. and Hilborn, R. (2006). Alternatives to statistical hypothesis testing in
ecology: a guide to self teaching. Ecological Applications, 16, 5–19.
Hossack, B. R. and Corn, P. S. (2007). Responses of pond-breeding amphibians to wildfire:
short-term patterns in occupancy and colonization. Ecological Applications, 17, 1403–10.
Hurlbert, S. N. (1984). Pseudoreplication and the design of ecological field experiments.
Hyde, E. J. and Simons, T. R. (2001). Sampling plethodontid salamanders: sources of
Johannson, M., Primmer, C. R., and Merilä, U. (2006). History vs. current demography:
explaining the genetic population structure of the common frog (Rana temporaria).
Johnson, D. H. (2008). In defense of indices: the case of bird surveys. Journal of Wildlife
Jung, R. E., Droege, S., Sauer, J. R., and Landy, R. B. (2000). Evaluation of terrestrial
and streamside salamander monitoring techniques at Shenandoah National Park.
Environmental Monitoring and Assessment, 63, 65–79.
Karraker, N. E., Gibbs, J. P., and Vonesh, J. R. (2008). Impacts of road deicing salt on the
demography of vernal pool-breeding amphibians. Ecological Applications, 18, 724–34.
Keller, C. M. E. and Scallan, J. T. (1999). Potential roadside biases due to habitat changes
along Breeding Bird Survey routes. The Condor, 101, 50–7.
Kiesecker, J. M., Blaustein, A. R., and Belden, L. K. (2001). Complex causes of amphib-
ian population declines. Nature, 410, 681–4.
Kroll, A. J., Risenhoover, K., McBride T., Beach E., Kernohan B. J., Light, J., and Bach, J.
(2008). Factors influencing stream occupancy and detection probability parameters of
stream-associated amphibians in commercial forests of Oregon and Washington, USA.
Forest Ecology and Management, 255, 3726–35.
Link, W. A. (2003). Nonidentifiability of population size from capture-recapture data
with heterogeneous detection probabilities. Biometrica, 59, 1123–30.
MacKenzie, D. I., Nichols, J. D., Hines, J. E., Knutson, M. G., and Franklin, A. B.
(2003). Estimating site occupancy, colonization, and local extinction when a species is
detected imperfectly. Ecology, 84, 2200–7.
Occurrence. Academic Press, Burlington, MA.
Mazerolle, M. J., Desrochers, A., and Rochefort, L. (2005). Landscape characteris-
tics influence pond occupancy by frogs after accounting for detectability. Ecological
Muths E., Jung, R. E., Bailey, L. L., Adams, M. J., Corn, P. S., Dodd, Jr, C. K., Fellers, G. M.,
Sadinski, W. J., Schwalbe, C. R., Walls, S. C. et al. (2005). The U.S. Department
of Interior’s Amphibian Research and Monitoring Initiative: a successful start to a
national program. Applied Herpetology, 2, 355–71.
Olson, D. H., Anderson, P. D., Frissell, C. A., Welsh, Jr, H. H., and Bradford, D. F.
(2007). Biodiversity management approaches for stream–riparian areas: perspectives
for Pacific Northwest headwater forests, microclimates, and amphibians. Forest Ecology
and Management, 246, 81–107.
Palen, W. J., Schindler, D. E., Adams, M. J., Pearl, C. A., Bury, R. B., and Diamond, S. A.
(2002). Optical characteristics of natural waters protect amphibians from UV-B in the
U. S. Pacific Northwest. Ecology, 83, 2951–7.
Pellet, J., Guisan, A., and Perrin, N. (2004). A concentric analysis of the impact of urban-
ization on the threatened European tree frog in an agricultural landscape. Conservation
Biology, 18, 1599–1606.
Peterjohn, B. G., Sauer, J. R., and Robbins, C. S. (1995). Population trends from the North
American Breeding Bird Survey. In T. E. Martin and D. M. Finch (eds), Ecology and
Management of Neotropical Migrants, pp. 3–39. Oxford University Press, New York.
Platt, J. R. (1964). Strong inference. Science, 146, 347–53.
Pounds, J. A. (2001). Climate and amphibian declines. Nature, 410, 639–40.
Royle, J. A. and Nichols, J. D. (2003). Estimating abundance from repeated presence–

absence data or point counts. Ecology, 84, 777–90.
Royle, J. A., Nichols, J. D., and Kéry, M. (2005). Modelling occurrence and abundance
of species when detection is imperfect. Oikos, 110, 353–9.
Schmidt, B. R. (2004). Declining amphibian populations: the pitfalls of count data in the
study of diversity, distributions, dynamics, and demography. Herpetological Journal,
14, 167–74.
Schmidt, B. R. (2005). Monitoring the distribution of pond-breeding amphibians
when species are detected imperfectly. Aquatic Conservation: Marine and Freshwater
Ecosystems, 15, 681–92.
Thompson, S. K. (1992). Sampling. John Wiley & Sons, New York.
Underwood, A. J. (1998). Design, implementation, and analysis of ecological and envir-
onmental experiments. In W. J. Resetarits, Jr. and J. Bernardo (eds), Experimental
Ecology: Issues and Perspectives, pp. 325–49. Oxford University Press, New York.
Van Buskirk, J. (2005). Local and landscape influence on amphibian occurrence and
abundance. Ecology, 86, 1936–47.
Weir, L. A. and Mossman, M. J. (2005). North American Amphibian Monitoring
Program (NAAMP). In M. Lannoo (ed.), Amphibian Declines: The Conservation Status
of United States Species, pp. 307–13. University of California Press, Berkeley, CA.
Welsh, Jr, H. H. and Droege, S. (2001). A case for using plethodontid salamanders
for monitoring biodiversity and ecosystem integrity of North American forests.
Werner, E. E. (1998). Ecological experiments and a research program in community
ecology. In W. J. Resetarits, Jr and J. Bernardo (eds), Experimental Ecology: Issues and
Perspectives, pp. 3–26. Oxford University Press, New York.
Williams, B. K., Nichols, J. D., and Conroy M. J. (2002). Analysis and Management of
Animal Populations. Academic Press, San Diego, CA.
24
Capture–mark–recapture, removal
sampling, and occupancy models
Larissa L. Bailey and James D. Nichols
24.1 Introduction
Understanding the distribution and abundance of organisms is frequently stated
as the objective of ecological investigations (Elton 1927; Krebs 1972). Similarly,
distribution and abundance are primary criteria used to classify the status of
species (e.g. threatened, endangered) for conservation purposes (Gardenfors
et al. 2001). Thus, both scientific and conservation perspectives lead us to select
abundance and occupancy as reasonable state variables for investigation. We
define abundance as the number of organisms present in some area of interest
and occupancy as the proportion of sample units or, more generally, of area, that
is occupied by a species of interest.
Estimation of these quantities is typically based on count statistics of some
sort. For example, abundance estimation may be based on the number of animals
caught in traps, detected by visual encounter surveys, or perhaps by auditory
means. Estimation of occupancy is typically based on detection/non-detection
data (frequently referred to as presence/absence data) representing counts of
sample units at which the focal species is (or is not) detected. Use of count sta-
tistics to draw inferences about such quantities as abundance and occupancy
requires attention to two important sources of variation in counts (Pollock et al.
2002; Williams et al. 2002; Lancia et al. 2005).
The first important source of variation is spatial, with counts, abundance,
and occupancy of organisms varying over space. It is frequently not possible to
count organisms over the entire area about which inference is sought, requiring
selection of a subset of geographic sample units for survey. Selection of these
units must be done in such a way that counts on these units permit inference
about the units not selected, even in the face of geographic variation. Many
such sampling designs have been developed (e.g. simple random sampling,
stratified random sampling, adaptive cluster sampling), and we refer the reader
to sources that describe these designs in some detail (Cochran 1977; Williams
et al. 2002).
The second important source of variation concerns the recognition that
counts, whether of organisms or occupied sample units, are nearly always incom-
plete, in that some organisms and species go undetected no matter how intensive
the sampling effort. Detection probability associated with the count must be
estimated to provide information needed to translate the count(s) into an infer-
ence about abundance or occupancy. Various means of estimating detection
probabilities have been developed for both counts of organisms (Seber 1982;
Williams et al. 2002; Lancia et al. 2005) and counts of sample units occupied by
a focal species (MacKenzie et al. 2002, 2006; Royle and Dorazio 2008).
Both scientific and conservation interests frequently focus not only on the
values of relevant state variables, but also on the estimation of the vital rates
responsible for the dynamics of those state variables. Rates of survival, reproduc-
tion, and movement in and out of the population are responsible for all changes
in abundance of organisms and are thus important topics of study. Similarly,
rates of local extinction and colonization determine the dynamics of species
occurrence across units of the landscape. Ecological investigations of popula-
tion and occupancy dynamics frequently focus on the environmental and eco-
logical variables that induce changes in the relevant vital rates. Conservation
requires actions that influence vital rates in the desired manner. As with esti-
mation of abundance, inference about vital rates requires models that explicitly
incorporate detection probabilities that reflect the incompleteness of virtually
all wildlife sampling.
In this chapter we describe methods that we believe should be useful for
estimating abundance, occupancy, and their respective vital rates in studies of
amphibian populations. Our intention is to provide the reader with information
about how these methods work, as well as an entrée to the scientific literature
that describes these methods in detail.
24.2 Estimating amphibian population size and

vital rates
24.2.1 Marking: tag type and subsequent encounter
Robust estimation of amphibian abundance and the vital rates that cause changes
in abundance (e.g. survival and movement) usually require capturing and marking
24 Abundance estimation and occupancy models | 449
individual animals from the population(s) of interest. Numerous methods have

been used to ‘mark’ amphibians and most require physical handling of indi-
viduals during an initial capture and marking occasion (see Chapters 8 and 11
in this volume for some methods). ‘Marking’ can involve applying a mark, tag,
or radiotransmitter, or simply photographing or otherwise recording individu-
als with unique natural patterns, or obtaining unique identification via genetic
material. Collectively, we group these marking methods into two sets: those
permitting unique, individual identification on subsequent encounters and
those where the individual can only be identified as marked or unmarked (so-
called batch marks). Batch marks are relatively easy to apply and may identify
the time, location, or life-history stage when the individual was first caught, but
batch marks are generally not as useful for estimating survival and movement
probabilities. In the next section we will discuss capture–mark–recapture mod-
els with the understanding that marked individuals can be uniquely identified.
We will discuss the use of batch-marked animals for abundance estimation in
the subsequent section on removal sampling (section 24.2.3).
24.2.2 Capture–mark–recapture
Capture–mark–recapture studies typically involve one to a few study areas or
populations. The duration of the sampling period, the intensity of sampling
within the period, and the length of time between successive sampling periods
can vary substantially depending on study objectives. These features (objectives
and sampling design) are critical components that determine which capture–
mark–recapture methods are most appropriate for a given study. Capturing,
marking, and recapturing individuals within a single day have been used to esti-
mate amphibian abundance at different life-history stages: egg masses (Grant
et al. 2005) and tadpoles (Jung et al. 2002). Alternatively, many studies of
amphibian populations involve capturing individuals during the species’ breed-
ing season over multiple years, where each sampling period may be 2–12 weeks
(Schmidt and Anholt 1999; Fretey et al. 2004; Church et al. 2007).
Data resulting from capture–mark–recapture studies are summarized into
individual capture histories, which are simply rows of ones and zeros indicating
whether an individual was captured (1) or not (0) during each sampling occasion.
For example, a study of terrestrial or stream-side salamanders might sample an area
for four consecutive days yielding individuals with the following capture history:
0 1 1 0. These individuals were first captured and marked on day 2, recaptured on
the third day, but not encountered on the fourth and final day of sampling. If we
assume that the population exposed to sampling does not change over the 4-day
period, then these data can be modeled using closed-population models (closed
population implies no births, deaths, immigration, or emigration; Williams et al.

2002). Here, we define pt as the capture probability for an unmarked individual in
the population exposed to sampling on day t, and ct as the probability of recapture
for previously marked animals. Assuming the capture events are independent, we
would model the above capture history, 0 1 1 0, as:
Pr(0 1 1 0)(1 p1 ) p2c3 (1c 4 )
Since all individuals in the exposed population are assumed to be present at the
initial sampling occasion, (1 p1)p2 denotes the probability a salamander was
missed on the first day and captured for the first time on day 2. The probability
of encountering an individual may change after first capture: thus the remain-
der of the expression indicates the probabilities associated with the events of
recapture on day 3, c3, but not on day 4, (1 c4).
We would compile similar probability expressions for each captured sala-
mander. The product of these probabilities over all captured salamanders would
constitute a model (i.e. the likelihood) for the entire data set, and estimates of
the capture and recapture probabilities could be obtained. In this model, the
likelihood is conditioned on the total number of individuals captured during
the entire study, MK 1, where K indexes the final sample occasion. Abundance
N can be estimated as a derived parameter, specifically as:
M K 1
Nˆ
⎡⎣1(1 pˆ1 )(1 pˆ2 ) (1 pˆ3 ) (1 pˆ4 )⎤⎦ (24.1)
Abundance is simply estimated as the total number of animals caught, divided

by the probability that a member of the population is caught at least once dur-
ing the study. This latter probability (the denominator of eqn 24.1) is obtained
as the complement of the probability that the animal is not captured at any
of the four sample occasions. Modern closed-population models include con-
ditional likelihood models (Huggins 1989, 1991) and full likelihood models
where the abundance parameter, N, remains in the likelihood (Otis et al. 1978;
see Williams et al. 2002 and Chao and Huggins 2005 for detailed reviews).
Parameter estimates can be obtained for various models using common, free
software packages such as programs MARK (White and Burnham 1999) and
CAPTURE (Rexstad and Burnham 1991).
In most studies related to amphibian ecology and conservation, the point esti-
mates of abundance and various capture probabilities are not of primary interest.
Rather, biologists are more interested in the relationships between these param-
eters and relevant environmental, habitat, or biological variables (i.e. covariates).
Usually, biologists have competing hypotheses about the influence of specific

covariates on model parameters, and these hypotheses are translated into models
that are then fit to data using covariate modeling. For example, suppose capture
and recapture probabilities were thought to be equivalent but to vary with tem-
perature on the day of the sampling occasion, and both capture probabilities and
local abundances were thought to differ between two groups of individuals (say
males and females). Capture probability can be modeled as a linear-logistic func-
tion of both temperature and sex:
b b x b x
e 0 1t 2 g
ptg ctg b b x b x (24.2)
1e 0 1 t 2 g
where xt is the temperature at sampling occasion t and xg is an indicator variable

denoting whether the individual is male (xg 0) or female (xg 1). This additive
model contains three parameters—b0, b1 and b2—with b1 and b2 defining the
relationships between capture probability and temperature and sex, respectively.
Abundance estimates for males and females can be calculated using eqn 24.1,
where both the numerator and denominator are gender-specific. Inference about
the relationship between capture probability and both temperature and sex can
be based on confidence intervals of the estimates of b1 and b2, and on compari-
sons of the model incorporating the relationship of eqn 24.2 with other models
that do not include these relationships. Such model comparisons could be based
on model selection statistics (Burnham and Anderson 2002) or on likelihood
ratio tests in the case of nested models (e.g. Williams et al. 2002).
Program MARK has numerous classes of closed-population models that
include conditional and full likelihood models, as well as models that allow for
individual heterogeneity in capture probabilities that cannot be modeled via cov-
ariates (Norris and Pollock 1996; Pledger 2000). Discrimination among com-
peting biological hypotheses is accomplished by first fitting their corresponding
models to a given set of capture histories and then computing model selection
statistics (Burnham and Anderson 2002). This approach to model selection
presupposes that at least one model in the model set fits the data adequately.
Goodness-of-fit tests have been developed for closed-population models (Otis
et al. 1978), but they do not always perform well. Thus, practitioners are encour-
aged to use model-averagee estimates of abundance based on supported models
in the candidate set (Stanley and Burnham 1998).
The application of likelihood-based closed-population capture–mark–
recapture models to amphibian populations is surprisingly rare (but see Bailey
et al. 2004b; Jung et al. 2002, 2005). Most inference about factors contributing
to variation in amphibian abundance is still based on relative abundance indices

(counts). The lack of use of closed-population models may relate to suspected
violations in the closure assumption and the belief that the population of amphib-
ians available for capture is likely to change between sampling occasions. Tests of
the closure assumption have been developed by Otis et al. (1978) and Stanley and
Burnham (1999). This topic is discussed in detail in Chapter 25, and practitioners
should consider carefully both the target population and associated components
of capture probability when designing abundance-related studies.
Often, it is the processes that cause change in a population that may be the
primary interest for amphibian research and conservation. Investigation of these
processes requires the use of capture–mark–recapture models that are “open”
to changes in population size and composition between sampling periods.
There are four fundamental demographic parameters responsible for changes
in amphibian abundance: reproduction, mortality, emigration, and immigra-
tion. The estimation of these parameters and their relationship to environmen-
tal and habitat variables is key to understanding the dynamics of amphibian
populations and predicting possible responses to management or conservation
actions. Methods for estimating amphibian reproduction are discussed in previous
chapters in this volume (Chapters 4 and 9), and here we focus primarily on esti-
mating survival and movement probabilities.
Consider our previous four-occasion study, but let’s assume that the four occa-
sions represent successive years. Now our observed capture history, 0 1 1 0, is
interpreted as an individual that was first captured, marked, and released in year
2, recaptured in year 3, and not seen in year 4. The first capture–mark–recapture
models for open populations were conditional on first release (year 2 in this case),
in the sense that the initial capture is used as a starting point and is not itself
modeled. The classic Cormack–Jolly–Seber (CJS) model (Cormack 1964; Jolly
1965; Seber 1965) contains two types of parameter: pt is the probability that an
amphibian present in the study area is captured during occasion t, and t is the
probability that an amphibian alive and present in the study area during occasion
t is still alive and in the study area at occasion t 1 (apparent survival probabil-
ity). Thus, we would write a model for the observed capture history, 0 1 1 0, as:
Pr (0 11 0 | released in year 2 )2 p3 ⎡⎣3 (1 p4 )(13 )⎤⎦ 2 p3 (13 p4 )
The first two terms, 2p3, represent the probability the amphibian survives
between periods 2 and 3 and is recaptured in year 3 (corresponding to the events 1
1 in the capture history). The remainder of the expression [3 (1 p4) (1 3)]
represents two possibilities: (1) the amphibian survived and remained in the study
area from year 3 to 4, but was not captured in year 4, or (2) the amphibian died
or permanently emigrated from the study area between year 3 and 4. Probability
statements such as this are compiled for each capture history (i.e. each individual)
and combined into a likelihood function. Estimates can be obtained via max-
imum likelihood using computer software such as programs MARK (White and
Burnham 1999) or M-SURGE (Choquet et al. 2005).
Many extensions are available for this sort of modeling. Parameters can be
constrained to be constant over time, they can be modeled as functions of group
(e.g. males and females) membership, and they can be modeled as functions of
time-specific or individual-specific covariates (Lebreton et al. 1992). So-called
unconditional modeling also permits the estimation of time-specific abun-
dance, recruitment, and even population growth rate (see Williams et al. 2002).
Program MARK (White and Burnham 1999) provides user-friendly software
for such modeling.
These models for open populations are capable of dealing with population
gains and losses between sampling periods, but they do not provide a means of
distinguishing movement from other processes. Indeed, under the CJS model,
the processes of emigration and immigration are confounded with survival
and capture probabilities. For example, if individuals permanently emigrate
from the study area, then the complement of ‘apparent survival’ probability
will include both mortality and permanent emigration (e.g. Burnham 1993). In
the case of random temporary emigration, CJS estimates of capture probability
represent the product of the probability that the individual is not a temporary
emigrant and the probability that the individual is caught, conditional on pres-
ence (Kendall et al. 1997; see also Chapter 25 in this volume). Temporary emi-
gration is particularly important in amphibian ecology, as individuals are often
unavailable for capture during specific life-history phases (i.e. post-metamorphic
juveniles or adults that skip breeding opportunities). Movement processes can
be separately estimated in some cases using multistate models (see below) or
models that incorporate extra sources of information (Burnham 1993; Kendall
et al. 1997; Fujiwara and Caswell 2002; Kendall and Nichols 2002; Williams
et al. 2002; Bailey et al. 2004a; Schaub et al. 2004).
Multistate models are open-population capture–mark–recapture models that
can be used when animals are categorized by some state variable that can be
assessed each time an animal is encountered (Arnason 1972, 1973; Brownie et al.
1993; Schwarz et al. 1993). When location or study site is the state, direct infer-
ences about movement among sampled locations are possible (e.g. Hestbeck et al.
1991; Brownie et al. 1993). In addition to location, state variables can include
characteristics of the captured animals themselves. For example, size class is a
reasonable state variable in some amphibian studies (Wood et al. 1998).
Multistate models can also be used when one or more of the states may be
unobservable. Such uses focus on situations in which animals move back and
forth between study sites, where capture and resighting efforts are made, and
other locations at which no sampling effort is expended. For example, when
amphibians are sampled at breeding sites, non-breeding animals are tempor-
ary emigrants and thus unobservable. In the simple two-state case, the model
contains parameters that are state-specific, r, where r O denotes the observ-
able or breeding state, and r U denotes the unobservable or non-breeding
o
state. Parameters include: pt , the aforementioned capture probability which is
non-zero for breeders only ( ptU 0, by definition); Str , the probability that an
individual in state r at time t survives until time t 1; and ct , the probability
rs
that an individual that is in state r at time t, and survives until time t 1, is in

state s at time t 1. Using this parameterization we would write a model for the
observed capture history, 0 1 1 0 1, as:
Pr (0 11 0 1 | released in year 2 )S2O c2OO p3O S3O ⎡⎣(c3OO (1 p4O )S4O c4OO c3OU S4U cUO
4 )⎤
O
⎦ p5
The fact that the individual was not captured in year 4 could result from one of
two processes: either the amphibian was in the study area but not captured, or
the amphibian was not in the study (or breeding) area in year 4, but returned to
the area in year 5. Note that the above parameterization assumes that survival
between sampling periods t and t 1 depends on state at occasion t, e.g. Str .
For situations in which this assumption is not appropriate, it is possible to com-
bine the survival and transition parameters as ft rs Str ctrs and to estimate this
survival-transition parameter ( ft rs ) directly.
Parameters of these multistate models can be estimated using maximum
likelihood methods and modeled as a function of relevant covariates. Useful
recent software includes programs MARK (White and Burnham 1999) and
M-SURGE (Choquet et al. 2005). Practitioners using these models should ver-
ify that all parameters are uniquely identifiable as there are some cases where
multiple parameter values will yield the same maximized likelihood values
(Kendall and Nichols 2002; Gimenez et al. 2004; Schaub et al. 2004; Bailey
et al. 2009; Hunter and Caswell 2009). Simplifying assumptions, such as
equating parameters over time or state, may be necessary (Kendall and Nichols
2002; Schaub et al. 2004; Bailey et al. 2009; Hunter and Caswell 2009), but the
models offer tremendous flexibility and the ability to estimate important demo-
graphic parameters that have long eluded researchers (Biek et al. 2002).
The original robust design (Pollock 1982; Kendall et al. 1997) combines open
and closed models and involves sampling at two temporal scales: each primary
period (e.g. year) consists of two or more secondary occasions that are closely
spaced in time, such that the population may be considered demographically
and geographically closed. For example, during the breeding season individuals
at a pool may be sampled multiple times over the course of a few days. Each inde-
pendent sample of the breeding population would be considered a secondary
occasion, and this process would be repeated for several years (primary periods),
over which time survival, breeding and/or movement probabilities could be esti-
mated (see Kendall 2004 for review of robust design models). The robust design
typically yields improved precision of vital rate parameter estimates, but can
also be used to estimate period- (year-) specific abundance, while accounting
for modeled or unmodeled heterogeneity in capture probabilities. The robust
design also permits direct estimation of probabilities of temporary emigration
(Kendall et al. 1997) and of the separate contributions of recruitment from
immigration and in situ reproduction (Nichols and Pollock 1990; Nichols et al.
2000; see Schmidt et al. 2005 for an amphibian example).
24.2.3 Removal sampling

Removal sampling is simply a special case of a closed-population abundance
model where recapture probability is zero (i.e. ct ≠ pt where ct 0; Otis et al.
1978; White et al. 1982). On each sampling occasion, captured individuals are
removed from the population in some manner. Removal can be accomplished by
either physically removing individuals from the site (e.g. temporarily retaining
individuals until sampling is completed) or by batch marking individuals. The
data then consist of the number of new individuals captured on each sampling
occasion. As with closed models, one can directly model the effect of sampling
effort (e.g. hours of search) or other time-specific covariates on the probability
of initial capture. For applications of removal models to estimate amphibian
abundances, see Bruce (1995), Jung et al. (2002), and Bailey et al. (2004b).
24.3 Estimating amphibian occupancy and vital rates

Macroecological studies focus on the distribution of amphibians across the land-
scape. Even in large-scale studies for which abundance is of interest, abundance
estimation at many study locations may be prohibitively expensive. In these and
other cases, biological hypotheses and conservation objectives may shift from
individual populations to the number or proportion of patches occupied by a
target amphibian species. In these cases it may be reasonable to use occupancy,

or the proportion of sites that are occupied, as a state variable.
Quantitative occupancy modeling has exhibited rapid development recently,
with much of this work motivated by amphibian systems (Muths et al. 2005).
These methods can be used for single or multiple species over one to many sea-
sons. All methods acknowledge the likely scenario that species are not always
detected when present and emphasize that reliable inference can still be made
from detection/non-detection information if detection and occupancy prob-
abilities are simultaneously estimated. MacKenzie and his colleagues offer an
initial review of these methods and apply them to several amphibian problems
(MacKenzie et al. 2006). Here we consider methods that estimate occupancy
and associated vital rates for a single species only.
24.3.1 Occupancy estimation

Occupancy studies typically involve large areas containing numerous sampling
units, or sites. These sites may be naturally occurring patches of habitat (i.e.
wetlands or stream reaches) or independent subunits, such as plots or transects,
within different habitat types. A subset of sites is chosen using a probabilistic
sampling method (e.g. simple or stratified random sample), and sites are sam-
pled repeatedly over a period of time during which there is no change in the
occupancy status at the sites (i.e. sites are either occupied or unoccupied by the
target species during the sampling period). The possible duration of the sample
period will depend on the species life history and behavioral characteristics.
During each site visit the target species is recorded as detected (1) or not (0),
creating a detection history for each site reminiscent of the individual capture
histories in capture–mark–recapture methods. The main difference between
capture–mark–recapture and occupancy scenarios is that sites can have detec-
tion histories consisting only of zeros. Such histories never occur in capture–
mark–recapture data, because the data are conditional on individuals that are
captured/detected at least once. As with abundance estimation, a probabilistic
model is used to describe the sequence of events that produced each detection
history.
MacKenzie et al. (2002, 2006) defined two types of parameter for occupancy
models: ci represents the probability that site i is occupied by the target species,
and pij is the probability of detecting the species at site i during the jth inde-
pendent visit to the site. Dropping the i notation for each site, we would write a
model for the observed detection history, 0 1 1 0 as:
Pr (0 1 1 0)c(1 p1 ) p2 p3 (1 p4 )
The site is clearly occupied by the target species, detected during the second and
third visit, but not detected during the first and fourth visit. A detection history
consisting of all zeros, 0 0 0 0, would have two possible explanations: either the
site was occupied, but the species was not detected during any visit, or the site
was unoccupied. Written as a mathematical expression:
Pr (0 0 0 0)c(1 p1 )(1 p2 )(1 p3 )(1 p4 )(1c )
Again, the detection data and the corresponding probability model are com-
bined to form a likelihood function, and estimates can be obtained via software
such as program MARK (White and Burnham 1999) or program PRESENCE
(MacKenzie et al. 2006; www.mbr-pwrc.usgs.gov/software/).
An alternative approach to estimation and modeling is to view the above
models hierarchically. One model component concerns the true spatial process
of species occurrence across the sampled sites. Conditional on this true spatial
process, a sampling component models the detection process. These hierarch-
ical components (process and sampling) are combined under a Bayesian frame-
work, and statistical inference is achieved using Markov chain Monte Carlo
(MCMC) methods (Royle and Dorazio 2008). Using either procedure, it is
possible to model occupancy and detection probability as functions of measured
covariates. Models incorporating different combinations of covariates represent
competing hypotheses about factors believed to influence amphibian distribu-
tion or probability of detection. These occupancy models have gained popular-
ity in amphibian studies around the globe (e.g. Bailey et al. 2004c; Royle 2004;
Mazerolle et al. 2005; Muths et al. 2005; Pellet and Schmidt 2005; Royle and
Link 2005) and represent an important improvement over logistic regression
models of amphibian–habitat relationships that ignore imperfect detection (Gu
and Swihart 2004; MacKenzie et al. 2006).
24.3.2 Estimation of occupancy vital rates:

extinction and colonization
Amphibian systems are often viewed as collections of sites that are sometimes
occupied by the target species, and sometimes not, depending on the dynamic
processes of extinction and colonization. These vital rates are the processes by
which occupancy status changes among sites throughout the landscape and can
be the primary parameters of interest for many amphibian conservation and
management programs (Muths et al. 2005). MacKenzie et al. (2003) extended
the so-called single-season occupancy models discussed in the previous section
to include these two dynamic parameters: εt is the probability that an occupied
site in season t becomes unoccupied in season t 1 (local extinction) and t is

the probability that an unoccupied site in season t is occupied by the target spe-
cies in season t 1 (colonization). These multi-season models still assume that
all (or a subset of) sites are visited multiple times within a season, over a period
where the occupancy state at each site is static (notice this design resembles
the robust design in capture–mark–recapture studies). Probability models and
likelihoods are developed in the usual fashion, and inference can be based on
either maximum likelihood or MCMC implementation of hierarchical models.
Covariates can be modeled and constraints can be imposed to address interest-
ing biological hypotheses about variables believed to influence extinction and
colonization probabilities (MacKenzie et al. 2003, 2006).
24.4 Summary and general recommendations

Understanding factors that influence amphibian abundances and distribu-
tions are fundamental to amphibian ecology and conservation. The estima-
tion methods we discussed here should be useful to both scientific studies and
conservation efforts, as they have the flexibility to accommodate both spatial
and temporal variation while accounting for imperfect detection of individuals
or species of interest. Scientific studies of amphibian population or occupancy
dynamics typically involve efforts to discriminate among competing biological
hypotheses about the processes that underlie such dynamics. These hypotheses
can be translated into models that include both the ecological processes of inter-
est and the sampling processes that give rise to the collected data. Conditional
on selection of appropriate study design and field methods, the competing mod-
els can be fit to the resulting data and evaluated (e.g. via model selection criteria;
Burnham and Anderson 2002).
Conservation decisions are also based on hypotheses, and their correspond-
ing models, about how studied populations respond to potential management
actions. The same process as described above for scientific studies is used to
determine the degrees of faith that should be placed in the different models
of population response to conservation or management action(s). In addition,
conservation efforts require estimates of state variables (e.g. abundance or occu-
pancy) for the purposes of making state-dependent decisions and assessing the
degree to which conservation objectives are being met. The inference methods
described in this chapter should be useful in the conduct of science and conser-
vation for amphibian populations.
Sampling design should be carefully considered when designing capture–mark–
recapture or occupancy studies. Pilot data or guesses about capture/detection
probabilities can be used with recently-developed simulation- or approximation-

based software to compare estimator precision for different sample sizes and sam-
pling scenarios (see Devineau et al. 2006 for capture–mark–recapture studies and
Bailey et al. 2007 for occupancy studies). Common capture–mark–recapture
recommendations include trying to achieve high ( 0.20) capture probabilities,
minimizing variation in p among individuals, and collecting important covari-
ates likely to influence capture probabilities. These recommendations are echoed
for the species detection probabilities used to estimate occupancy and associated
dynamic parameters (see MacKenzie et al. 2002, 2006).
Ultimately, sampling design, including data collection and selection of infer-
ence methods, should be inherited from the nature and objectives of the lar-
ger scientific or conservation program. Conditional on the larger program, the
methods discussed in this chapter can be tailored to meet program needs. The
evolution of capture–mark–recapture and occupancy inference methods has
been characterized by substantive interactions between statisticians and both
ecologists and conservation biologists, with model development being moti-
vated almost exclusively by the needs of ecologists. Many of the recent advances
in occupancy methods were motivated by collaborations between amphibian
researchers and statisticians (Muths et al. 2005; MacKenzie et al. 2006; Royle
and Dorazio 2008), and we expect such interactions to continue and to lead to
development of even more useful inference methods in the future.
24.5 Disclaimer
Any use of trade, product, or firm names is for descriptive purposes only and
does not imply endorsement by the US Government.
24.6 References
Arnason, A. N. (1972). Parameter estimates from mark-recapture experiments on two pop-
ulations subject to migration and death. Researches on Population Ecology, 13, 97–113.
Arnason, A. N. (1973). The estimation of population size, migration rates, and survival
in a stratified population. Researches on Population Ecology, 15, 1–8.
Bailey, L. L., Kendall, W. L., Church, D. R., and Wilbur, H. M. (2004a). Estimating
survival and breeding probability for pond-breeding amphibians: A modified robust
design. Ecology, 84, 2456–66.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004b). Comparing population size
estimators for plethodontid salamanders. Journal of Herpetology, 38, 370–80.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004c). Estimating site occupancy
Bailey, L. L., Hines, J. E., Nichols, J. D., and MacKenzie, D. I. (2007). Sampling design
trade-offs in occupancy studies with imperfect detection: examples and software.
Ecological Applications 17, 281–90.
Bailey, L. L., Kendall, W. L., and Church, D. R. (2009). Exploring extensions to multi-
state models with multiple unobservable states. In D. L. Thomson, E. G. Cooch, and
M. C. Conroy (eds), Modeling Demographic Processes in Marked Populations, pp. 693–
710. Springer Science+Business Media, New York.
Biek, R., Funk, W. C., Maxell, B. A., and Mills, L. S. (2002). What is missing in amphib-
ian decline research: Insights from ecological sensitivity analysis. Conservation Biology,
16, 728–34.
Brownie, C., Hines, J. E., Nichols, J D., Pollock, K. H., and Hestbeck, J. B. (1993).
Capture-recapture studies for multiple strata including non-Markovian transitions.
Biometrics, 49, 1173–87.
Bruce, R. C. (1995). The use of temporary removal sampling in a study of population
dynamics of the salamander Desmognathus monticola. Australian Journal of Ecology,
20, 403–12.
Burnham, K. P. (1993). A theory for combined analysis of ring recovery and recapture
data. In J. D. Lebreton and P. M. North (eds), Marked Individuals in the Study of Bird
Populations, pp. 199–213. Birkhauser-Verlag, Basel.
Burnham, K. P. and Anderson, D. R. (2002). Model Selection and Multimodel Inference.
Springer-Verlag, New York.
Chao, A. and Huggins, R. M. (2005). Modern closed-population capture-recapture
models. In S. C. Amstrup, T. L. McDonald, and B. F. J. Manly (eds), Handbook of
Capture-Recapture Analysis, pp. 58–87. Princeton University Press, Princeton, NJ.
Choquet, R., Reboulet, A. M., Pradel R., Gimenez, O., and Lebreton, J. D. 2005.
M-SURGE 1–8 User’s Manual. CEFE, Montepellier. ftp.cefe.cnrs.fr/biom/soft-cr/.
Church, D. R., Bailey, L. L., Wilbur, H. M., Kendall, W. L., and Hines, J. E. (2007).
Iteroparity in the variable environment of the salamander Ambystoma tigrinum. Ecology,
88, 891–903.
Cochran, W. G. (1977). Sampling Techniques. Wiley, New York.
Cormack, R. M. (1964). Estimates of survival from the sighting of marked animals.
Biometrika, 51, 429–38.
Devineau, O., Choquet, R., and Lebreton, J. D. (2006). Planning capture-recapture stud-
ies: straightforward precision, bias, and power calculations. Wildlife Society Bulletin,
34, 1028–35.
Elton, C. (1927). Animal Ecology. Sidgwick & Jackson, London.
Fretey, T., Cam, E., Le Garff, B., and Monnat, J. Y. (2004). Adult survival and temporary
emigration in the common toad. Canadian Journal of Zoology, 82, 859–72.
Fujiwara, M. and Caswell, H. (2002). A general approach to temporary emigration in
mark-recapture analysis. Ecology, 83, 3266–75.
Gardenfors, U., Hilton-Taylor, C., Mace, G. M., and Rodriguez, J. P. (2001). The appli-
cation of IUCN red list criteria at regional levels. Conservation Biology, 15, 1206–12.
Gimenez, O., Viallefont, A., Catchpole, E. A., Choquet, R., and Morgan, B. J. T.
(2004). Methods for investigating parameter redundancy. Animal Biodiversity and
Grant, E. H. C., Jung, R. E., Nichols, J. D., and Hines, J. E. (2005). Double-observer
approach to estimating egg mass abundance of pool-breeding amphibians. Wetlands
Gu, W. and Swihart, R. K. (2004). Absent or undetected? Effects of non-detection of spe-
cies occurrence on wildlife-habitat models. Biological Conservation, 116, 195–203.
Hestbeck, J. B., Nichols, J. D., and Malecki, R A. (1991). Estimates of movement and site
fidelity using mark resight data of wintering Canada geese. Ecology, 72, 523–33.
Huggins, R. M. (1989). On the statistical analysis of capture experiments. Biometrika,
76, 133–40.
Huggins, R. M. (1991). Some practical aspects of conditional likelihood approach to cap-
ture experiments. Biometrics, 47, 725–32.
Hunter, C. M. and Caswell, H. (2009). Rank and redundancy of multistate mark-
recapture models for seabird populations with unobservable states. In D. L. Thomson,
M. C. Conroy, and E. G. Cooch (eds), Modeling Demographic Processes in Marked
Populations, pp. 799–828. Springer Science+Business Media, New York.
Jolly, G. M. (1965). Explicit estimates from capture-recapture data with both death and
immigration-Stochastic model. Biometrika, 52, 225–47.
Jung, R. E., Dayton, G. H., Williamson, S. J., Sauer, J. R., and Droege, S. (2002). An
Jung, R. E., Royle, J. A., Sauer, J. R., Addison, C., Rau, R. D., Shirk, J. L., and Whissel, J. C.
(2005). Estimation of stream salamander (Plethodontidae, Desmognathinae and
Plethodontinae) populations in Shenandoah National Park, Virginia, USA. Alytes, 22,
72–84.
Kendall, W. L. (2004). Coping with unobservable and mis-classification state in capture-
recapture studies. Animal Biodiversity and Conservation, 27, 97–107.
Kendall, W. L. and Nichols, J. D. (2002). Estimating state-transition probabilities for
unobservable states using capture-recapture/resighting data. Ecology, 83, 3276–84.
Kendall, W. L., Nichols, J. D., and Hines, J. E. (1997). Estimating temporary emigration
using capture-recapture data with Pollock’s robust design. Ecology, 78, 563–78.
Krebs, C. J. 1972. Ecology. Harper and Row, New York.
Lancia, R. A., Kendall, W. L., Pollock, K. H., and Nichols, J. D. (2005). Estimating
the number of animals in wildlife populations. In C. E. Braun (ed.), Techniques for
Wildlife Investigations and Management, 6th edn, pp. 106–53. The Wildlife Society,
Bethesda, MD.
Lebreton, J. D., Burnham, K. P., Clobert, J., and Anderson, D. R. (1992). Modeling
survival and testing biological hypotheses using marked animals—a unified approach
with case-studies. Ecological Monographs, 62, 67–118.
MacKenzie, D. I., Nichols, J. D., Lachman, G. B., Droege, S., Royle, J. A., and Langtimm,
C. A. (2002). Estimating site occupancy rates when detection probabilities are less than
one. Ecology, 83, 2248–55.
MacKenzie, D. I., Nichols, J. D., Hines, J. E., Knutson, M. G., and Franklin, A. B. (2003).
Estimating site occupancy, colonization and local extinction probabilities when species
are not detected with certainty. Ecology, 84, 2200–7.
Occurrence. Academic Press, Boston, MA.
Muths, E., Jung, R. E., Bailey, L. L., Adams, M. J., Corn, P. S., Dodd, Jr, C. K., Fellers,
G. M., Sadinski, W. J., Schwalbe, C. R., Walls, S. C. et al. (2005). Amphibian Research
and Monitoring Initiative (ARMI): a successful start to a national program in the
United States. Applied Herpetology, 2, 355–71.
Nichols, J. D. and Pollock, K. H. (1990). Estimation of recruitment from immigration
versus in situ reproduction using Pollock’s robust design. Ecology, 71, 21–6.
Nichols, J. D., Hines, J. E., Lebreton, J.-D., and Pradel, R. (2000). Estimation of contri-
butions to population growth: a reverse-time capture-recapture approach. Ecology, 81,
3362–76.
Norris, J. L. and Pollock, K. H. (1996). Nonparametric MLE under two closed capture
recapture models with heterogeneity. Biometrics, 52, 639–49.
Otis, D. L., Burnham, K. P., White, G. C., and Anderson, D. R. (1978). Statistical-
inference from capture data on closed animal populations. Wildlife Monographs, 62,
1–135.
Pellet, J. and Schmidt, B. R. (2005). Monitoring distributions using call surveys: esti-
mating site occupancy, detection probabilities and inferring absence. Biological
Pledger, S. (2000). Unified maximum likelihood estimates for closed capture- recapture
models using mixtures. Biometrics, 56, 434–42.
Pollock, K. H. (1982). A capture-recapture design robust to unequal probability of cap-
ture Journal of Wildlife Management, 46, 757–60.
Pollock, K. H., Nichols, J. D., Simons, T. R., Farnsworth, G. L., Bailey, L. L., and
Sauer, J. R. (2002). Large scale wildlife monitoring studies: statistical methods for
design and analysis. Environmetrics, 13, 105–19.
Rexstad, E. and Burnham, K. P. (1991). User’s Guide for Interactive Program CAPTURE.
www.mbr-pwrc.usgs.gov/software/. Colorado Cooperative Fish and Wildlife Unit,
Fort Collins, CO.
Royle, J. A. (2004). Modeling abundance index data from anuran calling surveys.
Royle, J. A. and Link, W. A. (2005). A general class of multinomial mixture models for
anuran calling survey data. Ecology, 86, 2505–12.
Royle, J. A. and Dorazio, R. M. (2008). Hierarchical Modelling and Inference in Ecology.
Schaub, M., Gimenez, O., Schmidt, B. R., and Pradel, R. (2004). Estimating survival
and temporary emigration in the multistate capture-recapture framework. Ecology, 85,
2107–13.
Schmidt, B. R. and Anholt, B R. (1999). Analysis of survival probabilities of female com-
mon toads, Bufo bufo. Amphibia-Reptilia, 20, 97–108.
Schmidt, B. R., Feldmann, R., and Schaub, M. (2005). Demographic processes under-
lying population growth and decline in Salamandra salamandra. Conservation Biology,
19, 1149–56.
Schwarz, C. J., Schweigert, J. F., and Arnason, A. N. (1993). Estimating migration rates
using tag-recovery data. Biometrics, 49, 177–93.
Seber, G. A. F. (1965). A note on the multiple-recapture census. Biometrika, 52, 249–59.
Seber, G. A. F. (1982). The Estimation of Animal Abundance and Related Parameters.
Macmillian, New York.
Stanley, T. R. and Burnham, K. P. (1998) Estimator selection for closed-population
capture-recapture. Journal of Agricultural Biological and Environmental Statistics, 3,
131–50.
Stanley, T. R. and Burnham, K. P. (1999). A closure test for time-specific capture-recapture
data. Environmental and Ecological Statistics, 6, 197–209.
White, G. C. and Burnham, K. P. (1999). Program MARK: survival estimation from
populations of marked animals. Bird Study, 46, 120–39.
White, G. C., Anderson, D. R., Burnham, K. P., and Otis, D. L. (1982). Capture-recapture
and Removal Methods for Sampling Closed Populations. Los Alamos National Lab, Los
Alamos, NM.
Wood, K. V., Nichols, J. D., Percival, H. F., and Hines, J. E. (1998). Size-sex variation in
survival rates and abundance of pig frogs, Rana grylio, in northern Florida wetlands.
25
Quantifying abundance: counts,
detection probabilities, and estimates
Benedikt R. Schmidt and Jérôme Pellet
25.1 Background: imperfect detection in

amphibian ecology and conservation
Understanding temporal and spatial variation in distribution and abundance
has been, and will remain, a central goal in amphibian ecology and conser-
vation. Even though these two quantities, distribution and abundance, are so
fundamental, we usually cannot observe them directly in the field. We rarely
observe all individuals, all populations, or all species in an area of interest.
Imperfect detection is the rule rather than the exception and is a characteristic
that all field studies share.
Imperfect detection is a trivial fact and herpetologists are often aware of it.
For example, Hairston and Wiley (1993) attributed all fluctuations in sala-
mander counts to variation in weather conditions (salamanders tend to remain
underground during cold weather) and motivation of students (a class of highly
motivated students found an exceptionally high number of salamanders). This
implies that salamander detection was imperfect and, equally important, vari-
able among years (Hyde and Simons 2001).
Nevertheless, imperfect detection is all too often ignored when estimating
abundance or when temporal and spatial trends are analysed. In this chapter we
pinpoint patterns and the consequences of imperfect detection and show how to
deal with it in general (Chapter 24 introduces many of the estimators required to
deal with imperfect detection). We focus on the effects of imperfect detection on
population abundance estimation but we call attention to the fact that imperfect
detection also affects other ecological quantities, such as distribution and species
richness (e.g. Schmidt 2004; Royle and Dorazio 2006; Mazerolle et al. 2007).
We focus on how the issue of imperfect detection should be incorporated into the
study of amphibians populations; that is, how to sample amphibian populations

and how to estimate abundance. By doing so, we assume that clear study objec-
tives have been formulated in advance (Yoccoz et al. 2001) and sites appropriately
selected (see Chapter 23).
25.2 Imperfect detection

It is convenient to use a simple equation to conceptualize imperfect detection:
E (C ) Np
where N is the true value of the parameter of interest (i.e. number of individuals,
density, number of populations in an area, or species richness) and p is the detec-
tion probability (Gill 1985; Yoccoz et al. 2001; Pollock et al. 2002; Schmidt
2004). E( ) denotes a statistical expectation. The expectation E(C) is the average
of the count C over repeated realizations of the sampling process. This equation
has three major implications that we discuss below.
Sampling a population should be viewed as a stochastic process because it
involves a probability of detection. This is why there is a statistical expect-
ation E(C) in the above equation. Even under identical conditions we should
not expect to obtain the same result if we sample the same population mul-
tiple times. We should therefore expect variability in the counts. Variability in
counts (C) does not imply variation in abundance (N) or detection probability
(p); it can simply be random variation. Technically, counts are random varia-
bles. This is illustrated in Figure 25.1. The figure shows that, as expected under
25
p = 0.2 p = 0.5 p = 0.8
20
Frequency
15
10
0
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
Count
Fig. 25.1 Expected variation in counts of individuals under identical conditions for
100 repeated counts, a true population size N 20 and detection probabilities
p 0.2, 0.5, and 0.8. Data were simulated using the R code rbinom(n 100,
size 20, p x), where x had the values 0.2, 0.5, and 0.8, respectively).
25 Quantifying abundance | 467
binomial sampling, variability of the counts is greatest if p 0.5. The range of

likely counts is large no matter what the detection probability is.
25.2.1 Counts underestimate abundance

Because p can take any value between 0 and 1, counts almost always under-
estimate true abundance (or other biological parameters such as species rich-
ness or occupancy probabilities). Detection probabilities vary depending on the
methods used to sample a population. In studies using drift fences, detection
probabilities can be very close to 1 (Bailey et al. 2004a), meaning that very few
individuals escape detection. When using other methods, such as hand capture,
trapping, netting, or cover-board surveys, detection probabilities are usually far
below 1 (see, for example, case studies and estimates of detection probabilities in
Mazerolle et al. 2007). Consequently, the discrepancy between counts and true
abundances increases with decreasing detection probability. Table 25.1 shows
the correlations between counts and abundance where the true number of indi-
viduals was known. Counts and true abundances are somewhat correlated but
the proportion of variance explained can be quite low.
Pellet et al. (2007) analysed the relationship between the chorus counts,
number of captures and mark-recapture estimates of abundance in detail
(Table 25.2). Their study of the European treefrog Hyla arborea showed that
calling males represented a variable fraction of the total male population present
at breeding ponds. In two sites studied over 3 years, the proportion of males call-
ing varied between 0.32 and 0.65, suggesting that there is no solid link between
chorus activity and (estimated) population size. Similarly, their capture effort
(total number of males captured) did not truly reflect actual population size,
Table 25.1 Results of linear regressions between counts (C), population size esti-
mates (N̂), and censuses for tadpoles of two anuran species. R2 and F tests were
calculated using PROC GLM in SAS. Asterisks indicate significance at 0.05. The
Lincoln–Peterson estimator was used to estimate abundance. Table adapted from
Schmidt (2004). Data are from Jung et al. (2002).
Species Studied in Intercept Slope R2 F
Between count and census

Hyla arenicolor Natural ponds 13.92 1.73 0.66 34.21*
Scaphiopus couchii Mesocosms 12.44 1.17 0.96 283.1*
Between estimate and census

Hyla arenicolor Natural ponds 18.35 0.87 0.97 1005.0*
Scaphiopus couchii Mesocosms 15.25 0.94 0.99 1509.0*
Table 25.2 Counts and estimates of male European tree frogs (Hyla arborea) in
two breeding aggregations. Maximum and mean chorus size, as well as number of males
captured, correlate only weakly with actual (estimated) male population size, demon-
strating the effects of imperfect detection on population size estimates. Adapted from
Pellet et al. (2007).
Year Maximum Mean Total males Modeled male Proportion of
chorus size chorus captured population calling males
size size SE
Males in 2002 27 11.4 35 57.9 9 47%

population 2003 18 6.8 34 49.6 6.9 36%
Camp Romain 2004 20 12.5 75 62.1 15 32%
Males in 2002 25 6.9 29 38.5 5.7 65%
population Les 2003 15 7.2 30 30.8 1.3 49%
Mossières 2004 20 7.3 45 46.8 2.2 43%
thus demonstrating that population trends estimated from counts (callers or

captures) are likely to reflect a mixture of population and detectability trends.
In addition, Pellet et al. (2007) compared mean chorus counts and maximum
chorus counts. Mean counts were only 25–63% of the maximum counts. Such
variability probably arises from the fact that a maximum count is an extreme
value that is likely to be highly variable spatially and temporally. We argue that
the maximum count is not a suitable metric if the goal of a study is a comparison
between years and/or sites. If anything, mean counts would be better compar-
able across sites and years than maxima (but see Schmidt and Pellet 2005 for a
study where maximum counts predicted population persistence).
Funk et al. (2003) compared the efficiency of abundance estimates based on
visual encounter surveys, distance sampling, and mark–recapture methods for
monitoring population trends of forest-floor-dwelling Eleutherodactylus frogs.
They found that mark–recapture methods were best at estimating abundance
and the method had the greatest power to detect population declines. Like in
the examples above, the message is clear: estimates clearly outperform counts in
reflecting true population abundances.
25.2.2 Per-visit and cumulative detection probabilities

Detection probabilities come in two flavors: per visit and cumulative. Whereas
per-visit detection probabilities may be low, cumulative detection probabilities
are usually much higher. Cumulative detection probability, pc, is given by
pc 1(1 p )n
_
where p is the average per-visit detection probability and n is the number of
visits or capture events. Per-visit (i.e. per day) detection probabilities of the frog
Colostethus stepheni in the study of Funk and Mills (2003) ranged from 0 to 1
with a mean of 0.58 (W.C. Funk, personal communication). Because Funk and
Mills (2003) had six capture events during a short time period when the popu-
lation did not change, the cumulative capture probabilities were greater than
0.987. This implies that Funk and Mills (2003) achieved an almost complete
census of the frogs and suggests that multiple capture events are an obvious way
to deal with imperfect detection because cumulative detection probabilities can
be quite high.
25.2.3 Temporal and spatial variation in detection probabilities

Counts of animals are usually made to assess temporal or spatial variation in
population size. Such comparisons are only valid if E(p) remains constant in
time or space; however, this is hardly ever the case (MacKenzie and Kendall
2002), even under strictly standardized methods. If p varies temporally and/or
spatially then variation in p and N are confounded. In the form of an equation,
this may be expressed as
E (T )C1 /C 2 N 1 p1 /N 2 p2
Spatial or temporal variation in pi can lead to the detection of spurious temporal

or spatial trends E(T ). The contrary may also be true: one may miss true pat-
terns in abundance or (under- or) overestimate the effects of variables that are
thought to explain variation in abundance (Mazerolle et al. 2005). Obviously,
the absolute values of p matter. If p is high, then bias will be low. If p is low, how-
ever, then most variation in C is likely due to variation in p. For example, if p var-
ies between 80 and 90%, then variation in C will be relatively small. However,
if a low p varies by the same absolute amount, say between 10 and 20%, then
variation in C will be relatively greater.
Temporal and spatial variation is the rule rather than the exception. Tacitly,
Hairston and Wiley (1993) attributed variation in salamander counts entirely to
variation in detection probability. In the study of Funk and Mills (2003), p varied
from 0 to 1 depending on prevailing weather conditions (W.C. Funk, personal
communication). Daily means (across sites) ranged from 0.43 to 0.75 and different
sites had different means (across days) ranging from 0.43 to 0.90 (W.C. Funk, per-
sonal communication). Most variation in detection probabilities of the European
treefrog H. arborea in the study of Pellet et al. (2007) was between sites whereas
Schmidt et al. (2007) found that detection probabilities of Salamandra salamandra
were low in autumn and high in spring. Bailey et al. (2004a) used a drift fence to
capture salamanders (Ambystoma tigrinum) on their way to and from the pond.
Detection probabilities were high in most years (greater than 90%) but inexplicably
lower in one year (76%). Although detection probabilities of egg masses of Rana
sylvatica and Ambystoma maculatum were generally high (usually 80%), Grant
et al. (2005) documented substantial spatial and temporal variation in detection
probabilities that could bias estimates of population trends based on unadjusted
counts.
25.3 Components of imperfect detection

There are many reasons why detection is imperfect and it makes sense to decom-
pose detection probabilities into its components (Pollock et al. 2004; Nichols
et al. 2008):
E (C ) Npe
where e is exposure to sampling (also referred to as availability for sampling).

In this equation, p should best be called detection probability given exposure.
Nichols et al. (2008) describe how detection probability and exposure to sam-
pling can be further decomposed. Detection probability (p) and exposure to
sampling (e) are easy to distinguish at the conceptual level. In practice, the dis-
tinction may not always be obvious.
Under different names, exposure to sampling is a well-known phenomenon
of mark–recapture studies dealing with survival estimation. The two cases
are transients and temporary emigrants (Williams et al. 2002). Transients are
animals that show up only once in the study area and then leave. Temporary
emigrants are animals that leave the study area temporarily during the sam-
pling period and then return (e.g. animals that skip a breeding season). In other
words, these animals are only partly exposed to sampling and this can cause
negative bias in survival estimates.
Incomplete exposure to sampling also affects abundance estimation (Kendall
1999; Bailey et al. 2004b; Royle and Dorazio 2006). Non-exposure to sam-
pling can result from both biological and methodological reasons. Biological
reasons include brooding salamanders that are underground and not at the sur-
face during the time of survey or frogs that do not breed in a particular year
and hence do not migrate to the study site which may be a breeding site (e.g.
pond). Another biological reason may be that only a proportion of all males
present at a pond may be calling at a particular time. This would explain why
there is a difference between mean and maximum chorus counts (Pellet et al.
2007). Methodological reasons may play a part when funnel or minnow traps
are placed along the edge of a pond and some individuals spend all their time
near the center of the pond. Non-exposure to sampling may also result from
breeding phenology. Some individuals may breed early and some may breed
later in the season (e.g. Sinsch 1988). If the population is sampled only early in
the season, then many individuals will not be exposed to sampling. This might
be a case where both biology and method cause non-exposure to sampling.
Non-exposure to sampling can have profound consequences for abundance
estimation (Kendall 1999). Imagine that the goal of a study is to estimate abun-
dance in a particular area. If amphibians do not move in and out of the study
area, then an abundance estimator such as the Lincoln–Peterson estimator will
provide an unbiased estimate the number of amphibians in the area. However,
if some individuals move randomly in and out of the area (i.e. if they are not
always exposed to sampling), then the very same estimator will estimate a differ-
ent quantity. It will now estimate the size of the superpopulation where the super-
population is defined as the total number of amphibians exposed to sampling at
least once. This includes all amphibians that are residents in the study area, but
also all individuals that move in and out or that move through the study area.
Bailey et al. (2004b) encountered a related problem in their study of salamanders
in the Appalachian Mountains, USA. They argued that short-term studies where
salamander movement was negligible yielded estimates of the “surface” popu-
lation (i.e. salamanders exposed to sampling only). Long-term studies, where
salamanders had time to move from the surface to deeper ground and vice versa,
gave estimates of the total number of salamanders in the sampled area. Kinkead
and Otis (2007) describe a similar situation with breeding and non-breeding
ambystomatid salamanders that were sampled at the breeding site.
If many individuals are not exposed to sampling (low e), then the mismatch
between the spatial or temporal scale at which a population is sampled and the
desired temporal or spatial scale of inference is likely to be great. In conclusion,
study design (both the spatial and the temporal scales) and species behavior can
jointly determine which biological entity is quantified.
25.4 How to deal with imperfect detection

25.4.1 Estimation of abundance
Evidently, an elegant way to deal with imperfect detection is to estimate detec-
tion probability (p̂). An estimate of detection probability can then be used to
correct counts (C ) and estimate abundance (N̂ ):
Nˆ C / pˆ
This equation is the conceptual basis for all kinds of abundance estimators, be
they mark–recapture, distance sampling, point count, removal, or other meth-
ods (Williams et al. 2002; Mazerolle et al. 2007; see also Chapter 24 in this
volume). When non-exposure to sampling is a problem, then N̂ can only be
estimated if one knows the fraction of the population that is exposed to sam-
pling as well as the probability of detecting exposed individuals, populations,
or species:
Nˆ C / pe
ˆˆ
The best tool to estimate abundance in the presence of non-exposure to sam-

pling is the robust design (see Williams et al. 2002) which has been successfully
used with plethodontid and ambystomatid salamanders (Bailey et al. 2004b;
Kinkead and Otis 2007). The robust design allows estimating both abundance
and non-exposure (known as temporary emigration in this particular case).
In the previously cited case of the European treefrog (H. arborea), the robust
design also allowed demonstrating that annual male temporary emigration was
negligible (i.e. that males rarely skipped breeding seasons; Pellet et al. 2007). In
contrast, Bailey et al. (2004a) showed that temporary emigration in A. tigrinum
was substantial.
Other approaches to dealing with non-exposure were described by Royle and
Dorazio (2006) and Condit et al. (2007). Royle and Dorazio (2006) describe
a method for point counts that allows dealing with non-exposure to sampling
that arises from a mismatch between the scale at which data was collected and
the desired scale of inference. They describe a case where quadrats have only
been partially searched (in that study, only a fraction of the area of 1 km2 quad-
rats was surveyed). Their method works if there is a suitable covariate (such as
transect length) that can statistically link the exposed population to the true
population.
Condit et al. (2007) developed a method for estimating the size of a popula-
tion when individuals are asynchronously present. Their method may be par-
ticularly relevant for pond-breeding amphibians where the breeding season of
the population is longer than the breeding season of an individual (e.g. in the
natterjack toad Bufo calamita and the European treefrog H. arborea; Sinsch
1988; Friedl and Klump 2005).
Methods to estimate abundance are not without problems. Detection prob-
abilities can be low. This will have the effect that confidence intervals can be
very wide (Williams et al. 2002). It is possible to make confidence intervals
shrink with more effort or better capture techniques. That is, researchers should
either use better methods that increase per-visit detection probabilities or
increase the number of repeat visits such that cumulative detection probabilities
are increased. However, standard errors and confidence intervals are not a nuis-
ance. Rather, they are an advantage of estimation methods. Standard errors and
confidence intervals are a measure of uncertainty that allow an assessment of
the estimates’ reliability. Consequently, we view wide confidence intervals as an
honest statement whether a particular estimate is or is not particularly reliable
and useful. Wide confidence intervals are no reason to discard mark–recapture
estimates and to prefer the simple counts (C; e.g. Alford and Richards 1999).
There is also uncertainty associated with counts (because detection probability
p is unknown) but it is not explicit and is, in fact, unknowable.
Heterogeneity in detection probabilities among individuals can be a problem
in mark–recapture studies (Link 2003). Heterogeneity usually leads to nega-
tive bias in abundance estimates and in the worst case it may not be possible to
identify a best model that should be used for inference. We believe that amphib-
ian ecologists and conservationists should attempt to minimize detection prob-
ability heterogeneity among individuals by adopting methods that account
for variation in detectability among individuals (i.e. grouping individuals into
homogeneous sets by sexes, colour morphs, age classes).
25.4.2 Other approaches to dealing with imperfect detection

Amphibian ecologists have dealt with imperfect detection in many ways. Some
authors simply did not analyse data from species where detection was uncertain
and variable. For example, Pechmann et al. (1991) analysed data from ambysto-
matid salamanders that are unlikely to trespass a drift fence but they did not
analyse data from treefrogs that could easily trespass a drift fence.
The most common objection to the use of methods that allow estimating
population sizes and detection probabilities instead of counts is that they are
demanding in terms of money, human resources, and statistical knowledge.
This argument was not true in the detailed study of Funk et al. (2003) where
different methods were compared. We discard this argument because every con-
servation action based on counts is likely to be biased to the point where they
will be inefficient or, worse, counterproductive (Yoccoz et al. 2001). Moreover,
it will be impossible to evaluate the success of actions taken because no data will
be available to detect population trends accurately, and to adapt management
actions accordingly.
Some authors have expressed the view that there is no need to estimate detec-
tion probabilities and to adjust counts accordingly (i.e. estimate abundance).
One argument put forward in the context of long-term monitoring programs
is that variability in detection probability does not matter as long as there is no
temporal trend in detection probability (Bart et al. 2004; tacitly, this is also the
reasoning of Hairston and Wiley 1993). In such a situation, variation in detec-
tion probability likely causes extra variability (sampling variability) in the counts
(in comparison to variation in absolute abundance which is the phenomenon of
biological interest). Detectability-induced extra variation in the counts means
that a monitoring program loses power to detect temporal trends. However, one
should keep in mind that detection probabilities likely show temporal trends.
Reasons include, but are not limited to, habitat succession or changes through
time in the ability of the observer to detect the study species (Link and Sauer
1997).
One commonly held view is that field methods can be standardized to the
extent where detection probability is constant. If this is the case, then counts
or other estimates of relative abundance should serve well as proxies for abso-
lute abundance. Unfortunately, variation in detection probabilities is the rule
rather than the exception (MacKenzie and Kendall 2002). Whenever detec-
tion probabilities of amphibians have been estimated, they were found to be
variable both within and across seasons (Bailey et al. 2004a, Kinkead and Otis
2007; Mazerolle et al. 2007; Pellet et al. 2007; Schmidt et al. 2007). This
was the case even when researchers used standard(ized) methods; even when
drift fences were used—where the assumption is that detection probability is
1—there was variation in detection probability (Bailey et al. 2004a). Pellet
et al. (2007) used the same methods at two sites yet detection probabilities dif-
fered between sites by a factor of approximately two. Hyde and Simons (2001)
showed that counts obtained from applying four standard methods gave results
that were only weakly correlated. That is, the use of standardized methods does
not guarantee that detection probabilities are constant. We believe that stand-
ardization of field methods is important because it can help to keep variation
in detection probabilities within bounds, but it should certainly not be viewed
as a panacea.
Standardization of field methods is one solution to limit variation in detec-
tion probabilities. Another solution is to measure covariates that may affect
detection probabilities and use these covariates at the analysis stage to adjust
counts (Link and Sauer 1997, 1998). This approach may work well as long as the
important covariates are known and has been successfully used in large-scale
bird monitoring programs. However, it may be that the effect of a covariate on
the counts varies from one site to the next. Lauber (2004) counted alpine sala-
manders (Salamandra atra) along fixed transects at four sites and tested whether
weather covariates could be used to predict the salamander counts. An analysis
of covariance found no main effect of air humidity on counts but there was
100
Study sites:
Cholerenschlucht
Weissenburgbad
80 Rüschegg-Heubach
Salamander count Grünenbergpass
60
40
20
0
60 70 80 90 100
Air humidity (%)
Fig. 25.2 Geographic variation in the relationship between salamander (Salamandra

atra) counts and air humidity. Salamanders were counted multiple times at four
sites along transects. Analysis of covariance showed no significant main effect of air
humidity on counts but there was a significant interaction between air humidity and
site. Data were taken with permission from Lauber (2004).
a significant site-by-air-humidity interaction (Figure 25.2). Air humidity can

thus be used to adjust counts at some sites but not at others; a likely explanation
is that sites differ in overall humidity. Grant et al. (2005) also found that no sin-
gle explanatory variable or set of variables best explained variation in detection
probabilities across sites of egg masses of Rana sylvatica and Ambystoma macula-
tum. This implies that herpetologists must be very cautious when extrapolating
results from one study site to another. At the planning stage, it also implies that
one should try to replicate all experiments spatially and temporally to find out
whether biological patterns are universally applicable.
In conclusion, we believe that the use of standard methods is always valuable.
Because it does not always avoid variation in detection probabilities, it is better
to rely on adjusting counts than on the strong assumption that detection prob-
abilities are not showing any trends. We argue that one should assume a priori
that detection probabilities are less than one and that they are variable in space
and time. Thus, amphibian ecologists and conservationists should provide evi-
dence for their studies that the counts that they report are indeed reliable indices
of abundance. Because detection probabilities can vary from one study to the
next even under apparently highly similar conditions (see examples in Mazerolle
et al. 2007), the proof has to be provided every time anew.
25.5 Designing a sampling protocol

The adequate design of a sampling protocol is a fundamental aspect of
research, whether it be on individuals, populations, or species. Once the
research or monitoring question addressed has been explicitly formalized
(Yoccoz et al. 2001) and sites were appropriately selected (see Chapter 23 in
this volume), it is essential that aspects of imperfect detection are incorpo-
rated into the design. To do so, the procedures of data collection and data
analysis must be identified in advance. Having a sound knowledge of your
study species’ ecology will be instrumental in determining which biological
quantity (e.g. above-ground population size, breeders, super-population)
your estimator will represent.
Importantly, one must be fully aware that detection probability (p) is not
only a species trait (see examples in Mazerolle et al. 2007); it also depends
on methods, observers, year, site, and a myriad of other factors. Because p
(and exposure to sampling, e) are both variable in space and time, one can-
not apply values obtained in one study to another. For these reasons we rec-
ommend that detection probability is explicitly integrated in all amphibian
study protocols. Non-exposure to sampling is often hard to deal with at the
analysis stage (except when using the robust design). We therefore recom-
mend that researchers carefully plan a study such that all animals are exposed
to sampling.
25.6 Software
There are many computer programs freely available to estimate population
abundance while incorporating detection probability. The most versatile and
widely used of them is program MARK (White and Burnham 1999). This
software, available at www.phidot.org/, allows the analysis of a wide range of
capture–recapture-based data sets. Every new version incorporates the latest
development in capture–recapture and thus allows the user to choose from a
wide range of models the one that will fit its data best. As the name implies,
DISTANCE (www.ruwpa.st and.ac.uk/ distance/) is the software tool that
allows one to design and analyse distance sampling surveys. More recent devel-
opments have been integrated in statistical software such as R and WinBugs.
These tools have the inconvenience of being less user-friendly than the previously
listed programs, which benefit from a graphical user interface. There is soft-
ware that can be used when planning a mark–recapture study (Devineau et al.
2006; Zucchini et al. 2007).
25.7 Outlook
Imperfect detection is the feature that the vast majority of all amphibian surveys
have in common. Complete censuses where all amphibians that are present at
the study site are captured are not impossible, but require a lot of work (drift
fences: Bailey et al. 2004a; many capture events: Funk and Mills 2003; Pellet
et al. 2007). We argue that detection probabilities can be low and highly variable
among years and/or sites. While counts that are not adjusted for imperfect detec-
tion can certainly indicate negative population trends (Laurance et al. 1996),
variability in imperfect detection can seriously bias inference from surveys.
Amphibian ecologists and conservationists should therefore estimate detection
probabilities as the best tool to calibrate a survey and use robust methods for
abundance estimation (Williams et al. 2002; Chapter 24). Unfortunately, the
use of such methods is not yet widespread (Alford and Richards 1999).
The number of methods available for estimation of abundance that account
for imperfect detection has increased tremendously in the recent past. Existing
methods are constantly being refined, while new methods are being developed
(e.g. Royle 2004; Royle and Dorazio 2006). Still, all methods need to be used
with care as sampling design, the behavior of the species, and the estimator used
all determine which biological quantity is being estimated. Notwithstanding,
the quality of inference from methods that adjust for imperfect detection will be
stronger than inference from any other kind of method.
In the future, we ought to be able to estimate abundance with a precision and
freedom from bias that was not achievable in the past. We should now be able
to determine which factors influence abundance rather than study patterns of
an inseparable combination of abundance, detectability, and exposure to sam-
pling (such as counts). This will help us gain new insights into fundamental and
applied aspects of amphibian ecology and conservation.
25.8 References
ecology. Annual Reviews of Ecology and Systematics, 30, 133–65.
Bailey, L. L., Kendall, W. L., Church D. R., and Wilbur, H. M. (2004a). Estimating
survival and breeding probability for pond-breeding amphibians: a modified robust
design. Ecology, 85, 2456–66.
Bailey, L. L., Simons, T. R., and Pollock, K. H. (2004b) Comparing population size esti-
mators for plethodontid salamanders. Journal of Herpetology, 38, 370–80.
Bart, J., Droege, S., Geissler, P., Peterjohn, B., and Ralph, C. J. (2004). Density estima-
tion in wildlife surveys. Wildlife Society Bulletin, 32, 1242–7.
Condit, R., Le Boeuf, B. J., Morris, P. A., and Sylvan, M. (2007). Estimating population
size in asynchronous aggregations: A Bayesian approach and test with elephant seal
censuses. Marine Mammal Science, 23, 834–55.
Devineau, O., Choquet, R., and Lebreton, J. D. (2006). Planning capture-recapture stud-
ies: straightforward precision, bias, and power calculations. Wildlife Society Bulletin,
34, 1028–35.
Friedl, T. W. P. and Klump, G. M. (2005). Sexual selection in the lek-breeding European
treefrog: body size, chorus attendance, random mating and good genes. Animal
Behaviour, 70, 1141–54.
Funk, W. C. and Mills, L. S. (2003). Potential causes of population declines in forest frag-
ments in an Amazonian frog. Biological Conservation, 111, 205–14.
Funk, W. C., Almeida-Reinoso, D., Nogales-Sornosa, F., and Bustamante, M. R. (2003).
245–56.
Gill, D. E. (1985). Interpreting breeding patterns from census data: a solution to the
Husting dilemma. Ecology, 66, 344–54.
Grant, E. H. C., Jung, R. E., Nichols, J. D., and Hines, J. E. (2005). Double-observer
approach to estimating egg mass abundance of pool-breeding amphibians. Wetlands
Hairston, Sr, N. G. and Wiley, R. H. (1993). No decline in salamander (Amphibia: Caudata)
populations: a twenty-year study in the southern Appalachians. Brimleyana, 18, 59–64.
Hyde, E. J. and Simons, T. R. (2001). Sampling plethodontid salamanders: sources of
Jung, R. E., Dayton, G. H., Williamson, S. J., Sauer, J. R., and Droege, S. (2002). An
Kendall W. L. (1999). Robustness of closed capture-recapture methods to violations of
the closure assumption. Ecology, 80, 2517–25.
Kinkead, K. E. and Otis, D. L. (2007). Estimating superpopulation size and annual prob-
ability of breeding for pond breeding salamanders. Herpetologica, 63, 151–62.
Lauber, A. (2004). Methodenevaluation zum Monitoring der Alpensalamanderpopulation.
Diploma thesis, Eidgenössische Technische Hochschule Zürich (ETHZ), Zürich.
Laurance, W. F., McDonald, K. R., and Speare, R. (1996). Epidemic disease and the cata-
strophic decline of Australian rain forest frogs. Conservation Biology, 10, 406–13.
Link, W. A. (2003). Nonidentifiability of population size from capture-recapture data
with heterogeneous detection probabilities. Biometrics, 59, 1123–30.
Link, W. A. and Sauer, J. R. (1997). Estimation of population trajectories from count
data. Biometrics, 53, 488–97.
Link, W. A. and Sauer J. R. (1998). Estimating population change from count data:
Application to the North American Breeding Bird Survey. Ecological Applications, 8,
258–68.
MacKenzie, D. I. and Kendall, W. L. (2002). How should detection probability be incor-
porated into estimates of relative abundance? Ecology, 83, 2387–93.
(2007). Making great leaps forward: accounting for detectability in herpetological field
Nichols, J. D., Thomas, L., and Conn, P. B. (2008). Inferences about landbird abun-
dance from count data: recent advances and future directions. In D. L. Thomson,
E. G. Cooch, G. Evan, and M. J. Conroy (eds), Modeling Demographic Processes in
Marked Populations, pp. 203–38. Springer ScienceBusiness Media, New York.
Pechmann, J. H. K., Scott, D. E., Semlitsch, R. D., Caldwell, J. P., Vitt, L. J., and
Pellet, J., Helfer, V., and Yannic, G. (2007). Estimating population size in the European
tree frog (Hyla arborea) using individual recognition and chorus counts. Amphibia-
Reptilia, 28, 287–94.
Pollock, K. H., Nichols, J. D., Simons, T. R., Farnsworth, G. L., Bailey, L. L., and
Sauer, J. R. (2002). Large scale wildlife monitoring studies: statistical methods for
design and analysis. Environmetrics, 13, 105–19.
Pollock, K. H., Marsh, H., Bailey, L. L., Farnsworth, G. L., Simons, T. R., and
Alldredge, M. W. (2004). Separating components of detection probability in abun-
dance estimation: an overview with diverse examples. In W. L. Thompson (ed.),
Sampling Rare or Elusive Species, pp. 43–58. Island Press, Washington DC.
Royle, J. A. (2004). N-mixture models for estimating population size from spatially rep-
licated counts. Biometrics, 60, 108–15.
Royle, J. A. and Dorazio, R. M. (2006). Hierarchical models of animal abundance
and occurrence. Journal of Agricultural, Biological, and Environmental Statistics, 11,
249–63.
Schmidt, B. R. (2004). Declining amphibian populations: the pitfalls of count data in
the study of diversity, distributions, dynamics and demography. Herpetological Journal,
14, 167–74.
Schmidt, B. R. and Pellet, J. (2005). Relative importance of population processes and
habitat characteristics in determining site occupancy of two anurans. Journal of Wildlife
Schmidt, B. R., Schaub, M., and Steinfartz, S. (2007). Apparent survival of the sala-
mander Salamandra salamandra is low because of high migratory activity. Frontiers in
Zoology, 4, e19.
Sinsch, U. (1988). Temporal spacing of breeding activity in the natterjack toad, Bufo
calamita. Oecologia, 76, 399–407.
White, G. C. and Burnham, K. P. (1999). Program MARK: survival estimation from
populations of marked animals. Bird Study, 46, 120–38.
Yoccoz, N. G., Nichols, J. D., and Boulinier, T. (2001). Monitoring of biological diversity
in space and time. Trends in Ecology and Evolution, 16, 446–53.
Zucchini, W., Borchers, D. L., Erdelmeier, M., Rexstad, E., and Bishop, J. (2007). WiSP
1.2.4. Institut für Statistik und Ökonometrie, Georg-August-Universität Göttingen,
Göttingen.
26
Disease monitoring and biosecurity
D. Earl Green, Matthew J. Gray, and Debra L. Miller
26.1 Introduction
Understanding and detecting diseases of amphibians has become vitally import-
ant in conservation and ecological studies in the twenty-first century. Disease is
defined as the deviance from normal conditions in an organism. The etiologies
(causes) of disease include infectious, toxic, traumatic, metabolic, and neoplas-
tic agents. Thus, monitoring disease in nature can be complex. For amphib-
ians, infectious, parasitic, and toxic etiologies have gained the most notoriety.
Amphibian diseases have been linked to declining amphibian populations, are
a constant threat to endangered species, and are frequently a hazard in captive
breeding programs, translocations, and repatriations. For example, a group of
viruses belonging to the genus Ranavirus and the fungus Batrachochytrium den-
drobatidis are amphibian pathogens that are globally distributed and responsible
for catastrophic population die-offs, with B. dendrobatidis causing known spe-
cies extinctions (Daszak et al. 1999; Lips et al. 2006; Skerratt et al. 2007). Some
infectious diseases of amphibians share similar pathological changes; thus, their
detection, recognition, and correct diagnosis can be a challenge even by trained
veterinary pathologists or experienced herpetologists.
This chapter will introduce readers to the most common amphibian dis-
eases with an emphasis on those that are potentially or frequently lethal, and
the techniques involved in disease monitoring. It will also outline methods of
biosecurity to reduce the transmission of disease agents by humans. We start by
covering infectious, parasitic, and toxic diseases. Next, surveillance methods
are discussed, including methods for sample collection and techniques used in
disease diagnosis. Finally, biosecurity issues for preventing disease transmission
will be covered, and we provide protocols for disinfecting field equipment and
footwear.
26.2 Amphibian diseases of concern

Amphibians are susceptible to a variety of pathogens, including internal and
external parasites, viruses, bacteria, and fungi. Each of the three major life stages
of amphibians (embryos, larvae, and adults) has a distinct suite of diseases, with
some overlap between life stages. Aquatic amphibian embryos and larvae share
many diseases with fish, whereas post-metamorphic stages often share few infec-
tious diseases with earlier life stages. For detailed information on the amphibian
diseases, we recommend that readers consult recent reviews (e.g. Converse and
Green 2005a, 2005b; Green and Converse 2005a, 2005b) and the veterinary
literature (e.g. Wright and Whitaker 2001).
26.2.1 Infectious diseases

Major infectious diseases for each amphibian life stage are summarized in
Tables 26.1–26.3. Many viruses have been reported in amphibians, and include
Ranavirus, herpesvirus, and adenovirus (Converse and Green 2005a, 2005b;
Green and Converse 2005a, 2005b). Of these, Ranavirus has been the most sig-
nificant contributor to population declines, resulting in significant morbidity
and mass mortality (Daszak et al. 1999; Green et al. 2002; Cunningham et al.
2007). In North America, ranaviruses are responsible for the majority of cata-
strophic die-offs in ambystomid salamanders and late-stage anuran larvae, with
the number of reported cases each year exceeding all other pathogens by three to
four times (Green et al. 2002; Muths et al. 2006). Although many die-offs have
been with common species, declines in several species of conservation concern
(e.g. Rana muscosa, Rana aurora, Bufo boreas, and Ambystoma tigrinum steb-
binsi) have been reported (Jancovich et al. 1997; Converse and Green 2005a).
There is evidence that ranaviruses may function as a novel or endemic pathogen,
with the former likely associated with the movement of infected amphibians by
humans (Storfer et al. 2007). Anthropogenic stressors also may facilitate emer-
gence (Forson and Storfer 2006; Gray et al. 2007a). Additionally, subclinically
infected individuals (i.e. those that do not appear sick) may serve as reservoirs
for more susceptible amphibian species (Brunner et al. 2004).
Likewise, numerous bacteria have been cultured from anurans (Mauel et al.
2002). Of these, Mycobacterium liflandii, a mycolactone-producing myco-
bacteria, is of concern because it is closely related to the human pathogen
Mycobacterium ulcerans (Yip et al. 2007), which causes severe skin lesions in
humans. Nevertheless, Aeromonas hydrophila remains the most recognized bac-
terial pathogen in amphibians because of its association with red-leg disease
26 Disease monitoring and biosecurity | 483
Table 26.1 Significant diseases of amphibian eggs and embryos.

Disease agent Common Mortality Organ of Test methods
host species rate choice
Lucke tumor herpesvirus Rana pipiens only 0 Mesonephros Culture on Rana

or whole pipiens cell line;
embryo PCR
Chlamydomonas Ambystoma spp. 0 Egg capsule Gross or
(symbiotic alga)1 microscopic exam
Watermolds2 Bufo spp., Hyla Variable Egg capsule Culture;
spp., Pseudacris histology; DNA
spp., Rana spp., sequencing
Ambystoma spp.,
Taricha spp.
Microsporidium schuetzi Rana pipiens
10% Whole Histology;
swollen eggs electron
microscopy
Tetrahymena/Glaucoma Ambystoma spp. 15–25% Egg capsule, Submerged exam
(ciliated protozoa) brain and of eggs/embryos
subcutis of under dissecting
embryos/ microscope;
larvae histology; exam
by protozoologist
1
Chlamydomonas sp. is a symbiotic blue-green alga in the egg capsule of Ambystoma maculatum in
eastern North America and Ambystoma gracile in western North America, and not considered a
disease agent.
2
Watermold infections (oomycetes of several genera) referred to as saprolegniasis.
(Green and Converse 2005a). However, it is important to note that red-leg dis-
ease is a gross descriptor of a specific lesion (i.e. swollen red legs) and not specific
for a particular etiology. Many pathogens (e.g. Ranavirus, A. hydrophila, alveo-
lates) can cause edema (i.e. swelling) and erythema (reddening) in amphibians
(Figure 26.1a). This emphasizes the importance of diagnostic testing to deter-
mine the correct pathogen causing the disease.
Finally, numerous fungal and fungus-like organisms (Converse and Green
2005a, 2005b; Green and Converse 2005a, 2005b) and newly characterized
pathogens (Davis et al. 2007) are known to cause catastrophic mortality of
amphibian populations. B. dendrobatidis (Figure 26.1b) has resulted in global
population declines and species extinctions (Wake and Vredenburg 2008).
The newly discovered alveolate organism has only been diagnosed in a few
Table 26.2 Significant diseases of larval amphibians.
Disease agent Common host species Mortality rate Organ of choice Test methods
Ranaviruses Bufo spp., 50–99% Liver, skin ulcers Culture at 20–25°C;

Hyla spp., mesonephroi1 PCR on liver, spleen, skin ulcers,
Rana spp., mesonephroi
Pseudacris spp.,
Ambystoma spp.,
Notophthalmus spp.
Batrachochytrium dendrobatidis Rana spp., Pseudacris spp. 0% Oral disc, toe tips Histology; PCR; culture
1
Watermolds Bufo spp., Hyla spp., P Variable Oral disc, skin Culture; histology
seudacris spp., Rana spp.,
Ambystoma spp., Taricha spp.
Ichthyophonus sp. Pseudacris spp., Rana spp., 0–≈50% Skeletal muscle Histology
Ambystoma spp.
Perkinsus-like organism Rana spp., rarely Hyla spp., 5–99% Liver Histology; PCR
rarely Pseudacris spp.
Tetrahymena/glaucoma Ambystoma spp. 15–25% Egg capsule, brain and Submerged exam of eggs/embryos
(ciliated protozoa) subcutis of embryos/larvae under dissecting microscope; histology;
examination by protozoologist
Ribeiroia ondatrae Bufo spp., Pseudacris spp., Variable Skin around vent and Examination by parasitologist; PCR
Rana spp., Ambystoma spp. proximal hindlimbs
Other metacercariae Most aquatic genera Low Parasite Examination by parasitologist
Lernaea sp. (“anchorworm”) Rana spp. Low Parasite in skin Examination by parasitologist
Leeches Rana spp. Low Parasite Examination by parasitologist
1
Mesonephroi, “body kidneys” (versus pronephroi or “head kidneys” found only in larvae); “true” kidneys of reptiles, birds, and mammals are metanephroi.
2
Watermold infections (oomycetes of several genera) referred to as saprolegniasis.
Table 26.3 Significant diseases of post-metamorphic amphibians.
Disease agent Common host species Mortality rate Organ of choice Test methods
Ranaviruses Bufo spp., Hyla spp., Low in adults; variable Liver, skin ulcers Culture at 20–25°C;
Pseudacris spp., Rana spp., in recently mesonephroi PCR on liver, spleen, skin
Ambystoma spp., metamorphosed ulcers, mesonephroi;
Notophthalmus spp. amphibians electron microscopy
Lucke tumor Rana pipiens Variable in >2 yr old Mesonephroi Histology of tumors
herpesvirus Rana pipiens only
Batrachochytrium Most anuran genera Very high in many anurans Skin of pelvic patch, Histology; PCR; culture;
dendrobatidis especially at high elevations in toe webs electron microscopy
tropical latitudes
Ichthyophonus sp. Rana spp., Very low Skeletal muscle Histology
Ambystoma spp.,
Notophthalmus spp.
Amphibiothecum Bufo spp. 0 Ventral skin nodules Cytology of discharge; histology
penneri1 of nodule
Hepatozoon spp. Rana spp. Low Blood smear; liver Cytology; histology
Ribeiroia ondatrae Bufo spp., Pseudacris spp., 0 Skin around vent, Examination of metacercaria by a
Rana spp., Ambystoma spp. proximal hindlimbs parasitologist; PCR; radiographs of
and at tip of urostyle malformations
Rhabdias spp. Bufo spp. Unknown Lungs Visible at dissection; histology;
(nematode lungworm) Rana spp. examination by a
parasitologist
1
Amphibiothecum (formerly Dermosporidium) penneri, referred to as dermosporidiosis.
(a)
(b)
(c)
Fig. 26.1 (a) Tadpoles with swollen bodies and swollen red legs (arrow) are often
diagnosed as red-leg disease but the etiology is varied and may include Aeromonas
hydrophila, Ranavirus, and alveolates. (b) The amphibian fungus, Batrachochytrium
dendrobatidis (arrows), infects the keratin-producing cells of amphibians. Tadpole skin
is not keratinized; rather, only their ‘teeth’ contain keratin. Grossly, this is seen by loss
of pigmentation (upper inset) of the tooth rows. Lower inset is of a normal tadpole
for comparison. (c) Trematode cercaria encyst within the skin (arrows) and body
cavities of amphibians serving as a secondary host and may be easily seen grossly.
Histologically, the organisms are found in thin-walled cysts (inset).
isolated geographical areas so far (Davis et al. 2007). Still other organisms,
such as the watermolds Saprolegnia, may be beneficial (e.g. by facilitating
decomposition of dead eggs) but also have the potential to be opportunistic
pathogens of amphibians at any life stage (Converse and Green 2005a, 2005b;
Green and Converse 2005a, 2005b).
26.2.2 Parasitic diseases

As with any species, parasites are commonly found on and in amphibians
(Figure 26.1c). External parasites include leaches, anchorworms, and mites,
whereas internal parasites include various trematodes, cestodes, nematodes,
and protozoa (Converse and Green 2005a, 2005b; Green and Converse 2005b;
Wright and Whitaker 2001). Many species of helminthes (trematodes, ces-
todes, nematodes) have been documented in amphibians, and often they are
considered incidental (Miller et al. 2004), but their presence may be an indi-
cator of stress or aquatic food-web restructuring related to human land use
(Johnson and Lunde 2005; Gray et al. 2007b). Likewise, many protozoans
(e.g. myxozoa) are often considered incidental findings but their numbers may
increase when amphibians are stressed, and they may potentially contribute to
morbidity.
26.2.3 Toxins
Contaminants in the environment may kill larvae or post-metamorphs (Relyea
2005, 2009), and may have non-lethal impacts including reducing growth,
impacting metamorphosis, disrupting gonadal development and secondary sex
characteristics, or causing musculoskeletal, skin, and visceral malformations
(Boone and Bridges 2003; Davidson et al. 2007; Ouellet et al. 1997; Storrs
and Semlitsch 2008). Often these changes are not detected by external exami-
nations until metamorphosis is complete or until the animals attain a size for
reproduction. Amphibians are often considered sentinels or bio-indicators of
environmental quality because they are sensitive to toxins and many species
have the potential to be exposed to stressors in aquatic and terrestrial systems
due to their typical biphasic life cycle (Blaustein and Wake 1995).
26.3 Disease monitoring: detection and diagnosis

26.3.1 Disease surveillance
Recently, the World Animal Health Organization (the OIE; www.oie.int/eng/
en_index.htm) included two amphibian diseases (chytridiomycosis and rana-
viral disease) on their listing of reportable diseases. The OIE listing provides
the impetus for disease surveillance and required testing of amphibians prior
to transport among states or between nations. The need for required testing of
amphibians for pathogens has been expressed by several researchers (Gray et al.
2007a; Griffiths and Pavajeau 2008; Picco and Collins 2008). Historically,
pre-transport pathogen testing and health certification for amphibians has
been essentially non-existent, unlike for domestic livestock and pets and
some wild mammals (e.g. cervids). The OIE has established guidelines for
surveillance and requirements necessary for countries to declare Ranavirus-
free status (www.oie.int/eng/norms/fcode/en_chapitre_2.4.2.htm#rubrique_
ranavirus) and B. dendrobatidis-free status (www.oie.int/eng/normes/fcode/
en_chapitre_2.4.1.htm#rubrique_batrachochytrium_dendrobatidis). The
OIE-approved methods for conducting surveillance and diagnosis of infection
are in development. In the meantime, guidelines from the 2006 OIE Manual of
Diagnostic Tests for Aquatic Animals (www.oie.int/ eng/normes/fmanual/A_
summry.htm) and from the fisheries industry (USFWS and AFS-FHS 2005)
can be helpful for general monitoring of amphibian population health. In gen-
eral, the criteria of a population health assessment should include (1) determin-
ation of the status and trends of amphibian pathogens, (2) determination of
the risk of disease for threatened or endangered amphibians, (3) investigation
of unexplained population declines, (4) evaluation of populations following a
morbidity or mortality event, (5) detection of pathogens in non-indigenous spe-
cies, (6) evaluation of a site or population prior to translocation, (7) evaluation
of sympatric amphibians prior to release of captive-raised animals, and (8) the
potential for amphibians and their diseases to “piggy back” with fish translocation.
Disease testing should not focus on one pathogen. For surveillance programs,
we recommend that animals are tested for infection by at least the OIE patho-
gens: ranaviruses and B. dendrobatidis. For diagnosis of morbid or dead individ-
uals, we encourage a full diagnostic work-up (i.e. necropsy, histology, bacterial
culture, virus testing, and parasite testing) to attempt to identify all etiologic
agents. It is important to note that simultaneous infection by multiple patho-
gens is possible. Further, histological examination of organs often is required to
determine which of the pathogens identified are causing the changes responsible
for the diseased state (Miller et al. 2008, 2009). Histological examination is also
important in discovering introduced pathogens or pathogens that have not been
described previously (Longcore et al. 1999; Davis et al. 2007).
Population health assessments can include non-lethal or lethal collection of
tissue samples from individual amphibians (Greer and Collins 2007), and col-
lection of environmental samples (e.g. water, soil; Walker et al. 2007). Ideally,
we recommend that tissue samples are collected from all species in a community
and from pre- and post-metamorphic life stages. Amphibian species differ in sus-
ceptibility to pathogens, and some age classes may serve as a reservoir (e.g. larval
for B. dendrobatidis and adults for Ranavirus; Daszak et al. 1999; Brunner et al.
2004; Schock et al. 2008). Further, some infectious diseases become evident
only after the post-metamorph has overwintered (e.g. Lucke tumor herpesvirus,
Amphibiothecum (formerly Dermosporidium) penneri). The lack of gross signs
of disease also does not imply healthy populations. We and others have found
tadpoles with no signs of illness but that are infected with ranaviruses (Gray et al.
2007a; Harp and Petranka 2006; Miller et al. 2009). Laboratory studies have
demonstrated that amphibian pathogen infection and mortality rates frequently
track each other (e.g. Brunner et al. 2007); thus, high prevalence in a population
could signal that a die-off is imminent.
In some cases, it may not be possible to collect sufficient tissue for dis-
ease testing. For example, small amphibians (e.g. Bufo larvae) may not have
adequate tissue for tests, especially for toxicological analysis. Also, non-lethal
testing may be required because a species is listed as a conservation concern.
We found that testing for Ranavirus from tail clips results in about 20% false-
negatives (D. L. Miller and M. J. Gray, unpublished results). In cases when a
small amount of tissue is collected, multiple individuals within a species could
be pooled to acquire sufficient tissue for testing. If contaminants are suspected
as the cause of a die-off, we also recommend collecting and testing water and
sediment at the amphibian breeding site.
Monitoring for malformations can be challenging, because typically mal-
formed individuals have low survival. Although amphibian malformations
have been documented for many years (Rostand, 1958), an increase in mal-
formation rates occurred in the late twentieth century (Johnson and Lunde
2005). Generally, malformation studies have targeted recently metamorphosed
amphibians (Meteyer et al. 2000), because metamorphs with prominent abnor-
malities are quickly removed from the population by predation or starvation.
Additionally, the bony skeleton of metamorphosed amphibians is more con-
ducive for radiographically visualizing deformities compared to the cartilagin-
ous skeleton of larvae. However, monitoring of larval abnormalities is needed
because it is likely that some abnormalities prevent metamorphosis, thus are not
detected in post-metamorphic cohorts.
Finally, comprehensive disease surveillance should include captive amphib-
ians in zoological and ranaculture facilities, because disease transmission can
occur between captive and free-ranging populations. Maintenance of health
in zoological facilities is especially important for rare species or in captive
breeding populations intended for release. High densities in ranaculture facil-
ities, pet shops, and stores that sell amphibians (e.g. Ambystoma tigrinum) for
fishing bait can be cauldrons for disease transmission and pathogen evolution
(Picco and Collins 2008). Ranaviruses isolated from ranaculture facilities and
bait shops appear to be more virulent than wild strains (Majji et al. 2006;
Storfer et al. 2007). This emphasizes the importance for disease monitoring at
facilities with captive amphibians. In the event of a die-off in a captive facil-
ity, freshly dead animals should be submitted for diagnostic evaluation. Live
animals that are infected should be euthanized or treated if a treatment exists

(discussed in section 26.4.3), and facilities decontaminated with bleach or an
equivalent disinfectant (discussed in section 26.4.2).
26.3.2 Sample size

Determination of statistically appropriate sample sizes for amphibian disease
surveillance remains in its infancy. Although not established for amphibians,
health assessment of fish is based on the minimum assumed pathogen preva-
lence level (APPL). The commonly used APPLs in aquatic health investiga-
tions are 2, 5, and 10% (Lavilla-Pitogo and de la Pena 2004; USFWS and
AFS-FHS 2005). The APPL is used with an estimate of amphibian population
size to determine the number of individuals that should be tested to have 95%
confidence in pathogen detection. If it is assumed that APPL is 10%, required
sample size ranges from 20 to 30 depending on the size of the amphibian popu-
lation (Table 26.4). Required sample size increases with decreasing APPL
(Table 26.4). Unpublished findings of the US Geological Survey National
Wildlife Health Center suggest that APPL for ranaviruses, B. dendrobatidis,
and alveolates is 10% or less (Table 26.5). In Tennessee, USA, health moni-
toring of two common anuran species inhabiting farm ponds revealed 29%
prevalence for Ranavirus, 0% for B. dendrobatidis and alveolates, and 43% for
parasites (Miller et al., 2009). Ranavirus prevalence in plethodontid salaman-
ders in the southern Appalachian Mountains can range from 3 to 81% depend-
ing on the watershed (M. J. Gray and D. L. Miller, unpublished results). Thus,
we recommend that biologists determine required sample sizes for amphibian
disease monitoring using either 5 or 10% APPL (Table 26.4).
26.3.3 Sample collection and shipment

Sample collection may include whole live animals, dead animals, sections of
tissues, swabs of lesions or orifices, environmental samples, or sympatric spe-
cies. It is important to wear disposable gloves when handling amphibians and to
change gloves between animals. This is necessary to prevent disease transmis-
sion between amphibians and to protect biologists from zoonotic diseases (dis-
cussed in section 26.4). Gutleb et al. (2001) and Cashins et al. (2008) reported
that disposable gloves (especially latex gloves) may be toxic to amphibian larvae.
Therefore, when handling amphibians, biologists and researchers should use
disposable vinyl gloves that have been rinsed with distilled or sterilized water
(Cashins et al. 2008).
Mortality events involving all amphibian species should be investigated, even
if it is not part of a disease surveillance program. There is a paucity of information
Table 26.4 Sample size (i.e. number of amphibians) to assure

95% confidence in detection of pathogens in a population
(modified from USFWS and AFS-FHS 2005).
Estimated population size Number of amphibians for
5% APPL 10% APPL
100 45 23
500 55 26
2000 60 27
10 000 60 30
APPL, assumed pathogen prevalence level.
Table 26.5 Previously unreported low disease prevalence in free-ranging amphibian

populations in the USA (US Geological Survey National Wildlife Health Center).
Pathogen Host species Life US Sample Prevalence Test Case
stage state size (number method number
positive)
Ranavirus R. catesbeiana L OR 15 7% (1) Culture 44276

Pseudacris L WY 11 18% (2) Histology 4779
maculata
Perkinsus-like R. catesbeiana L OR 12 8% (1) Histology 44276
organism R. sphenocephala L FL 15 27% (4) Histology 4864
R. sphenocephala L LA 12 8% (1) Histology 18626
R. sphenocephala RM MD 14 21% (3) Histology 18761
R. sphenocephala L MS 11 9% (1) Histology 18642
Ichthyophonus R. sphenocephala L FL 19 11% (2) Histology 4864
R. grylio L FL 16 19% (3) Histology 4864
R. sphenocephala L LA 12 8% (1) Histology 18626
R. clamitans L ME 16 31% (5) Histology 4824
L, larvae; RM, recently metamorphosed.
on the occurrence of pathogen-related die-offs in amphibian populations. The

majority of samples submitted to diagnostic laboratories are from biologists that
encountered a dead or morbid amphibian during other work activities. Morbid
or freshly dead amphibians are preferred, because amphibians decompose rap-
idly. Decomposed carcasses are not suitable for cultures, histology, and parasito-
logical examinations, but may have limited diagnostic usefulness for molecular
tests that detect pathogens and for toxicological analyses. In general, we recom-
mend that amphibians be collected live or within 24 h of death. Mummified
(i.e. desiccated) carcasses with dry, leathery, and stiff digits or limbs usually have
limited diagnostic usefulness.
Dead amphibians should be collected, put individually in plastic bags (e.g.
Nasco Whirl-Pak® bags), and placed on ice for transport. Live amphibians can be
placed in separate plastic containers and humanely euthanized (Baer 2006) via
transdermal exposure for 10 min to tricaine methanesulfonate (100–250 mg/L)
or benzocaine hydrochloride ( 250 mg/L or 20% benzocaine over-the-counter
gel; Oragel, Del Paharmaceuticals, Uniondale, New York, USA) after returning
from the field. It is important that amphibians are bagged separately to prevent
cross-contamination of samples. Biologists that are experienced in blood collec-
tion may collect blood antemortem from the ventral vein in adult anurans or tail
vein in salamanders, or collect blood antemortem or immediately postmortem
from the heart of larvae or adults (Wright and Whitaker 2001). Blood can be
tested for various biochemical parameters and examined for cellular compos-
ition, blood parasites, and viral inclusions (discussed in section 26.3.4).
We recommend that half of the individuals collected are frozen immediately
for cultures and molecular tests. Samples can be frozen in a standard 20°C
freezer if stored for short duration (
1 month); otherwise, samples should be
stored in a 80°C freezer. The other half of samples should be promptly fixed
in 75% ethanol or 10% neutral buffered formalin for histology. For the first
2–4 days of fixation, the volume of fixative should be 10 times the volume of the
animals. After this initial fixation, carcasses can be stored in a smaller volume of
fixative that is sufficient to cover the tissues. The body cavity of amphibians that
are more than 1 g in body mass should be cut along the ventral midline prior to
immersion in fixative to assure rapid fixation of internal organs. Body cavities of
frozen individuals should not be opened.
Special processing is required for amphibians with skin, digital, limb, head,
or vertebral abnormalities. Whenever possible, amphibians with suspected
malformations should be submitted alive for examinations. Dead individuals
should be promptly frozen until time exists to properly fi x individuals. Fixation
can be done with ethanol or formalin but should be done in a pan so that car-
casses can be positioned on a flat surface with limbs and digits extended from
the body during fi xation. Positioning amphibians in the standard museum
configuration is ideal. Digits and limbs may be taped in position prior to fi x-
ation. Amphibians should be covered with fi xative and additional fi xative
added if a significant amount evaporates. Placing a cover over the pan will help
reduce evaporation. After 2–4 days of fi xation, the carcass and limbs will be
hardened in position and may be stored in a smaller volume of fi xative. The
positioning described above is necessary for radiographic examination of the

malformation.
Given that it often is not possible to obtain collection permits for threat-
ened amphibians, alternative sampling may be necessary. Two alternatives
are (1) capture–release studies and (2) collection of “sentinel” sympatric
amphibian species. Capture–release studies can be used to collect swabs of
external tissues, blood, or fecal samples. Swabs appear to be a reliable tech-
nique to test for B. dendrobatidis (Kriger et al. 2006); however for ranaviruses,
false-positive and -negative test results are greater than for tail clips and both
of these non-lethal techniques have more false results than testing internal
organs (D. L. Miller and M. J. Gray, unpublished results). Swabs are typ-
ically performed in the oral then cloacal regions, and the swab stored in its
packaging container or a microcentrifuge tube. Swabs should be put on ice
and frozen similar to tissues. An accepted protocol for swabbing amphibians
for B. dendrobatidis testing using PCR (discussed in section 26.3.4) has been
reported by Brem et al. (2007) and can be found at www.amphibianark.org/
chytrid.htm. Briefly, the amphibian should be gently but firmly swabbed in
a sweeping motion five times at each of the following five locations (for a
total of 25 times): rear feet (toe webbing), inner thighs, and ventral abdomen.
Occasionally, modifications to this technique are necessary for salamanders
(Brem et al. 2007). Swabs for B. dendrobatidis testing by PCR may be stored
in 70% ethanol. Collecting common sympatric species for health assessment
can provide insight into the presence of amphibian pathogens at a site, but
does not allow for direct health assessment of the species of concern, which
may differ in susceptibility.
Shipment of live, freshly dead, or frozen specimens must be via an overnight
courier and according to the specific courier guidelines. For fixed specimens,
overnight shipment is unnecessary. General guidelines for shipment include tri-
ple packaging and labeling each layer of packaging with a waterproof writing
utensil. Commonly, the first package layer is a specimen in a Whirl-Pak® bag.
The second layer is a larger sealable plastic bag in which multiple specimens
are placed. If the first package layer contains liquid (e.g. ethanol), paper towel
should be added to the second package to absorb any liquid if a spill occurs. The
third package typically is a padded box or shipping cooler. For frozen specimens,
adequate ice packs or dry ice should be added around the secondary package.
It is vital that the package contains a detailed list of all contents, a description
of requested services, and the contact information of the shipper. The tracking
information should be provided to the recipient prior to package arrival.
26.3.4 Diagnostics
Several tools are available for diagnosing amphibian diseases but generally
require some level of specialized expertise to perform. Commonly used diagnos-
tic tools include necropsy, histology, cytology, bacterial culture, virus isolation,
fecal floatation, electron microscopy, molecular modalities, and radiology. Most
of these tests can be performed on samples collected from dead or live amphib-
ians. Fresh or frozen tissues can be used for most tests, and are necessary for virus
isolation. Frozen tissues are not appropriate for histology or cytology; rather,
preserved tissues are used. Although formalin-fixed specimens are preferred for
histological examination, ethanol-fixed specimens may also be used. Blood can
be used for cell counts to assess immune function and to look for inclusion bod-
ies that can be diagnostic for certain pathogens. Blood also may be tested for the
presence of antibody response to various diseases. Examples of laboratories that
currently test for amphibian diseases in Australia, New Zealand, Europe, and
the USA include Australian Animal Health Laboratories (AAHL), Geelong,
Victoria, Australia (www.csiro.au/places/aahl.html), Gribbles Veterinary
Pathology, Australia and New Zealand (www.gribblesvets.com/), Exomed,
Berlin, Germany (www.exomed.de/), Hohenheim University (R. Marschang),
Stuttgart, Germany, Wildlife Epidemiology, Zoological Society of London
(ZSL), London, UK, The University of Georgia Veterinary Diagnostic and
Investigational Laboratory, Tifton, GA, USA (www.vet.uga.edu/dlab/tifton/
index.php), University of Florida (J. Wellehan), College of Veterinary Medicine,
Gainesville, FL, USA, and National Wildlife Health Center, Madison, WI,
USA (www.nwhc.usgs.gov/).
There are advantages and disadvantages to the various tests available
(Table 26.6). Necropsy allows for identification and documentation of exter-
nal and internal gross changes. Histological and cytological examination
allows for identification of changes at the cellular level and is generally neces-
sary to document disease versus infection. Virus isolation is the process of
culturing a virus which is necessary to determine the presence of live virus
and to perform some molecular tests used in identifying viral species (e.g.
sodium dodecyl sulphate/polyacrylamide-gel electrophoresis (SDS/PAGE)
and restriction fragment length polymorphism (RFLP)). One caveat is that
some viruses are difficult to culture, thus infection cannot be ruled out based
solely on negative isolation results. Electron microscopy is used for identify-
ing key features of parasites or other infectious agents (e.g. B. dendrobatidis,
Ranavirus, herpesvirus), documenting intracellular changes or changes to the
cellular surface, and confirmation of cultured virus. Electron microscopy can
be performed on fresh, fi xed, or paraffin-embedded tissues. Radiology allows
Table 26.6 Advantage and disadvantages of diagnostics tests for amphibian pathogens
given the type of sample.
Sample type Tests Pathogen Advantages Disadvantages
Live animal Necropsy, histology, Viruses, bacteria, Observe Difficulty in

cytology, fungi, parasites, behavior, least transport,
hematology, virus toxins chance of stressful to
isolation, bacterial contamination, animal
culture, toxicological blood collection
analysis, is possible
parasitology, PCR
Fresh tissue Necropsy if whole Viruses, bacteria, Can isolate live If advanced
(including animal, histology, fungi, parasites, pathogens postmortem
whole dead cytology, virus toxins autolysis, then of
organisms) isolation, bacterial limited value
culture, toxicological
analysis,
arasitology, PCR
Frozen tissue Virus isolation, Viruses, bacteria, Can isolate live Limited value for
bacterial culture, fungi, parasites, pathogens histology
PCR toxins (freeze artifact)
Swab Virus isolation, Viruses, bacteria, Non-lethal, may False positives
bacterial culture, fungi detect shedders and negatives are
PCR possible
Fixed tissue Histology, PCR Parasites, Can see cellular Cannot isolate
bacteria, fungi, changes due to live pathogens
viruses, disease
for documentation of bone structure or the presence of foreign bodies, includ-

ing certain parasites (e.g. Ribeiroia metacecariae).
Molecular testing is becoming increasingly popular and affordable for disease
diagnostics. It is especially useful for endangered species, as non-lethal sampling
can yield accurate results. Specifically, it can be performed on fresh, fixed or
paraffin-embedded tissues, swabs, blood, and feces. For testing via PCR, one
caveat is that a positive PCR result only confirms the presence of the pathogen
whether it is dead or alive. Thus, it is important to perform supportive tests (e.g.
virus isolation, histological examination) to differentiate between infection and
disease. Either conventional or real-time PCR (qPCR) may be used, depending
on the availability of known primer sequences and the purpose of the test. For
quantifying viral presence and infection (Yuan et al. 2006; Storfer et al. 2007),
qPCR is most ideal (Brunner et al. 2005; Pallister et al. 2007). However, if
sequencing is desired, which is often necessary to identify the species of a patho-

gen, conventional PCR is necessary.
There are three standard methods for characterizing amphibian malforma-
tions: (1) dissection, (2) radiography, and (3) clearing and staining. Dissection
of a carcass is tedious, as it usually requires careful removal of muscles from
limbs, head, and axial skeleton. Dermestid beetles (Dermestes maculatus) might
be used to remove muscles and soft tissues from an amphibian carcass, but re-
assembly of bones of vertebrae, limbs and digits can be very challenging and
time-consuming. Radiography is the preferred diagnostic method for investi-
gating and documenting abnormalities of the skeleton (Meteyer et al. 2000).
A major limitation of radiography is that cartilage is invisible; hence, detection
of abnormalities of cartilage is not possible. Instead, the clearing-and-staining
method commonly used in teratological studies of embryos is recommended
when both cartilage and bone need to be examined (Kimmel and Trammell
1981; Schotthoefer et al. 2003). This method involves “clearing” the skin, mus-
cles, and viscera by immersion in potassium hydroxide. The bones are stained
red with Alizarin Red, and cartilage is stained blue using Alcian Blue stain.
Clearing and staining is the preferred method to evaluate larval amphibian skel-
etal abnormalities.
Regardless of the diagnostic tests employed, interpretation of the test
results must be done with caution and knowledge of the amphibian patho-
gen that is being tested (Table 26.6). The type of sample must be consid-
ered when targeting a pathogen. For example, infection of Ranavirus is best
diagnosed from internal organs otherwise environmental contamination (e.g.
water or soil) cannot be ruled out. Nonetheless, documentation of ranavi-
ruses from skin surfaces does provide evidence of environmental exposure.
In contrast, B. dendrobatidis is commonly tested from skin surfaces in adults
or mouth parts in larvae, because this pathogen infects only keratinized tis-
sue (Kriger et al. 2006; Skerratt et al. 2008). However, histology is generally
required to distinguish between B. dendrobatidis exposure and infection when
gross lesions are not observed. In contrast, some pathogens (e.g. alveolates,
Ichthyophonus spp., Ribeiroia ondatrae) may not be identifiable from external
swab preparations and often specialized techniques are required (e.g. clearing
or radiography for R. ondatrae). In addition, one must keep in mind that there
is a difference between malformations and deformities. Malformations are
those abnormalities that arise during growth and development (organogen-
esis) in which the organ or structure fails to form normally. A deformity is an
abnormality that naturally occurs to a normal organ or structure, such as an
amputation or wound. It is often difficult to determine, even in radiographs,

whether an abnormality is a malformation or a deformity.
26.4 Biosecurity: preventing disease transmission

Lethal infectious diseases of amphibians may be endemic and emerge in response
to stressors, whether anthropogenic or natural (Carey et al. 2003). Disease emer-
gence also may occur through geographical transport of pathogens (Jancovich
et al. 2005; Storfer et al. 2007). Ranaviruses, B. dendrobatidis, the alveolate
organism, and Ichthyophonus spp. are well established in many regions of the
world; however, it is likely that some amphibian species have never been exposed
to these agents. Further, in areas with multiple endemic pathogen strains or
species (e.g. ranaviruses), slight variations in genetic coding can increase viru-
lence (Williams et al. 2005). Thus, an endemic strain may function as a novel
pathogen to amphibian populations outside the region where the pathogen
evolved. This may be especially true with amphibian pathogens given the lim-
ited mobility of their host. Hence, prevention of the spread of endemic diseases
to naïve populations or species remains a high conservation priority. Health
examinations of amphibian populations and good biosecurity methods need
to be employed because often little is known about the life cycles of infectious
diseases, modes of transmission, and the persistence of the pathogen within and
outside the amphibian host.
Preventing mechanical transmission of pathogens and contaminants from
one location to another by equipment, supplies and people is the purpose of
biosecurity. Biosecurity involves three equally important aspects: (1) safety of
the humans and animals in the area, (2) decontamination or disinfection of field
equipment, and (3) restriction on transporting amphibians among watersheds.
26.4.1 Human and animal safety

Whenever sampling amphibians for disease, the priority must be personal safety
and health. For standard monitoring, biologists should wear gloves and water-
proof footwear that can be easily disinfected (e.g. rubber boots). If a die-off is
observed, it is important to note whether other vertebrates (e.g. birds, fish) are
dead or appear morbid. If so, there is a greater chance the animal deaths are due
to toxins, which may present a significant human health risk. In cases with a
multiple wildlife taxa die-off, field personnel should leave the site immediately
without collecting specimens and notify the nearest public health department
and wildlife agency. Persons leaving a multiple-taxa mortality site should wash
and disinfect boots, waders, nets, and field equipment and change clothes before
entering a vehicle and leaving the site (discussed in section 26.4.2).
Few infectious diseases of amphibians are contagious to humans. Potential
zoonotic diseases that may be carried by amphibians include certain Salmonella
spp., Yersinia spp., Chlamydophila spp. (formerly Chlamydia)), and some tox-
in-producing mycobacteria (e.g. Mycobacterium liflandii) that can cause skin
ulceration. In addition, Gray et al. (2007c) demonstrated that Rana catesbe-
iana metamorphs were suitable hosts for the human pathogen Escherichia coli
O157:H7. We also demonstrated recently that tadpoles could maintain this
pathogen in aquatic mesocosms (M. J. Gray and D. L. Miller, unpublished
results). Thus, disposable gloves should be worn whenever handling amphib-
ians, and hands washed thoroughly with soap and warm water after removing
gloves. In the field, hands can be soaked in a 2% clorhexidine solution for 1 min
or disposable antibacterial wipes used. Avoid exposure of surface water to soaps
and disinfectants, as they may negatively affect local flora and fauna. Clothing
that becomes stained with feces or skin secretions should be removed as soon as
possible and washed in color-safe bleach.
The skin secretions of many amphibians contain potent irritants and toxins.
For example, newts (Salamandridae), toads (Bufonidae), and poison-dart frogs
(Dendrobatidae) exude toxic skin secretions. Skin secretions of certain newts (e.g.
Taricha) may cause temporary blindness lasting several hours if the secretions get
into the eyes. The parotoid gland secretions of giant toads (Bufo marinus), if
ingested, can rapidly cause heart malfunction in humans and animals. When
handling toads, it is best to avoid touching the parotoid glands. After handling
amphibians, avoid touching your eyes or mouth prior to washing hands.
26.4.2 Washing and disinfecting equipment

Cleaning equipment and waders is recommended when leaving any amphibian
breeding site, whether it is known that pathogens are present or not (see also
www.nwhc.usgs.gov/). Cleaning is a three-step process: (1) washing with a soap
or detergent, (2) rinsing thoroughly with clean water, and (3) disinfecting of the
objects via a chemical disinfectant. Common soaps or detergents are not disin-
fectants but are useful in removing sediments and vegetation. Biodegradable
soaps should be used in the field and not discarded into surface waters, as many
are toxic to amphibians, fish, and invertebrates. Chemical disinfectants need to
remain in contact with cleaned and rinsed surfaces for several minutes to kill
microorganisms.
Common disinfectants used are chlorhexidine and sodium hypochlorite
(bleach). Bleach is often preferred because it is cost effective, easily obtained,
and effective against most bacteria and many viruses. The US Fish and Wildlife
Service and American Fisheries Society – Fish Health Section (USFWS and
AFS-FHS) (2005) recommend 10 min of exposure of a 0.05% bleach solution
(i.e. 28.4 g of 6.15% sodium hypochlorite in 3.8 L of clean water) for disinfection
of field equipment and surfaces for B. dendrobatidis, and, although not conclu-
sive, a 0.5% solution (i.e. 312 g of 6.15% sodium hypochlorite per 3.8 L of water)
is recommended to destroy myxosporeans. However, bleach is not very effective
at inactivating Ranavirus, and requires at least a 3% concentration (Bryan et al.
2009). It should be noted that this concentration can be toxic to amphibians. In
contrast, chlorhexidine used at a dosage that is safe for amphibians (0.75% for
a 1 min exposure) has been shown to inactivate Ranavirus (Bryan et al. 2009).
Further, it is important to keep in mind that the shelf-life of bleach solutions is
influenced by exposure to light, air, and organic material, and solutions should
be discarded after 5–7 days. After disinfection, equipment may be allowed to air
dry or rinsed with fresh, clean water. Alternatively, if carrying large quantities of
water is not possible because multiple fields sites are to be visited, surface water
from the subsequent site (i.e. where the equipment will be used next) can serve as
the rinse water. If mountain systems with stream watersheds are sampled, we rec-
ommend that researchers begin sampling at higher elevations and work towards
lower sites. If a disease agent is present at higher elevations, it is likely to be at
lower elevations due to downstream transmission. Hence, if accidental transmis-
sion occurs during travel on fomites, it is less likely to be a novel introduction.
26.4.3 Movement of animals and disease management

Introducing captive-raised or moving wild amphibians into new locations may
be necessary because of population declines or extirpations. It is important to
understand the initial cause of the die-off to ensure the factor no longer exists.
In the case of diseases, environmental testing for the etiologic agent should be
done before reintroductions or translocations. For pathogens, existing amphib-
ian species also should be tested to ensure they are not functioning as a reservoir.
It hinders conservation efforts to release species with high susceptibility if the
pathogen remains at a site. Simultaneously, testing of the source population
should be performed prior to reintroduction to avoid introduction of pathogens
into the wild. Non-lethal testing as described previously generally can be used.
Alternatively, in the case of translocations, lethal testing of common closely
related resident species from the donor environment can provide some assurance
that the target species is not infected.
Amphibians (dead or alive) from a mortality site should be considered con-
tagious specimens. Morbid animals and carcasses should not be released or
discarded at the same or other sites because this may facilitate the spread or
persistence of infectious diseases. Dead amphibians that are not used for testing
should be placed in double-layered plastic trash bags and disposed by burial or
incineration. Removal of carcasses is a good strategy to help thwart the spread
of infectious diseases.
While some serious infectious diseases of amphibians (e.g. B. dendrobatidis,
nematode lungworms (Rhabdias spp.)) are readily treated and eliminated
from captive populations, some important infectious diseases have no known
treatments (e.g. ranaviruses, alveolates) or no practical treatment in the wild.
Treatment of any disease varies by the pathogen involved as well as the host.
Some pathogens are resistant to many treatments (e.g. antibiotic-resistant bac-
teria) and some hosts may be sensitive to a particular treatment (e.g. Methylene
Blue may be toxic to tadpoles at concentrations over 2 mg/ml). Generally, it is
best to contact a veterinarian with experience in amphibians for proper treatment
of disease. However, some treatments (i.e. elevated temperature for B. dendroba-
tidis or dermosporidium, sea salt or Methylene Blue for Saprolegnia, chlorhexi-
dine for bacteria and Ranavirus) may be attempted by the non-veterinarian and
treatment guidelines can be found in Wright and Whitaker (2001) and Poole
(2008). As a general rule, treatment for disease is only applicable to captive
environments; however, it can be a valuable conservation tool for amphibians
slated for release.
In the event that animals destined for release test positive for a treatable dis-
ease, the animal and any others that may have been exposed should be treated.
Following treatment, a minimum of two negative test results with 1 month
between tests should be obtained. If the animal does not test negative, the treat-
ment should be repeated. Only animals that test negative should be released into
the wild. In addition, if one animal in a group of 10 housed together tests posi-
tive for a pathogen, all of the animals should be treated, regardless of individual
test results. Current guidelines for treatment and release have been established
by the Association of Zoos and Aquariums (Poole 2008). Testing at the appro-
priate life stage for the host and disease agent is important.
26.5 Conclusions
Amphibians are declining globally and emerging infectious diseases are one of
the causes. Natural resource agencies and conservation organizations should
consider establishing amphibian disease surveillance programs that moni-
tor populations for at least the two pathogens linked to catastrophic die-offs:
Ranavirus and B. dendrobatidis. Further, the OIE has listed these pathogens as
notifiable diseases, mandating that Ranavirus- and B. dendrobatidis-free status

be verified prior to movement of amphibians for commerce. Herein, we have
provided guidance on collection, storing, and shipping protocol of amphibians
to diagnostic laboratories for disease testing. We encourage readers to use the
Internet to locate a wildlife diagnostic laboratory in your area.
Given that pathogens can cause significant mortality that have trickle-down
effects on ecosystem processes (Whiles et al. 2006), biologists must be prudent
to decontaminate field equipment and footwear when moving among amphib-
ian breeding sites. We also recommend that natural resource agencies consider
implementing wildlife laws that prevent the use of amphibians as fishing bait.
Transmission of Ranavirus in western North America has been attributed to
the movement and sale of A. tigrinum larvae (Storfer et al. 2007; Picco and
Collins 2008). We also encourage natural resource agencies to develop public
educational brochures on the threat of amphibian diseases and the benefits of
decontaminating recreational gear when leaving watersheds. Finally, prudent
land stewardship undoubtedly reduces the likelihood of disease emergence by
decreasing the effect of anthropogenic stressors. We encourage support of exist-
ing or development of new conservation programs that help landowners estab-
lish undisturbed buffers around amphibian breeding sites.
26.6 References
Baer, C. K. (2006). Guidelines for Euthanasia of Nondomestic Animals. American
Association of Zoo Veterinarians, Yulee, FL.
Blaustein, A. R. and Wake, D. B. (1995). The puzzle of declining amphibian populations.
Scientific American, 272, 52–7.
Boone, M. D. and Bridges, C. M. (2003). Effects of carbaryl on green frog (Rana clami-
tans) tadpoles: timing of exposure versus multiple exposures. Environmental Toxicology
and Chemistry, 22, 2695–2702.
Brem, F., Mendelson, III, J. R., and Lips, K. R. (2007). Field-Sampling Protocol for
Batrachochytrium dendrobatidis from Living Amphibians, using Alcohol Preserved
Swabs, version 1.0 (18 July 2007). www.amphibianark.org/pdf/Field%20sampling%
20protocol%20for%20amphibian%20chytrid%20fungi%201.0.pdf. Conservation
International, Arlington, VA.
Brunner, J. L., Schock, D. M., Davidson, E. W., and Collins, J. P. (2004). Intraspecific reser-
voirs: complex life history and the persistence of a lethal ranavirus. Ecology, 85, 560–6.
Brunner, J. L., Richards, K., and Collins, J. P. (2005). Dose and host characteristics influ-
ence virulence of ranavirus infections. Oecologia, 144, 399–406.
Brunner, J. L., Schock, D. M., and Collins, J. P. (2007). Transmission dynamics of the amphib-
ian ranavirus Ambystoma tigrinum virus. Diseases of Aquatic Organisms, 77, 87–95.
Bryan, L. K., Baldwin, C. A., Gray, M. J., and Miller, D. L. (2009). Efficacy of select
disinfectants at inactivating Ranavirus. Diseases of Aquatic Organisms, 84, 89–94.
Carey, C., Bradford, D. F., Brunner, J. L., Collins, J. P., Davidson, E. W., Longcore, J. E.,
Ouellet, M., Pessier, A. P., and Schock, D. M. (2003). Biotic factors in amphibian
declines. In G. Linder, S. K. Krest, and D. W. Sparling (eds), Amphibian Declines: an
Integrated Analysis of Multiple StressorEffects, pp. 153–208. Society of Environmental
Toxicology and Chemistry, Pensacola, FL.
Cashins, S. D., Alford, R. A., and Skerratt, L. F. (2008). Lethal effect of latex, nitrile, and
vinyl gloves on tadpoles. Herpetological Review, 39, 298–301.
Converse, K. A. and Green, D. E. (2005a). Diseases of tadpoles. In S. K. Majumdar,
J. Huffman, F. J. Brenner, and A. I. Panah (eds), Wildlife Diseases: Landscape
Epidemiology, Spatial Distribution, and Utilization of Remote Sensing Technology,
pp. 72–88. Pennsylvania Academy of Science, Easton, PA.
Converse, K. A. and Green, D. E. (2005b). Diseases of salamanders. In S. K. Majumdar,
Cunningham, A. A., Hyatt, A. D., Russell, P., and Bennett, P. M. (2007). Experimental
transmission of a ranavirus disease of common toads (Bufo bufo) to common frogs
(Rana temporaria). Epidemiology and Infection, 135, 1213–16.
Daszak, P., Berger, L., Cunningham, A. A., Hyatt, A. D., Green, D. E., and Speare, R.
(1999). Emerging infectious diseases and amphibian population declined. Emerging
Infectious Diseases, 5, 735–48.
Davidson, C., Benard, M. F., Shaffer, H. B., Parker, J. M., O’Leary, C., Conlon, J. M.,
and Rollins-Smith, L. A. (2007). Effects of chytrid and carbaryl exposure on survival,
growth and skin peptide defenses in foothill yellow-legged frogs. Environmental Science
and Technology, 41, 1771–6.
Davis, A. K., Yabsley, M. J., Keel, M. K., and Maerz, J. C. (2007). Discovery of a novel
alveolate pathogen affecting southern leopard frogs in Georgia: description of the dis-
ease and host effects. EcoHealth, 4, 310–17.
Forson, D. D. and Storfer, A. (2006). Atrazine increases ranavirus susceptibility in the
tiger salamander, Ambystoma tigrinum. Ecological Applications, 16, 2325–32.
Gray, M. J., Miller, D. L., Schmutzer, A. C., and Baldwin, C. A. (2007a). Frog virus 3
prevalence in tadpole populations inhabiting cattle-access and non-access wetlands in
Tennessee, U.S.A. Diseases of Aquatic Organisms, 77, 97–103.
Gray, M. J., Smith, L. M., Miller, D. L., and Bursey, C. R. (2007b). Influence of agricul-
tural land use on trematode occurrence in southern Great Plains amphibians, U.S.A.
Herpetological Conservation and Biology, 2, 23–8.
Gray, M. J., Rajeev, S., Miller, D. L., Schmutzer, A. C., Burton, E. C., Rogers, E. D.,
and Hickling, G. J. (2007c). Preliminary evidence that American bullfrogs (Rana cat-
esbeiana) are suitable hosts for Escherichia coli O157:H7. Applied and Environmental
Microbiology, 73, 4066–8.
Green, D. E. and Converse, K. A. (2005a). Diseases of amphibian eggs. In S. K. Majumdar,
Green, D. E. and Converse, K. A. (2005b). Diseases of frogs & toads. In S. K. Majumdar,

Green, D. E., Converse, K. A., and Schrader, A. K. (2002). Epizootiology of sixty-four
amphibian morbidity and mortality events in the USA, 1996–2001. Annals of the New
York Academy of Sciences, 969, 323–39.
Greer, A. L. and Collins, J. P. (2007). Sensitivity of a diagnostic test for amphibian ranavi-
nus varies with sampling protocol. Journal of Wildlife Diseases, 43, 525–32.
Griffiths, R. A. and Pavajeau, L. (2008). Captive breeding, reintroduction, and the con-
servation of amphibians. Conservation Biology, 22, 852–61.
Gutleb, A. G., Bronkhorst, M., van den Berg, J. H. J., and Murk, A. J. (2001). Latex
laboratory gloves-an unexpected pitfall in amphibian toxicology assays with tadpoles.
Environmental Toxicology and Pharmacology, 10, 119–21.
Harp, E. M. and Petranka, J. W. (2006). Ranavirus in wood frogs (Rana sylvatica): poten-
tial sources of transmission within and between ponds. Journal of Wildlife Diseases, 42,
307–18.
Jancovich, J. K., Davidson, E. W., Morado, J. F., Jacobs, B. L., and Collins, J. P. (1997).
Isolation of a lethal virus from the endangered tiger salamander Ambystoma tigrinum
stebbinsi. Diseases of Aquatic Organisms, 31, 161–7.
Jancovich, J. K., Davidson, E. W., Parameswaran, N., Mao, J., Chinchar, V. G., Collins, J. P.,
Jacobs, B. L., and Storfer, A. (2005). Evidence of emergence of an amphibian iridoviral
disease because of human-enhanced spread. Molecular Ecology, 14, 213–24.
Johnson, P. T. J. and Lunde, K. B. (2005). Parasite infection and limb malformation: a
growing problem in amphibian conservation. In M. Lannoo (ed.), Amphibian Declines:
the Conservation Status of United States Species, pp. 124–38, University of California
Press, Berkeley, CA.
Kimmel, C. A. and Trammell, C. (1981). A rapid procedure for routine double staining of
cartilage and bone in fetal and adult animals. Stain Technology, 56, 271–3.
Kriger, K. M., Hines, H. B., Hyatt, A. D., Boyle, D. G., and Hero, J.-M. (2006).
Techniques for detecting chytridiomycosis in wild frogs: comparing histology with
real-time Taqman PCR. Diseases of Aquatic Organisms, 71, 141–8.
Lavilla-Pitogo, C. R. and de la Pena, L. D. (2004). Diseases in Farmed Mud Crabs Scylla
spp.: Diagnosis, Prevention, and Control. SEAFDEC Aquaculture Department, Iloilo,
Philippines.
Lips, K. R., Brem, F., Brenes, R., Reeve, J. D., Alford, R. A., Voyles, J., Carey, C., Livo, L.,
Pessier, A. P., and Collins, J. P. (2006). Emerging infectious disease and the loss of bio-
diversity in a neotropical amphibian community. Proceedings of the National Academy
of Sciences USA, 103, 3165–70.
Longcore, J. E., Pessier, A. P., and Nichols, D. K. (1999). Batrochochytrium dendrobatidis
gen.et sp. nov., a chytrid pathogenic to amphibians. Mycologia, 91, 219–27.
Majji, S., LaPatra, S., Long, S. M., Sample, R., Bryan, L., Sinning, A., and Chinchar, V. G.
(2006). Rana catesbeiana virus Z (RCV-Z): a novel pathogenic ranavirus. Diseases of
Aquatic Organisms, 73, 1–11.
Mauel, M. J., Miller, D. L., Frazier, K., and Hines, II, M. E. (2002). Bacterial pathogens
isolated from cultured bullfrogs (Rana catesbeiana). Journal of Veterinary Diagnostic
Investigation, 14, 431–3.
Meteyer, C. U., Loeffler, I. K., Fallon, J. F., Converse, K. A., Green, E., Helgen, J. C.,
Kersten, S., Levey, R., Eaton-Poole, L., and Burkhart, J. G. (2000). Hind limb malfor-
mations in free-living northern leopard frogs (Rana pipiens) from Maine, Minnesota,
and Vermont suggest multiple etiologies. Teratology, 62, 151–71.
Miller, D. L., Bursey, C. R., Gray, M. J., and Smith, L. M. (2004). Metacercariae of
Clinostomum attenuatum in Ambystoma tigrinum mavortium, Bufo cognatus and Spea
multiplicata from west Texas. Journal of Helmithology, 78, 373–6.
Miller, D. L., Rajeev, D., Brookins, M., Cook, J., Whittington, L., and Baldwin, C. A.
(2008). Concurrent infection with Ranavirus, Batrachochytrium dendrobatidis, and
Aeromonas in a captive anuran colony. Journal of Zoo and Wildlife Medicine, 39, 445–9.
Miller, D. L., Gray, M. J., Rajeev, S., Schmutzer, A. C., Burton, E. C., Merrill, A. and
Baldwin, C. (2009). Pathological findings in larval and juvenile anurans inhabiting
farm ponds in Tennessee, U. S.A. Journal of Wildlife Diseases, 45, 314–24.
Muths, E., Gallant, A. L., Campbell, E. H. C., Battaglin, W. A., Green, D. E., Staiger, J. S.,
Walls, S. C., Gunzburger, M. S., and Kearney, R. F. (2006). The Amphibian Research
and Monitoring Initiative (ARMI): 5-year report. U.S. Geological Survey Scientific
Investigations Report 5224. U.S. Geological Survey, Reston, VA.
Ouellet, M., Bonin, J., Rodrigue, J., DesGranges, J. L., and Lair, S. (1997). Hindlimb
deformities (ectromelia, ectrodactyly) in free-living anurans from agricultural habitats.
Journal of Wildlife Diseases, 33, 95–104.
Pallister, J., Gould, A., Harrison, D., Hyatt, A., Jancovich, J., and Heine, H. (2007).
Development of real-time PCR assays for the detection and differentiation of Australian
and European ranaviruses. Journal of Fish Diseases, 30, 427–38.
Picco, A. M. and Collins, J. P. (2008). Amphibian commerce as a likely source of patho-
gen pollution. Conservation Biology, 22, 1582–9.
Poole, V. A. (2008). Amphibian Husbandry Resource Guide, edn 1.1. Association of Zoos
and Aquarium’s Amphibian Taxon Advisory Group, Yulee, FL.
Relyea, R. A. (2005). The lethal impact of Roundup on aquatic and terrestrial amphib-
ians. Ecological Applications, 15, 1118–24.
Relyea, R. A. (2009). A cocktail of contaminants: how mixtures of pesticides at low con-
centrations affect aquatic communities. Oecologia, 159, 363–76.
Rostand. J. (1958). Les Anomalies des Amphibiens Anoures. Sedes, Paris.
Schock, D. M., Bollinger, T. K., Chinchar, V. G., Jancovich, J. K., and Collins, J. P.
(2008). Experimental evidence that amphibian ranaviruses are multi-host pathogens.
Copeia, 2008, 133–43.
Schotthoefer, A. M., Koehler, A. V., Meteyer, C. U., and Cole, R. A. (2003). Influence of
Ribeiroia ondatrae (Trematoda: Digenea) infection on limb development and survival
of Northern leopard frogs (Rana pipiens): Effects of host-stage and parasite exposure
level. Canadian Journal of Zoology, 81, 1144–53.
Skerratt, L. F., Berger, L., Speare, R., Cashins, S., McDonald, K. R., Phillott, A. D.,
Hines, H. B., and Kenyon, N. (2007). Spread of chytridiomycosis has caused the rapid
global decline and extinction of frogs. EcoHealth, 4, 125–34.
Skerratt, L. F., Berger, L., Hines, H. B., McDonald, K. R., Mendez, D., and Speare, R.
(2008). Survey protocol for detecting chytridiomycosis in all Australian frog popula-
tions. Diseases of Aquatic Organisms 80, 85–94.
Storfer, A., Alfaro, M. E., Ridenhour, B. J., Jancovich, J. K., Mech, S. G., Parris, M. J., and
Collins, J. P. (2007). Phylogenetic concordance analysis shows an emerging pathogen
is novel and endemic. Ecology Letters, 10, 1075–83.
Storrs, S. I. and Semlitsch, R. D. (2008). Variation in somatic and ovarian develop-
ment: predicting susceptibility of amphibians to estrogenic contaminants. General and
Comparative Endocrinology, 156, 524–30.
USFWS and AFS-FHS (US Fish and Wildlife Service and American Fisheries Society – Fish
Health Section) (2005). AFS-FHS Blue Book: suggested procedures for the detection and
identification of certain finfish and shellfish pathogens, 2005 edition. American Fisheries
Society – Fish Health Section, Bethesda, MD.
Wake, D. B. and Vredenburg, V. T. (2008). Are we in the midst of the sixth mass extinc-
tion? A view from the world of amphibians. Proceedings of the National Academy of
Sciences USA, 105, 11466–73.
Walker, S. F., Salas, M. D., Jenkins, D., Garner, T. W.J., Cunningham, A. A., Hyatt, A. D.,
Bosch, J., and Fisher, M. C. (2007). Environmental detection of Batrachochytrium den-
drobatidis in a temperate climate. Diseases of Aquatic Organisms, 77, 105–12.
Whiles, M. R., Lips, K. R., Pringle, C. M., Kilham, S. S., Bixby, R. J., Brenes, R.,
Connelly, S., Colon-Gaud, J. C., Hunte-Brown, M., Huryn, A. D., Montgomery, C.,
and Peterson, S. (2006). The effects of amphibian population declines on the struc-
ture and function of Neotropical stream ecosystems. Frontiers in Ecology and the
Environment, 4, 27–34.
Williams, T., Barbosa-Solomieu, V., and Chinchar, V. G. (2005). A decade of advances
in iridovirus research. In K. Maramorosch and A. Shatkin (eds), Advances in Virus
Research, vol. 65, pp. 173–248. Academic Press, New York.
Wright, K. M. and Whitaker, B. R. (2001). Amphibian Medicine and Captive Husbandry.
Krieger Publishing, Malabar, FL.
Yip, M. J., Porter, J. L., Fyfe, J. A. M., Lavender, C. J., Portaels, F., Rhodes, M., Kator, H.,
Colorni, A., Jenkin, G. A., and Stinear, T. (2007). Evolution of Mycobacterium ulcerans
and other mycolactone-producing mycobacteria from a common Mycobacterium mari-
num progenitor. Journal of Bacteriology, 189, 2021–9.
Yuan, J. S., Reed, A., Chen, F., and Stewart, Jr, C. N. (2006). Statistical analysis of real-
time PCR data. BMC Bioinformatics, 7, 85–97.
27
Conservation and management
C. Kenneth Dodd, Jr
27.1 Introduction
Considering the large number of amphibian species and the diversity of life
histories, potential conservation strategies are likewise diverse and may be com-
plex. Beyond a few guiding principles, such as the need for planning and set-
ting objectives, understanding life-history constraints, and protecting existing
habitats and ecosystem function, there are no standardized approaches that
will be applicable to all situations. In the brief discussion that follows, I outline
some important considerations and various management approaches that have
proved successful in particular instances, and I provide references that con-
tain more extended discussions. Specific information on management options
for amphibian conservation are available in Beebee (1996), Gent and Gibson
(1998), Scoccianti (2001), Kingsbury and Gibson (2002), Semlitsch (2003),
Bailey et al. (2006), and Mitchell et al. (2006).
Amphibian conservation requires an integrated landscape approach to man-
agement, rather than solely a species-oriented focus. The reason for this is simple:
amphibians do not live in a biotic or physical vacuum in nature. Lindenmayer
et al. (2007) present a number of conceptual ideas for landscape conservation
that are apropos when focusing on amphibians; these apply equally to all man-
agement options (Table 27.1). Amphibian conservation options range from the
rather simple and inexpensive to the very complex and expensive. When plan-
ning, the overriding consideration should be “first, do no harm” to the species,
its community, or its habitat. High-technology-based approaches may work no
better than simple and inexpensive approaches, and care should be exercised to
maximize the benefits from the human resources and funds available. The pri-
mary objectives of management should always focus on the species or commu-
nity of concern, and not on peripheral or extended objectives, such as positive
publicity.
Table 27.1 Conceptual ideas for landscape conservation that are important when
planning for amphibian management. Adapted from Lindenmayer et al. (2007) and
Gardner et al. (2007).
Conceptual ideas important for amphibian management
Plan for habitat connectivity, particularly between aquatic and terrestrial habitats
Develop a long-term approach, both in management and in the evaluation of objectives
Manage all aspects of the habitat, not individual components, regarding landscape features,
species, or stressors
Manage for change, not a static condition
Manage species and ecosystems using explicit experimental approaches and at multiple
scales, rather than resorting to “cookbook” approaches
27.1.1 Statutory protection

The simplest approach to amphibian conservation is to legislate protection
for species or habitats. In some cases, this can be extremely effective at pre-
venting habitat destruction or alteration, adverse trade, prevention of take, or
similar threats. Legislative approaches work best where there is a direct threat
to amphibians or crucial habitats, and where the laws or regulations will be
enforced. Measures designed to decrease or eliminate threats also may result
from legislative protection, such as mandated mitigation requirements or direc-
tives to develop recovery plans. Active management and research must accom-
pany legislative protection; simply putting a species on a list could lead to a false
sense of security concerning its protection and, without enforcement, offers no
long-term conservation benefits.
27.1.2 Protecting habitats

Coupled with legislative protection is the need for control over habitats where
imperilled amphibian species or communities reside. Habitat protection can be
achieved through outright purchase by governments or private organizations or,
in some cases, by executive order transferring land from one level of controlled
use to a more restrictive category (such as by changing land-zoning categories).
Habitat purchase allows direct control and incorporation of land into parks,
preserves, refuges, or wildlife management units that may allow for multiple use
as long as it does not conflict with the conservation objectives outlined in the
purchase plan.
In addition to direct purchase, lands may be protected via conservation ease-
ments or agreements, tax incentives, or by payment subsidies. Agreements can
be flexible to ensure conservation yet allow landowners access and use of their
27 Conservation management | 509
lands. In all cases, however, management plans must ensure that long-term
research and monitoring is carried out, and that land use occurs in accord-
ance with initial agreements. All land acquisition and alternative conservation
programs require early, careful, fiscal planning in order to ensure long-term
continuity of protection, management and assessments of how well objectives
are being achieved.
People residing adjacent to protected lands can be incorporated into long-
term management programs to provide local or regional economic incentives to
protection. Such an approach would be especially desirable where key amphib-
ian habitats are located in proximity to economically disadvantaged human
settlements. Providing educational programs, outlets for local arts, and even
minimal health care could go a long way to ensure that crucial habitats are
protected from vandalism or incursion. This approach works well in both devel-
oped and underdeveloped regions, for example, where local residents have been
hired as caretakers. Residents thus come to have a stake in the protected area and
in the species being protected.
27.2 Managing amphibian populations

I suggest that there are five cardinal rules when developing management pro-
grams for amphibians (Dodd et al. 2009).
• Understand amphibian life histories. Conservation and management can
only proceed with knowledge of amphibian biology, including but not
limited to habitat requirements, spatial use of habitats, reproduction, diet,
predators, and population biology. Life-history and demographic variables
form the biological constraints to management efforts.
• Address causes of decline. Species cannot be managed or recovered unless
the causes of decline or need for conservation are understood and corrected.
It makes no sense to place imperilled individuals in a continuing position of
threat.
• Consider human-based constraints. Finances, personnel, logistics, time,
regulations, local sensitivities and collaboration, and administrative pol-
icy must be factored into a management program prior to its initiation.
The best of intentions will fail unless these human-based constraints are
adequate to complete the program in accord with the biological require-
ments of the species.
• Have achievable biological objectives, plans, and measures of success.
These factors should be clearly stated, and management efforts should
be directed at achieving specific goals. Management plans are guidelines,

readily changed to accommodate new insights; they should not be altered
to meet political objectives.
• Monitor conservation outcomes. Assessing the effectiveness of conserva-
tion research and management is essential to determining effectiveness.
Monitoring should be considered as fundamental research designed to
address specific questions, and must be long-term in nature.
27.3 Wetland breeding sites

In considering management options to conserve wetland-breeding amphib-
ians, Semlitsch (2000a) suggested the following: (1) maintenance or restoration
of temporary wetlands with a diverse array of hydroperiods, (2) protection of
terrestrial buffer zones of natural vegetation to protect core breeding sites, (3)
protection of amphibian communities by invasion of fish predators (primarily
in reference to seasonal wetlands), (4) protection of the integrity of ecological
connectivity across a landscape, (5) restriction of chemical use on site, but espe-
cially in ditches, streams, and wetlands, (6) prohibiting the release of captive or
exotic specimens, and (7) identification and resolution of current management
practices with those necessary for amphibian conservation.
In general, wetlands of all sizes serve as breeding sites or habitats for adult
amphibians, from small tree holes to large lakes. A variety of regulations pro-
tect rivers, streams, and lakes in many parts of the world, but it is important
to recognize the value of small, seasonal wetlands to amphibians (Semlitsch
2000b); these small isolated wetlands are usually overlooked and regulations
rarely protect them. Any management plan on a landscape level must take them
into account, and plans for their protection, restoration, or creation are vital to
the survival of many amphibian species.
27.3.1 Wetland integrity and hydroperiod

Managing wetlands for amphibians involves the maintenance of breeding and
dormancy sites (e.g. lake or pond bottoms, small streams) in conditions favorable
for habitation and larval development. Conservation involves preventing physical
disturbances, regulating the influx of pollutants, preventing the establishment of
fish or non-indigenous predators and vegetation, and controlling vegetative suc-
cession where habitats are spatially restricted or limited in distribution.
Wetlands may be surrounded by plastic sheeting to minimize the temporary
inflow of siltation and pollutants, such as during road construction, but in cases
where disruption to adjacent habitats is permanent, walls or barriers must be built.
Walls can be constructed of pre-cast concrete or other materials (section 27.5.2).

Barriers might incorporate passages to allow some animals to cross, or walls could
be sloped in such a manner (/——\) as to allow animals to crawl up and across
the barrier. This design would block silt or contaminated inflow yet not serve as a
barrier to migration.
Wetland ditching is counterproductive to amphibian management, even
though some agricultural ‘best management practices’ advocate it under the
guise of ‘precision levelling’ (Trauth et al. 2006). Maintaining hydroperiod (the
length of time water is available within a wetland) is critically important. Many
amphibians breed in temporary or seasonal wetlands free of fish predators.
Periodic drying ensures that fish do not become established. Many processes
disrupt the hydroperiod, such as ditching adjacent habitats or pumping under-
ground water to supply municipal needs. In such cases, water can be diverted
back to the wetland during the time necessary for amphibian development.
Ground water may have different pH and chemical properties than surficial
wetlands, however, so it may be necessary to experimentally test developmental
rates and success prior to augmenting breeding ponds with ground water (Seigel
et al. 2006).
Non-indigenous predators (other amphibians, fish, crayfish) are extremely
detrimental to amphibians breeding in seasonal wetlands; they should never
be intentionally released. Fish may be killed by draining water from temporary
ponds during the non-breeding season, or by applying a piscicide such as rote-
none. In streams, non-native fish have been removed by using a combination
of rotenone and direct netting. However, some piscicides may be detrimental
to larval amphibians, so care must be taken when choosing chemical applica-
tions. Never apply a chemical unless there is published data showing that it is
safe. Otherwise, it will be necessary to conduct experimental evaluation prior
to field use.
Invasive vertebrates such as Cuban treefrogs (Osteopilus septentriona-
lis), American bullfrogs (Rana catesbeiana), and wild hogs (Sus spp.) may be
extremely detrimental to amphibians at breeding sites. Hogs and cattle (Bos
spp.) can be excluded by fencing, but fencing may be counterproductive for
bullfrogs inasmuch as native amphibians also need to enter or leave the wetland.
Species with extended larval periods can be removed by draining ponds to kill
larvae, but adult American bullfrogs must be shot or manually removed. Once
removed, high and fine-meshed perimeter fences can be used to prevent immi-
gration, since bullfrogs move over large distances, even in desert conditions.
Crayfish also must be manually removed or trapped because of their burrowing
ability.
The larvae of Cuban treefrogs have detrimental impacts on native tadpoles

(Smith 2005), and adults consume or out-compete native species for retreat
sites. PVC pipes can be used to artificially establish retreat sites and, once colo-
nized, individuals can be removed and euthanized. Whether this technique will
reduce populations is unknown, so such a removal plan may require extended
effort. Although not exotic to their locale, manual removal of water snakes
(Natrix maura) has proved effective in helping to recover and re-establish popu-
lations of Alytes muletensis on Mallorca (Román 2003).
Vegetation encroaching upon wetlands may seriously impact them by extract-
ing additional water, changing the chemistry and pH, physically changing the
habitat structure of calling and egg-deposition sites, and enclosing the canopy,
thus altering thermal regimes. Invasive plants have detrimental effects on lar-
val development, and the secondary compounds they produce may be toxic to
some species (Maerz et al. 2005a, 2005b; Brown et al. 2006). Vegetation may
be mechanically or selectively removed or thinned, but care should be taken
when using herbicides inasmuch as they or their surfactants may have lethal or
sublethal effects on developing larvae (e.g. Hayes 2004).
27.3.2 Wetland creation and restoration

Wetland creation has proven very successful in providing oviposition sites for
many pond-breeding amphibians, for both rare and common species (Table
27.2; Anonymous undated; Aplin et al. 2000). Newly created breeding ponds
are usually placed in areas near former breeding sites, the exact location depend-
ent on existing canopy cover, land use, and vegetation. Created ponds should
be located in proximity to non-breeding habitats, and dispersal corridors must
be included to accommodate terrestrial foraging activity and winter/dry season
retreats. Ponds may be dug by machine or by hand, with subsequent seeding of
native vegetation around the shoreline in such a manner to accommodate the
breeding habits of targeted species. For example, some species require shallow
emergent vegetation for egg deposition, whereas others require arboreal and
structured vegetation as calling sites. Zooplankton may be added to provide a
food source for tadpoles.
Once dug, many created ponds will require a liner of clay, bentonite, foil, rub-
ber or plastic to allow water retention. The ability to seal ponds may dictate the
size of the pond or wetland to be created inasmuch as small wetlands are easier to
line than large wetlands. Clay works well in some habitats, but often deteriorates
through time, especially if the ponds are seasonal and the bottom is exposed to
air for long periods. A “permanent” liner may not allow for seasonal water fluc-
tuation, however, except through evaporation. Thus, many newly created ponds
Table 27.2 Programs featuring breeding pond creation or restoration as part of amphibian habitat management.
Location Species Type of constructed pond Reference
New South Wales, Australia 10 frogs Constructed farm ponds Hazell et al. (2004)
Prince Edward Island, Canada 6 frogs Dredged wetlands Stevens et al. (2002)
3500 ponds, Denmark 7 frogs, 2 salamanders Dredging existing ponds, dug ponds Fog (1997)
Samsø, Denmark Bufo viridis Dredged ponds, vegetation removal Amtkjaer (1995)
Lolland, Denmark Hyla arborea Dredging existing ponds, dug ponds Hels and Fog (1995)
France Pelobates fuscus Use of bentonite and plastic-liners Eggert and Guyetant (2002)
Muensterland, Germany 5 frogs, 3 salamanders Dug and dredged ponds Schwartze (2002)
Salzburg, Germany 3 frogs, 4 salamanders Foil-lined permanent ponds Kyek et al. (2007)
Arribes del Duero, Spain 4 frogs, 4 salamanders Permanent dug ponds Alarcos et al. (2003)
Gipuzkoa, Spain Hyla meridionalis Permanent dug ponds Rubio and Etxezarreta (2003)
UK Rana temporaria Dug ponds, mostly seasonal Williams (2005)
Arkansas, USA Rana sylvatica Permanent dug wildlife ponds Cartwright et al. (1998)
Colorado, USA Bufo boreas Dug ponds, 1.5–6 m deep, little or no vegetation Pearl and Bowerman (2006)
restoration
Illinois, USA 11 frogs, 6 salamanders Earthen dams in shallow valley, ponds permanent Palis (2007)
Minnesota, USA 7 frogs, Ambystoma tigrinum Filled ditches, reflooded basins Lehtinen and Galatowitsch (2001)
North Carolina, USA Rana sylvatica, Seasonal and permanent ponds, modified golf Petranka et al. (2003)
Ambystoma maculatum course ponds, blocked stream channels
Ohio, USA 7 frogs, 2 salamanders Complex ponds with shallow and deep sections, Weyrauch and Amon (2002)
revegetation, zooplankton and some amphibians stocked
South Carolina, USA 13 frogs, 4 salamanders Plastic-lined permanent dug ponds Pechmann et al. (2001)
are permanent and require vigilance to ensure that fish predators are not inad-
vertently (through sheet flow) or purposely introduced. Weyrauch and Amon
(2002) provided an innovative U-shaped design for created ponds whereby one
arm of the U is dug shallow and the other deep. Natural evaporation allows the
water to fluctuate in hydroperiod between the arms, and provides amphibians
with a choice of water depth.
Once ponds are dug, they are filled naturally by precipitation. Thus, ponds
are best created prior to the season when most precipitation occurs, allowing
pond chemistry and limnology to stabilize before the breeding season. Water
may be added to ensure that at least some amount is available between con-
struction and the onset of rainy periods. Most invertebrates and amphibians
will naturally colonize newly created ponds, often very rapidly. However, some
researchers have combined pond creation with stocking or relocation in order
to speed-up the colonization process or to target certain declining species. As
Petranka et al. (2003) noted, perturbations (disease, drought) sometimes hin-
der efforts to establish amphibians at created ponds, and monitoring needs to
continue for approximately 5 years or more to evaluate success. The longer the
pond is maintained, the more likely that additional species will colonize it, and
that amphibian populations will become established.
Existing breeding ponds may be restored or enhanced, usually by mechanic-
ally deepening the pond to remove infill or muck and by removing encroaching
vegetation. Trees, such as willows, are cut within the pond basin and along
the pond’s perimeter. Herbicides must be used with extreme care at amphibian
breeding sites, since these or their surfactants may have detrimental effects on
amphibians. In fire-dependent habitats, controlled burns through pond basins
may help to periodically re-establish breeding ponds by burning off muck and
vegetation. Typically, pond restoration is carried out during the non-breeding
time of the year when individuals are dispersed. Biebighauser (2002) has pre-
pared a free guidebook to creating and maintaining vernal ponds.
27.3.3 Core habitat and buffer zones

Key amphibian habitats include wetland breeding sites, retreat sites, dispersal
corridors, and unique or limited habitats, such as caves, rock faces, steeply sided
slopes, or areas where populations may be restricted in distribution, such as on
mountain tops or in the spray zone of waterfalls. Each such habitat constitutes
a critical area where the integrity of the landscape must be fully protected to
ensure population survival. Although Semlitsch and Jensen (2001) advanced
the idea of core habitats for wetland-breeding amphibians, the concept may be
expanded to include these other unique habitats in terms of management.
Apart from the critical conservation area, zones of protection may extend
outward depending on the extent to which amphibians travel away from the
site. The critical area requires a buffer to ensure its integrity, and the “core”
habitat to be protected becomes the critical area plus its associated buffer zones.
For wetland-breeding amphibians, a core area for conservation would include
the wetland site, plus both aquatic and terrestrial buffer zones that would
ensure a functioning community (see Figure 27.1). For a terrestrial salamander
habitat specialist, a core conservation area might include the immediate steep
slopes or rock face on which it resides, plus terrestrial buffer zones to allow for
complete canopy cover over the slope or rock face, allowing retention of shade
and high humidity. Thus, the concept of a “buffer” zone should be expanded
to include as much habitat as necessary to maintain functionality, rather than
F2 F1
AB CH
CW
Wil
dlife
Cor
rido
r
TB
F3 F4
Fig. 27.1 Protection of a pond-breeding amphibian community at a landscape scale

in a tree plantation. A core wetland (CW) is surrounded by three areas of protection,
the aquatic buffer zone (AB), a core habitat (CH; ABan area where most individuals
disperse), and a terrestrial buffer zone (TB). The CW is linked to other wetlands via
wildlife corridors. The surrounding tree farm is managed in rotations, from currently
harvested (F1) to various-age stands (F2–F4). The level of human use varies from
total protection in CW, to recreational use in CH, to light selective harvesting in TB,
to rotational clear- or selective cutting (F1–F4). This configuration provides a mosaic
of habitats that allows wetland-breeding amphibian to persist through time, yet
permit some human recreational and commercial activity.
be synonymous with a minimum area necessary to protect the critical area from
destruction or alteration.
Certainly core areas and associated buffer zones need to be managed essen-
tially free from adverse effects on resident amphibians. However, some human
uses may be allowed within a series of outermost concentric buffer zones such
that layers of allowed uses surround the core conservation area. For example, a
core protection area might extend 100 m from a wetland breeding site (a core area
plus aquatic and terrestrial buffer zones) where no forestry would be allowed.
However, very selective cutting could be allowed at distances of 100–150 m,
with even less restrictions on the types of activity at further than 150 m (see
Figure 27.1). Given that amphibians are not evenly distributed around a core site
or present at all times of the year, developing an effective core plus buffer zone
conservation and management area requires good data on spatial distribution.
A common question is, how much habitat must be included as conservation
buffer zones for amphibians? As Semlitsch and Jensen (2001) noted, there is no
one-size-fits-all value, and the size of buffer zones will vary depending on spe-
cies and community. In temperate habitats, amphibians often routinely travel
more than 500 m from breeding sites; thus, buffer zones may need to extend
much farther than some previous authorities have suggested. Based on extensive
life-history and movement data, for example, Semlitsch and Jensen (2001) sug-
gested that a core plus buffer zone of 164 m would provide substantial protec-
tion for 95% of the salamander community that they studied.
27.3.4 Vegetation structure, composition, and canopy cover

Amphibians require a variety of vegetative structure around a wetland; often,
the structure of the vegetation is more important than the species composition.
Knowledge of the breeding habits of local species is the best guide to vege-
tation management; that is, whether amphibians need elevated calling sites,
shallow emergent vegetation for cover, or woody debris within a pond around
or on which to deposit eggs. In complex amphibian communities, managers
should plan for a diversity of vegetative structure both within wetlands and in
associated uplands. Canopy cover is particularly important as it affects thermal
regimes. Many species do not breed in enclosed canopy, so tree and vegetation
removal will be necessary as wetlands undergo succession.
27.3.5 Water quality

Each species of wetland breeding amphibian has its unique water-quality pre-
requisites for tadpole development. Part of successfully managing amphibians is
knowing these requirements and monitoring them during the course of surveys
(Chapter 7). Acidity is particularly important, as it affects the amount of oxygen

available to the developing embryo. As vegetation succession occurs, layers of
peat and muck accumulate and may change the pH to non-favorable levels. Such
muck can be mechanically removed to raise pH to pre-accumulation levels, or
ponds may be limed (with CaCO3 or Ca(OH)2)to increase pH to circumneutral
levels (Bellemakers and van Dam 1992; Beattie et al. 1993).
27.4 Terrestrial habitats

Terrestrial habitats are critically important for amphibians, although they are
often overlooked. Managing amphibians terrestrially involves knowledge of
species richness and movement patterns, especially seasonal migratory pathways
and orientation during breeding seasons. Species that do not breed in water (e.g.
many tropical frogs and plethodontid salamanders) or spend only part of their
life cycle on land need moist or humid sites containing cover, surface debris,
subsurface burrows, and retreats during cold or dry seasons. Activities which
remove natural canopy cover, alter leaf litter, do not mimic natural disturbances
(e.g. fires in non fire-evolved ecosystems), or disturb subsurface retreats (e.g.
roller-chopping and disking) should be avoided.
27.4.1 Contiguous habitats and edge effects

Large contiguous habitats are ideal for conservation, whether they be rainforest,
deciduous woodlands, or prairie grasslands. The more such habitats are frag-
mented, diminished, or shredded, the less desirable they are for maintaining
amphibian biodiversity. Maximum effort should be placed on preserving such
habitats. However, nearly all remnants of native habitats are likely to contain
some amphibians, depending on location and spatial scale. Since edges expose
amphibians to barriers for dispersal, predators, and altered biophysical condi-
tions, landscape configurations designed to minimize edges and maintain inter-
ior areas far from edges have the best chances of protecting amphibians through
time. A mosaic of breeding and retreat sites must be included within contiguous
areas, and corridors of various widths, lengths, and vegetation structure should
connect principle amphibian habitats.
27.4.2 Silviculture
One of the most contentious issues in amphibian conservation management is the
effect of silvicultural practices on amphibians (deMaynadier and Hunter 1995).
However, there are still important principles which should be kept in mind.
Certainly, practices which completely destroy wetland breeding sites should be
prevented, and buffer zones need to be implemented around streams and wet-
lands to avoid desiccation and changes in thermal regimes and water chemistry.
The most extreme form of forestry is clear-cutting all trees and vegetation.
Many species of terrestrial amphibians, particularly salamanders and ground-
dwelling tropical frogs, decline or disappear after clear-cutting, and surviving
populations may take many decades to return to pre-cut abundance. If left to
succession in the absence of mechanical roller chopping, however, even such
severely degraded sites can recover if amphibian source populations are nearby
and wetlands with buffers remain intact. However, these sites cannot recover
if wetlands are destroyed, large areas are clear-cut, roller-chopping eliminates
all subsurface retreats, such as root tunnels, and all surface debris is eliminated
manually or by burning. Herbicides and fertilizers also may be detrimental to
many species.
One solution to these problems is to manage for a mosaic of habitats. Some
areas (including breeding sites and associated buffer zones) would be off limits
to cutting, whereas other areas would be cut on a rotational basis, thus pro-
viding different aged stands at various locations in the landscape (see Figure
27.1). Patches would have to be sufficiently large to be economical yet retain the
diversity of amphibian species and allow for inter-patch migration. If plantation
forestry is involved, patches of native vegetation could be retained within the
landscape mosaic. A second solution to forest management would be to employ
selective cutting that allows for canopy cover, yet does not significantly affect
surface and subsurface conditions.
27.4.3 Restoring degraded lands

Natural succession may lead to eventual restoration of degraded amphibian habi-
tats, as long as source populations are within colonizing distance and the habitat
has not been critically disturbed (e.g. wetlands, water table, subsurface retreats).
However, some habitats will not recover without extensive renovation, such as
mine sites, degraded or silted streams, or sites where vegetation has changed to
the extent where specialized amphibians can no longer occur. In such cases,
extensive and usually expensive vegetation and habitat restoration (including
wetland creation, altering water chemistry, and building retreat sites) will be
required. Habitat generalists have the best success in repopulating recovered
terrestrial sites, but unaided recolonization and recovery, especially by terrestrial
species, may take decades. Still, there have been successes. For example, endan-
gered heath sites have been reclaimed in Britain for Bufo calamita by removing
encroaching terrestrial scrub vegetation and intensively managing the pH of
wetlands (Beebee 1996).
27.5 Migratory and dispersal routes

One of the primary goals of amphibian conservation management at the land-
scape scale is to ensure connectivity between breeding sites, foraging areas, and
winter/dry season retreat sites. Amphibians are sometimes capable of long-
distance movements, and roads (even unpaved), habitat edges, and open areas
have profound effects on movement propensity by some species, but not others.
27.5.1 Corridors between habitat fragments

In forested habitats, the best way to ensure amphibian connectivity is to retain
forested corridors between breeding sites and fragmented larger areas. Forested
corridors should also be retained along the riparian zones of creeks, streams,
and rivers; corridors should extend on both sides of the stream. The size of the
corridor will be different in different habitats, but should allow a swath of at
least 30–50 m on each side to reduce edge effects, particularly desiccation and
exposure to sunlight.
When open areas surround fragmented forests, many amphibians will not
cross the ecotonal boundary. However, patches of forest trees or dense shrubs
can be planted at regular intervals between patches to serve as cover sites. Small
created or restored wetlands interspersed along the corridor also facilitate move-
ment between core fragments. In savannas or grasslands, heaps of rocks, woody
debris (log piles), and created wetlands can also be used to facilitate movement
between habitat patches. The objective is to create a mosaic of complex habitats
along the way to provide amphibians a stepping-stone pattern across unfavor-
able or hostile habitats (Chan-McLeod and Moy 2007).
27.5.2 Crossing transportation corridors

One of the most formidable barriers to amphibians are the literally thousands
of kilometres of roads, highways, rail lines, and other transportation corridors
which fragment habitats. Inasmuch as many species migrate between breed-
ing sites and foraging/retreat sites, crossing highways leads to mass mortality
which affects population persistence. The effects of roads may extend hundreds
of meters away from the actual roadway. As a result, an extensive literature has
evolved on methods to allow amphibians and other species to cross.
The most effective way to assist amphibians across roads is to construct a
barrier wall to prevent access, coupled with a series of underpasses, culverts, or
tunnels to allow transit to the other side (Figure 27.2a–d). The walls serve to
prevent trespass, while at the same time funnelling the migrating amphibians
toward the underpass or culvert. Walls may be made of pre-cast concrete using
(a) (b)
(c) (d)
Fig. 27.2 (a) Overview of drift fence/culvert system around an amphibian

breeding pond; Arcegno, Canton of Ticino, Switzerland. Photograph by P. Schlup,
with permission. (b) Combining a tunnel with water to allow for a high-humidity
microclimate during transit underneath a roadway; near Wiedlisbach, Canton of
Berne, Switzerland. Photograph by S. Zumbach, with permission. (c) A recently
constructed tunnel/barrier wall allowing water flow. Note the seamless connection
between the tunnel and barriers; near Belp, Canton of Berne, Switzerland.
Photograph by B. Lüscher, with permission. (d) Combined barrier wall and culvert
system, Paynes Prairie State Preserve, Alachua County, FL, USA. The lip at the
top of the wall discourages trespass by small treefrogs. However, the lack of
vegetation maintenance severely reduces the effectiveness of the wall by allowing
small vertebrates (frogs, snakes, mammals) to climb up and over the barrier lip.
Photograph by C. Kenneth Dodd, Jr, October 2008.
a variety of lip (or overhang) designs to discourage climbing amphibians from

crossing over the wall; slick, rather than rugose, surfaces also help to discourage
climbing. The size of the culverts has varied among projects, but most provide
for light and moisture access (to prevent animals from desiccating and provide
an incentive for them to enter) and a moist dirt, mud or water substrate.
Culverts or tunnels may be built well under a raised road bed, or even level with
a roadway surface with a grate forming the ceiling when raised road beds are not
present; a permanent barrier wall need not be present. In such cases, drift fences
constructed of highway cloth or sheet plastic can be used to funnel amphibians
toward the underpass. Some amphibians will enter even narrow tunnels and
successfully cross a highway, but others may require culverts of more than 1 m
diameter. The length of the culvert may have important effects of the propen-
sity of an amphibian to enter and cross successfully. For long culverts, as across
more than four lanes, the underpass might have to be constructed in sections,
and access to light and moisture become critical. Entrances should allow easy
access in both directions, and brush or other woody debris may provide retreat
sites and refuges from predators. Excellent discussions showing diagrams and
photographs of various types of amphibian passages are provided in Langton
(1989), ALASV (1994), Percsy (1995), Glandt et al. (2003), Brodziewska
(2005), and in the Proceedings of the International Conferences on Ecology
and Transportation (www.icoet.net/) (also see Figure 27.2a–d). Overpasses
generally do not work as well for amphibians as they do for mammals.
27.6 Intensive manipulation of individuals

27.6.1 Captive breeding
Captive breeding of imperilled amphibians is a high-technology-based option of
last resort; many programs are very costly, and few have been successful (Dodd
2005; Griffiths and Pavajeau 2008). Although many amphibians have high
fecundity, short generation times, and can be kept in reasonable numbers, other
species do not have these characteristics, especially those in decline, or they are
very difficult to maintain and breed in captivity. Prior to undertaking captive
breeding for conservation or reintroduction, a careful plan must be developed
that takes into consideration housing, cleanliness, disease screening, specialized
cover, diet and biophysical requirements, and the potential effects of long-term
captivity on fitness. For example, Kraaijeveld-Smit et al. (2006) showed that
predator escape behavior and heterozygosity of Alytes muletensis could be main-
tained for several generations in captivity, but that these traits start to deteri-
orate over time, thus potentially affecting long-term conservation programs.
Conservation is best carried out in situ unless the threats become so dire that
extinction is likely without intervention and removal to a secure location.
27.6.2 Relocation, repatriation, translocation (RRT)

Moving animals out of harm’s way (relocation), into areas currently or previously
occupied (repatriation), or even to areas not previously occupied (transloca-
tions) has been popular in wildlife management. In some cases, RRT programs
have been undertaken without consideration of alternative options or with a
firm understanding of the full requirements of RRT projects; they are often
expensive and labor-intensive. As a result, many amphibian RRT programs have
been unsuccessful, or the ultimate fate of the RRT is still unknown (Dodd
and Seigel 1991; Dodd 2005). There have been some successful RRT programs
(Román 2003; Rubio and Etxezarreta 2003; Bell et al. 2004; Kinne 2004), and
such projects are still often advocated.
There are a number of criteria to consider, in addition to funding and staff-
ing, prior to undertaking RRT projects (Dodd and Seigel 1991): (1) the causes
of the original decline should be known, and steps must be taken to rectify
them, (2) the biological constraints on conservation relating to life history, habi-
tat, and demography should be understood, (3) population genetic and social
structure should be evaluated, (4) individuals should be screened for disease,
and (5) long-term habitat protection and monitoring should be a part of the
RRT project. These criteria need to be considered for both the donor and recipi-
ent individuals and populations, where appropriate, especially when a resident
population remains at or in the vicinity of a recipient site. To avoid problems of
half-way technology, RRT should be considered only after other less manipula-
tive options have been considered.
Various life stages have been used in RRT programs and to augment exist-
ing populations. Captive-reared animals may been used as alternative sources
for all life stages, particularly juveniles and adults (Bloxam and Tonge 1995)
for RRT, rather than removing individuals from existing natural populations.
Most projects use larvae or eggs for RRT, as these are the easiest life-history
stages to obtain and most abundant for mass movement. Presumably, larvae
metamorphosing from a breeding site are more likely to remain in the vicinity
than adults or subadults found or reared in one location but released in another.
Adult mortality may be high in unfamiliar surroundings, and adults generally
are scarcer or less available than eggs or larvae for RRT.
Most successful RRT programs employ a stage-based release protocol through
time, rather than planning for a single release. This helps ensure that individuals
become established, especially when presented with stochastic environmental
conditions. Animals may have to be moved for years prior to successful estab-
lishment. Pre-movement activities could include population censuses, assess-
ment and preparation of suitable habitat at recipient sites, disease and genetic
screening, predator removal, and education.
27.6.3 Disease and biosecurity

Diseases and parasites, including bacteria (Aeromonas, Citrobacter), fungi
(Batrachochytrium dendrobatidis, Saprolegnia), mesomycetozoa (Amphibi-
ocystidium, Ichthyophonus), a newly described alveolate pathogen (Davis et al.
2007), and trematodes (Ribeiroia ondatrae) have been shown to seriously affect
amphibian species and populations (see Chapter 26). In terms of management,

biosecurity protocols aim to stop the spread of these diseases throughout a land-
scape and between continents.
Pathogens may be transported by wildlife, moved when hosts are used as bait,
stocked with fish, sold as pets, deliberately released, inadvertently moved on vehi-
cles, boats, and other forms of human transportation, used in medical tests or for
educational and research purposes, or transported during catastrophic events, such
as floods and tropical storms. Pathogens may move as organisms or propagules,
and they can be transported in water, air, or on human and wildlife vectors.
In some instances, regulations can be established which prohibit the deliberate
release of nonindigenous species or of individuals held in captivity. Shipments of
stocked fish should be screened for amphibians, and both water and fish treated
to ensure that diseases are not spread; such treatments are often already mandated
when endangered fish species are released. Likewise, amphibians in the pet trade
and food industry, particularly the American bullfrog (R. catesbeiana), should be
screened for amphibian chytrid since they are vectors of this insidious disease. In
cases where ponds are known to be infected by amphibian pathogens, public entry
may be prohibited and attempts should be made to limit access by wildlife.
27.7 Conclusion
Preventing the decline of amphibians is challenging in the global economy of
the twenty-first century, one that requires biologists to use a complex array of
innovative and integrated approaches. The greatest threat to amphibians remains
the loss or alteration of habitats, so primary emphasis should be placed on main-
taining large tracts of undisturbed areas and to link already fragmented habitats.
Conservationists must use tested and scientifically based options, and research
must address questions in the field, laboratory, and through statistical model-
ling. Approaches need to focus on the parsimonious (what is most likely to work)
rather than on the latest technological advancement, unless they are synonym-
ous. Amphibian conservation is in triage mode; researchers are competing with
many others for scarce resources. It is our responsibility to use these resources
wisely for amphibians as we navigate through the Earth’s sixth great extinction.
27.8 References
Alarcos, G., Ortiz, M. E., Lizana, M., Aragón, A., and Fernández Benéitez, M. J. (2003).
La colonización de medios acuáticos por anfibios como herramienta para su conser-
vación: el ejemplo de Arribes del Duero. Munibe, 16, 114–27.
ALASV(ArbeitsgruppeunterLeitungderAbteilungStraßenbaudesVerkehrsministeriums)
(1994). Amphibienschutz. Leitfaden für Schutzmaßnahmen an Straßen. Schrifttenreihe
der Straßenbauverwaltung Baden Württemberg, Heft, 4.
Amtkjaer, J. (1995). Increasing populations of the green toad (Bufo viridis) due to a pond
project on the island of Samsø. Memoranda Societatis pro Fauna et Flora Fennica, 71,
77–81.
Anonymous (undated). Garden Ponds as Amphibian Sanctuaries. British Herpetological
Society, Montrose.
Aplin, K., Paino, A., and Sleep, L. (2000). Building Frog Friendly Gardens. Western
Australian Museum, Perth.
Bailey, M. A., Holmes, J. N., Buhlmann, K. A., and Mitchell, J. C. (2006). Habitat man-
agement guidelines for amphibians and reptiles of the southeastern United States. Technical
Publication HMG-2. Partners for Amphibian and Reptile Conservation, Montgomery,
AL.
Beattie, R. C., Aston, R. J., and Milner, A. G. P. (1993). Embryonic and larval survival
of the common frog (Rana temporaria) with particular reference to acidic and limed
ponds. Herpetological Journal, 3, 43–8.
Beebee, T. J. C. (1996). Ecology and Conservation of Amphibians. Chapman & Hall,
London.
Bell, B. D., Pledger, S., and Dewhurst, P. L. (2004). The fate of a population of the
endemic frog Leiopelma pakeka (Anura: Leiopelmatidae) translocated to restored habi-
tat on Maud Island, New Zealand. New Zealand Journal of Zoology, 31, 123–31.
Bellemakers, M. J. S. and van Dam, H. (1992). Improvement of breeding success of the
moor frog (Rana arvalis) by liming of acid moorland pools and the consequences of lim-
ing for water chemistry and diatoms. Environmental Pollution, 78, 165–71.
Biebighauser, T. R. (2002). A Guide to Creating Vernal Ponds. U.S. Department of
Agriculture, Forest Service, Morehead, KY.
Bloxam, Q. M. C. and Tonge, S. (1995). Amphibians: suitable candidates for breeding-
release programmes. Biodiversity & Conservation, 4, 636–44.
Brodziewska, J. (2005). Wildlife tunnels and fauna bridges in Poland: past, present,
and future, 1997–2013. International Conference on Ecology and Transportation 2005
Proceedings, pp. 448–60.
Brown, C. J., Blossey, B., Maerz, J. C., and Joule, S. J. (2006). Invasive plant and experi-
mental venue affect tadpole performance. Biological Invasions, 8, 327–38.
Cartwright, M. E., Trauth, S. E., and Wilhide, J. D. (1998). Wood frog (Rana sylvat-
ica) use of wildlife ponds in northcentral Arkansas. Journal of the Arkansas Academy of
Science, 52, 32–4.
Chan-McLeod, A. C. A. and Moy, A. (2007). Evaluating residual tree patches as stepping
stones and short-term refugia for red-legged frogs. Journal of Wildlife Management, 71,
1836–44.
Davis, A. K., Yabsley, M. J., Keel, M. K., and Maerz, J. C. (2007). Discovery of a novel
alveolate pathogen affecting southern leopard frogs in Georgia: description of the dis-
ease and host effects. EcoHealth, 4, 310–17.
deMaynadier, P. G. and Hunter, Jr, M. L. (1995). The relationship between forest manage-
ment and amphibian ecology: a review of the North American literature. Environmental
Review, 3, 230–61.
Dodd, Jr, C. K. (2005). Population manipulations. In M. Lannoo (ed.), Amphibian

Declines. The Conservation Status of United States Species, pp. 265–70. University of
Dodd, Jr, C. K. and Seigel, R. A. (1991). Relocation, repatriation, and translocation of
amphibians and reptiles: are they conservation strategies that work? Herpetologica, 47,
336–50.
Dodd, Jr, C. K., Loman, J., Cogalniceanu, D., and Puky, M. (2009). Monitoring amphib-
ian populations, in press. In H. H. Heatwole and J. W. Wilkenson (eds), Amphibian
Biology, vol. 9. Conservation and Decline of Amphibians. Surrey Beatty & Sons, Chipping
Norton, NSW.
Eggert, C. and Guyetant, R. (2002). Safeguard of a spadefoot toad (Pelobates fuscus)
population: a French experience. Atti del terzo Convegno “Salvaguardia Anfibi”, Lugano,
23–24 giugno 2000, Cogecstre Ediz. Penne, 2002, 47–52.
Fog, K. (1997). A survey of the results of pond projects for rare amphibians in Denmark.
Memoranda pro Societatis Fauna et Flora Fennica, 73, 91–100.
Gardner, T. A., Barlow, J., and Peres, C. A. (2007). Paradox, presumption and pitfalls in
conservation biology: the importance of habitat change for amphibians and reptiles.
Gent, T. and Gibson, S. (1998). Herpetofauna Worker’s Manual. Joint Nature Conservation
Committee, Peterborough.
Glandt, D., Schneewei, N., Geiger A., and Kronshage, A. (eds.) (2003). Beiträge zum
Technischen Amphibienschutz. Zeitschrift für Feldherpetologie, Supplement 2.
Griffiths, R. A. and Pavajeau, L. (2008). Captive breeding, reintroduction, and the con-
servation of amphibians. Conservation Biology, 22, 852–61.
Hayes, T. B. (2004). There is no denying this: defusing the confusion about atrazine.
BioScience, 54, 1138–49.
Hazell, D., Hero, J.-M., Lindenmayer, D., and Cunningham, R. (2004). A comparison
of constructed and natural habitat for frog conservation in an Australian agricultural
landscape. Biological Conservation, 119, 61–71.
Hels, T. and Fog, K. (1995). Does it help to restore ponds? A case of the tree frog (Hyla
arborea). Memoranda pro Societatis Fauna et Flora Fennica, 71, 93–5.
Kingsbury, B. and Gibson, J. (2002). Habitat management guidelines for amphibians and
reptiles of the Midwest. Technical Publication HMG-1. Partners for Amphibian and
Reptile Conservation, Montgomery, AL.
Kinne, O. (2004). Successful re-introduction of the newts Triturus cristatus and T. vulgaris.
Endangered Species Research, 4, 1–16.
Kraaijeveld-Smit, F. J. L., Griffiths, R. A., Moore, R. D., and Beebee, T. J. C. (2006).
Captive breeding and the fitness of reintroduced species: a test of the responses to preda-
tors in a threatened amphibian. Journal of Applied Ecology, 43, 360–5.
Kyek, M., Maletzky, A., and Achleitner, S. (2007). Large scale translocation and habitat
compensation of amphibian and reptile populations in the course of the redevelopment
of a waste disposal site. Zeitschrift für Feldherpetologie, 14, 175–90.
Langton, T. E. S. (ed.) (1989). Amphibians and Roads. Proceedings of the Toad Tunnel
Conference, Rendsburg, Federal Republic of Germany, 7–8 January 1989. ACO Polymer
Products, Shefford.
Lehtinen, R. M. and Galatowitsch, S. M. (2001). Colonization of restored wetlands by

amphibians in Minnesota. American Midland Naturalist, 145, 388–96.
Lindenmayer, D., Hobbs, R. J., Montague-Drake, R., Alexandra, J., Bennett, A.,
Burgman, M., Cale, P., Calhoun, A., Cramer, V., Cullen, P. et al. (2007). A checklist
for ecological management of landscapes for conservation. Ecology Letters, 10, 1–14.
Maerz, J. C., Blossey, B., and Nuzzo, V. (2005a). Green frogs show reduced foraging
success in habitats invaded by Japanese knotweed. Biodiversity & Conservation, 14,
2901–11.
Maerz, J. C., Brown, C. J., Chapin, C. T., and Blossey, B. (2005b). Can secondary com-
pounds of an invasive plant affect larval amphibians? Functional Ecology, 19, 970–5.
Mitchell, J. C., Breisch, A. R., and Buhlmann, K. A. (2006). Habitat management guide-
lines for amphibians and reptiles of the northeastern United States. Technical Publication
HMG-3. Partners for Amphibian and Reptile Conservation, Montgomery, AL.
Palis, J. G. (2007). If you build it, they will come: herpetofaunal colonization of con-
structed wetlands and adjacent terrestrial habitat in the Cache River drainage of south-
ern Illinois. Transactions of the Illinois State Academy of Science, 100, 177–89.
Pearl, C. A. and Bowerman, J. (2006). Observations of rapid colonization of constructed
ponds by western toads (Bufo boreas) in Oregon, USA. Western North American
Naturalist, 66, 397–401.
Pechmann, J. H. K., Estes, R. A., Scott, D. E., and Gibbons, J. W. (2001). Amphibian
colonization and use of ponds created for trial mitigation of wetland loss. Wetlands, 21,
93–111.
Percsy, C. (1995). Les Batraciens sur nos Routes. Ministère de la Région Wallonne. Brochure
technique no. 1. Service de la Conservation de la Nature et des Espaces verts,
Wallonne.
Petranka, J. W., Murray, S. S., and Kennedy, C. A. (2003). Responses of amphibians to
restoration of a southern Appalachian wetland: perturbations confound post-restoration
assessment. Wetlands, 23, 278–90.
Román, A. (2003). El ferreret, la gestión de una especie en estado crítico. Munibe, 16,
90–9.
Rubio, X. and Etxezarreta, J. (2003). Plan de reintroducción y seguimiento de la ranita
meridional (Hyla meridionalis) en Mendizorrotz (Gipuzkoa, País Vasco) (1998–2003).
Munibe, 16, 160–77.
Schwartze, M. (2002). Neuanlage und Verbesserungen von Kleingewässern für den
Laubfrosch und andere Amphibien – eine Untersuchung im östlichen Münsterland
(NRW). Zeitschrift für Feldherpetologie, 9, 61–73.
Scoccianti, C. (2001). Amphibia: aspetti di ecologia della conservazione/Amphibia: Aspects
of Conservation Ecology. WWF Italia, Sezione Toscana, Editore Guido Persichino
Grafica, Florence.
Seigel, R. A., Dinsmore, A., and Richter, S. C. (2006). Using well water to increase
hydroperiod as a management option for pond-breeding amphibians. Wildlife Society
Bulletin, 34, 1022–7.
Semlitsch, R. D. (2000a). Principles for management of aquatic-breeding amphibians.
Semlitsch, R. D. (2000b). Size does matter: the value of small isolated wetlands. National
Wetlands Newsletter, 22, 5–6, 13.
Semlitsch, R. D. (ed.) (2003). Amphibian Conservation. Smithsonian Books, Washington
DC.
Semlitsch, R. D. and Jensen, J. B. (2001). Core habitat, not buffer zone. National Wetlands
Newsletter, 23, 5–6, 11.
Smith, K. G. (2005). Effects of nonindigenous tadpoles on native tadpoles in Florida:
evidence of competition. Biological Conservation, 123, 433–41.
Stevens, C. E., Diamond, A. W., and Gabor, T. S. (2002). Anuran call surveys on small
wetlands in Prince Edward Island, Canada restored by dredging of sediments. Wetlands,
22, 90–9.
Trauth, J. B., Trauth, S. E., and Johnson, R. L. (2006). Best management practices and
drought combine to silence the Illinois Chorus Frog in Arkansas. Wildlife Society
Bulletin, 34, 514–18.
Weyrauch, S. L. and Amon, J. P. (2002). Relocation of amphibians to created seasonal
ponds in southwestern Ohio. Ecological Restoration, 20, 31–6.
Williams, L. R. (2005). Restoration of ponds in a landscape and changes in common frog
(Rana temporaria) populations, 1983–2005. Herpetological Bulletin, 94, 22–9.
Index
Figures and tables are indexed in bold. See under
abiotic factors 96, 152, 216, 218, 268, 285, agar models 370f20.1, 378, 390, 391, 392–3,
286, 300, 302–3, 304, 307, 309 394f21.2, 397f21.4, 399, 400,
abundance 24, 28, 55, 76, 79, 233f13.3, 287–8, 400f21.6, 401, 401f21.7
300, 302, 309, 321, 325, 330, see also models; physical models
334, 436, 438, 439, 440, 441, agile frogs 418
449, 458, 518 Agranat, I. D. 292
amphibians 143, 152, 233–4, 235–6, 240–1, agriculture 16, 185, 341, 511
248–50, 254, 255, 256, 259, 271, Aguirre, A. A. 267
307, 364, 437, 448–9, 452, 465 Aichinger, M. 271
estimation (counts) 287, 294, 329, 333, 432, Aide, T. M. 334
447, 448, 449, 450–2, 455, 456, Alford, R. A. 25, 143, 148, 157, 198, 207,
458, 465–6, 467, 468, 469, 470, 215, 367, 388, 390, 395, 400,
471–4, 476, 477; 436, 473, 477,
indices 235, 241, 288, 322 algae 16, 73, 76, 77–8, 79, 80, 82, 95, 107
relative 247, 248, 251, 259, 271, 288, 326, algal blooms/food 73, 307
452, 474, 475 alkalinity 110, 111, 308
sampling 451–2, 455 alleles 407, 414–15, 417, 418, 421
species 25, 60, 259, 281–2, 326, 327, Allen, C. R. 287
329, 346 Allmon, W. D. 250
transformation, effect of 322–3, 323t18.2 alpine newts 352
true 271, 467 alpine salamanders 414, 474–5
variation 441, 452, 467, 469 alternative hypotheses 22, 24
see also frogs, abundance; salamanders, Altig, R. 39, 40, 42, 49, 50, 51, 71, 72, 74,
abundance 82, 105
Acevedo, M. M. A. 189 Altwegg, R. 204, 207, 219, 222
ACI (amphibian calling index) 284, 288, 291 aluminium 111, 113, 309
acid rain 23, 110 alveolates 483, 486f26.1, 496, 497, 500, 522
acidity 517 Alytes 12, 45
Acris 11, 218 Alytes muletensis 512, 521
Acris crepitans 179, 284, 285f16.1 Ambystoma 43f3.2, 44, 151, 166, 216
blanchardi 179 Ambystoma laterale 353
Adams, M. H. 143, 145 Ambystoma macrodactylum 131, 420
Adams, M. J. 64, 65, 231, 442 Ambystoma maculatum 8, 43f3.2, 44, 100, 133,
adaptive variation 423–5 135–6, 145, 150f9.2, 151, 154–5,
Adenomera 14 197, 198, 348f19.5, 348f19.5, 352,
Adis, J. 175 470, 475
Adolf, E. F. 372 Ambystoma mexicanum 8, 44
Aeromonas 522 Ambystoma opacum 8, 135
Aeromonas hydrophila 284, 483, 486f26.1 Ambystoma talpoideum 44, 59f4.1, 217f12.4
AFDM (ash-free dry mass) 75, 76, 77 Ambystoma tigrinum 8, 106, 189, 196, 196f11.3,
AFLP (amplified fragment length 232, 353, 470, 472, 482, 489, 501
polymorphism) 411, 413, 415f22.1, Ambystomatidae 39, 42, 44, 45, 62, 77, 106,
416–17, 424 153, 220, 236, 314, 350f19.6, 424,
Africa 5, 7, 11, 13, 14, 15, 168, 250, 421 472, 473
African clawed frogs 10, 112 American bullfrogs 46f3.3, 106, 108, 132,
Afrixalus 13 511, 523
Agalychnis 12 American toads 215f12.3
530 | Index
Amezquita, A. 128 Archey’s frog 134

ammonia 114, 308 Arctic Circle 108
Amon, J. P. 514 area-based surveys 247–8, 248t14.1, 252, 257,
amphibian calling index, see ACI 259–60
Amphibian Conservation Action Plan 180 amphibian counts 252
amphibians 263, 509, 521 and weather conditions 256
conservation 341–2, 507, 510 design 254–9
management 507–8, 508t27.1, 509–11, plots 249, 250, 253, 254–5, 257f14.1,
516, 517, 519 259, 310
decline 15–16, 21, 180, 263, 295, 363, 364, quadrats 249, 252, 253, 254, 255, 256,
388, 436, 500, 509, 514, 521, 257f14.1, 258f14.2, 259, 270, 271,
522, 523 310, 312, 313, 314
specific causes 23, 203 transects 249, 253, 254–5, 257f14.1,
inventory 28, 263, 265–6, 287 258f14.2, 259, 271–2
taxonomy 273 nocturnal 251–2, 254
transport of 487, 493, 497, 501 vegetation 310
see also abundance, amphibians Arenophryne 15
Amphibiocystidium 522 Aresco, M. J. 235
Amphiuma 8–9, 137 Argentina 12f1.2, 14
Amphiumidae 7, 42–3, 234 arid habitats 108, 373
amplified fragment length polymorphism, see AFLP Arntzen, J. W. 30, 135, 188, 232, 236, 414
anesthesia 66, 129, 133, 137, 189–90, 381 ARS (automated recording systems) 281,
Anderson, D. R. 431, 432, 438, 439, 451, 458 288–94
Anderson, M. T. 109, 168 Arthroleptidae 49, 223
Anderson, S. H. 167 Arthroleptis 250
Andersson, G. 145, 147, 153, 157 arthropods 16, 167–8, 174
Andren, C. 111 seasonal changes 179–80
Andrias davidianus 8 Ascaphidae 145
Andrias japonicus 8 Ascaphus 45, 106, 421
Aneides aeneus 148–9 Ascaphus montanus 421
Aneides lugubris 10 Ascaphus truei 10, 286, 421, 424
Angerbjorn, A. 80 ash-free dry mass, see AFDM
Anholt, B. R. 449 Asia 5, 8, 11, 13
annelids 167 Assa darlingtoni 14
annuli 4, 40–1 Atelopus 11
Anodonthyla 13 Atelopus carbonerensis 128
Ansonia 11 Atelopus elegans 128
Anthony, B. P. 283 see also ARS; MCS
anthropogenic noise 183–4, 286–7 auditory monitoring 281, 283–4, 294–5, 438
ants 168, 169, 172, 173, 178 Augustin, N. H. 354
Anura 3, 4f1.1, 5, 40 Auld, J. R. 40, 51, 94
anurans, see frogs Aun, L. 174
Aplin, K. 512 Ausden, M. 174
aquatic mesocosms, see mesocosms Australasia 5
aquatic species 10, 196f11.3, 218, 234, 237, Australia 5, 10, 11, 14, 15, 147, 231, 282, 344,
250, 270, 311, 374, 375, 376 393, 395, 398f21.5, 400, 400f21.7,
breeding 251, 304 419, 423
habitats 304, 305t17.1, 307–9 automated recording systems, see ARS
larval stages 374, 382 autotrophs 79
see also terrestrial species
Araujo, C. 370, 378, 390 Babbitt, K. J. 252, 306, 307
Araújo, M. B. 344 Babik, W. 421
arboreal species 10, 12–13, 145, 174, 203, 217, bacteria 95, 113, 482–3, 500, 522
218, 373, 389 Baer, C. K. 492
Index | 531
BAF (Basal Area Fraction) 312 Berman, E. N. 391

Bailey, L L. 133, 135, 143, 254, 272, 439, 440, Bernstein, B. B. 24
441, 451, 453, 454, 455, 457, 459, Biebighauser, T. R. 514
467, 469, 470, 471, 472, 474, 477 Bieck, R. 203, 454
Bailey, M. A. 507 biodiversity 15, 180, 263–4, 265, 268, 269, 379
Baillie, B. R. 314 biomass 330
Baker, J. 123 biomes 203
Bakken, G. S. 389 biosecurity 481, 497, 522–3
Baldwin, C. 489 biotelemetry 123
Baldwin, J. A. 193 biotic factors 29, 96, 216, 218, 285, 300, 304,
Baldwin, R. F. 187, 188, 192, 193, 195, 197 307, 507
Balinsky, J. B. 112 biphasic life cycles 169, 341, 487
banding, see tagging birds 16, 292, 439
Bandoni de Oliviera, F. 310 cerulean warbler 292
Banks, M. S. 113 ivory-billed woodpeckers 289
Banning, J. L. 148 Black, J. H. 111
Barbeau, T. R. 402 Blair, W. F. 282, 286
Bardsley, L. 417 Blais, D. P. 187, 188, 189, 190, 192
Bardwell, L. V. 30 Blanc, M. 273
Bargelloni, L. 416 Blanchard’s cricket frogs 179
Barichivich, W. J. 64, 234, 289 Blaustein, A. R. 15, 363, 436, 487
Barr, G.E. 252, 307 Blaylock, L. A. 373
Bart, J. 439, 474 Bligh, E. G. 82
Bartelt, P. E. 372, 374, 378, 379, 389, 390, Blomquist, S. M. 193, 198, 199, 220
401, 402 Bloom, S. A. 322
Barton, K. 144 Blouin, M. S. 419
Basal Area Fraction, see BAF Bloxam, Q. M. C. 522
basking 366, 388, 399 bluegill sunfish 100
Batchelor, C. L. 311 body temperatures 402–3
Batie, R. D. 286 amphibians 5, 16, 108, 365, 366, 367, 370,
Batrachochytrium dendrobatidis 16, 50, 481, 483, 371, 376–7, 379, 387–8, 389, 390,
486f26.1, 488, 490, 493, 494, 496, 391, 393, 399
497, 499, 500, 522 instruments 370–1
Batrachoseps 112 spot measurements 367
Beals, E. W. 331 variation in 376, 387
Beard, K. H. 217, 221, 223 Bogert, C. M. 366, 367
Beattie, R. C. 517 Bolitoglossa 10
Beck, C. W. 216, 219, 220 Bombina 11, 45, 151
Becker, C. G. 23 Bombina bombina 152
Beckman, N. G. 206, 216, 222, 223 Bombinatoridae 45
Beebee, T. J. C. 206, 207, 209, 407, 412, 418, Bonham, C. D. 313, 314
420, 421, 423, 507, 518 Bonin, A. 424
beetles 179 Boone, M. D. 90, 92, 93, 94, 96, 97, 100, 204,
behavior 222–3, 365, 366, 377, 408, 423 205, 207, 219, 487
activities 366, 373, 376 Borass, M. 65
species 300, 367, 477 Bos, D. H. 424
Belgium 353 Bottrell, H. H. 75
Bell, B. D. 522 Boughton, R. G. 237, 238
Bellemakers, M. J. S. 516 Boulinier, T. 324, 466, 473, 476
Benke, A. C. 75 Boutilier, R. G. 374
Bennett, D. H. 230, 232 Boulton, A. J. 78
Bentley, P. J. 364 Bowen, S. H. 77
Bergallo, H. G. 174 Bowers, D. G. 283
Berkson, J. 216 box/pipe samplers 58–60
532 | Index
Bradfield, K. S. 134 Bufo boreas 46f3.3, 482

Bradford, D. F. 374, 378, 390 Bufo bufo 128, 236, 352
branding techniques 130, 130t8.4, 131 Bufo calamita 152, 353, 418, 472, 518
Brassil, T. 188, 189, 197 Bufo canorus 145, 419
Brattstrom, B. H. 364, 366 Bufo castaneoticus 11
Brauer, A. 41 Bufo fowleri 48f3.4, 128, 421
Bray-Curtis Similarity Index 331, 332f18.2, Bufo granulosus 128
333, 334 Bufo marinus 344, 400, 423, 498
Brazil 11, 15 Bufo terrestris 98
breeding 236, 421, 424, 441, 449, 454, 455, Bufo viridis 112
516, 519 Bufo woodhousii 288
amphibians 501, 511, 514 Bufonidae 14, 46, 47, 49, 153, 250, 372,
captive 521 389, 498
adults 421 Bulka, K. C. 251, 254
migrations 232 Bull, E. L. 131
phenology 152–3, 270, 471 bullfrogs 97, 100, 106, 113, 344, 371
seasons 514 American 46f3.3, 106, 108, 132, 511, 523
see also ponds, breeding Burggren, W. W. 105, 363, 382
Breitenbach, G. L. 155 Burnham, K. P. 450, 451, 452, 453, 454, 457,
Brem, F. 493 458, 476
Breviceps 15 Burns, E. L. 419
Bridges, A. S. 28, 289, 293 burrows 15, 108, 207, 209, 216, 223, 376, 517
Bridges, C. M. 100, 487 ampibians 108, 173, 220
Bridges-Britton, C. M. 93 capture 220
Briggs, J. L. 128 mammals 314
Britain 420, 423, 518 see also caecilians, burrows
Britto-Pereira, M. C. 168 Burt, W. H. 195
Brockelman, W. Y. 88 Bury, R. B. 145, 231, 270, 271, 437, 442
Brodkin, M. 111 Buttemer, W. A. 387, 389, 392, 402
Brodziewska, J. 521 Buzas, M. A. 431, 432, 433, 437, 439
bromeliads 13, 14, 55, 309 Byram, J. K. 381
Brook, B. W. 25
Brooks, P. D. 308 Cabe, P. R. 353
Brose, U. 326 caecilians 3, 4f1.1, 39, 73, 75, 76, 133, 137, 169,
Brower, J. E. 78, 175 220, 223, 288, 294, 364, 376
Brown, G. W. 441 aquatic 6–7
Brown, L. R. 67 burrows 4, 6, 7, 376
Brownie, C. 453 calling behaviors 288
Bruce, R. C. 455 combination land/aquatic 7
brumation 375, 379 diets 169
Brunner, J. L. 482, 488, 489, 495 ecology 223
Bryan, L. 499 eggs 6, 7
buccopharyngeal apparatus 71, 73 feeding 71, 72
Buchan, A. 132 field enclosures 216
Buck, J. C. 94 habitats 4, 6
Buckland, S. T. 269 gestation 6
Budge, S. M. 82 larvae 6, 50
Buech, R. R. 231 aquatic 39; size 73
Buergeria buergeri 198 larval 40–1, 41f3.1, 42, 71, 72, 74
buffer zones 515–16, 518 lifestyles 6–7
see also home range reproduction 6
Bufo 11, 98, 106, 151, 216, 237, 489 sampling methods 252
Bufo americanus 88, 100 tails 4
Bufo arenarum 4f1.1 terrestrial/fossorial 7
Index | 533
cages 65, 80f6.1, 88, 89f6.1, 96–7, 101, centrolenids 12–13

20612.1, 209 Ceratophrys 50
field 94, 96 Chalcraft, D. R. 207
ponds 96 Chalmers, R. J. 63, 143, 148
terrestrial 223 Chan-McLeod, A. C. A. 519
calcium 111, 112 Chandler, C. R. 237, 238
Caldwell, J. P. 49, 179 Channing, A. 42, 49
Calef, G.W. 64 Chao, A. 275, 333, 450
Calfee, R. D. 113 Chase, J. M. 105
Calhoun, A. J. K. 198 Chazal, A. C. 204, 206, 221, 222
call detection, see auditory monitoring Chazdon, R. L. 326, 331, 333
calling behaviors/patterns 285, 286, 288, 291, chemicals 94, 114, 116, 220, 307–8, 309,
291f16.2, 467–8, 512 381, 511
see also frogs, calling behaviors Chesson, J. 78
calling surveys 143, 151, 158, 270, 288, 437 Chester Jones, I. 375
Callulops 15 Child, T. 372, 374
Cam, E. 326 Chile 11, 14
Camp, C. D. 109 China 8
Campbell, H. W. 270, 271 Chirixalus eiffingeri 13
Campephilus principalis 280 Chlamydophila 498
Camper, J. D. 136, 137 Choquet, R. 453, 454
Canada 10, 108, 282, 286, 420 chorus, see calling behaviors
Canberra (dissimilarity) 331, 333 Christian, K. A. 389, 392
cane toads 112, 344 Christiansen, J. L. 179
Cano-Martinez, A. 44 Christman, S. P. 270, 271
Cao, Y. 175 Church, D. R. 449
canopy cover 100, 101, 209, 250, 251, 252, 303, chytrid/chytridiomycosis 16, 31, 388, 401,
307, 310, 311–13, 512, 515, 516, 424, 523
517, 518 Citrobacter 522
captive-reared animals 521–2, 523 Citron-Pousty, S. 354
capture 221–2, 229, 230, 235, 236, 237, 240–1, CJS (Cormack-Jolly-Seber) model 452
270, 327–8, 331, 449, 451, 452, 453, Clark, V. C. 178
467, 469, 472, 477 Clarke, K. R. 321, 330, 331, 334
capture-mark-recapture 221–2, 449, 450, 451, Clarke, L. B. 251, 254
452–3, 456, 458–9, 467 Clarke, R. D. 128
capture-recapture sampling 154, 439–40, Clausen, J. 99
441, 476 climate change 344, 364, 377, 380, 436
carbon 78–9, 80, 81, 309 global 15, 35, 403
Cardona, L. 176 clones 416
Carey, C. 497 closed-population models 449–50, 451, 452, 455
carnivores 50, 71, 75, 167–8, 169 cluster dendrograms 329–30, 331, 332f18.2
Carpenter, S. R. 87, 90, 113 clutches 5, 11, 13, 144, 149, 150–1, 152, 153,
carpenter’s frogs 111 156, 219, 222
Cascades frogs 419 coarse woody debris, see CWD
Cashins, S. 490 coastal tailed frog 442
Caswell, H. 453, 454 Cochran, W. G. 259, 432, 448
cations 112, 309 Coddington, J. A. 326, 331
cattle tanks 88, 89f6.2, 90, 93–4, 98–9, 101 Cohen, M. P. 207
Caudata, see salamanders Colberg, M. E. 189, 190
cause-effect relationships 88, 90, 96 Collembola 171, 174
Cecala, K. K. 189, 190 collaboration 266–7, 423, 459
census techniques, see traps, active Colli, G. R. 173
Central America 11, 145, 283 Collins, J. P. 16, 487, 488, 489, 501
Central Europe 339 Collins, J. T. 108
534 | Index
Collura, R. V. 273 Costa Rica 13, 26

Colombia 15, 4219–20 Cottam, G. 310
colonization 88–9, 356, 412, 418, 448, 457–8, Couch’s spadefoot toads 108
514, 518, 514, 518 counts 329, 331, 334, 439, 440–1, 473, 477
coloration 42, 44, 49 frogs 467f25.1, 468
Colostethus stepheni 469 tree 468t25.2
Colwell, R. K. 275, 324, 326, 331, 333 salamanders 465, 474
commercial over-exploitation 15 air humidity 475f25.2
common frog 180, 423, 440 sample units 310–11, 313, 433, 435, 442,
common toads 352 447, 448
community 340f19.1 variation in 447–8, 466, 466f25.1, 467, 474
composition 334 see also abundance, estimation
diversity 324 coverboards 229, 236–8, 238f13.4, 238–40,
ecology 98–9, 247, 248, 334, 377 270, 467
structures 27, 89–90, 91, 93, 98, 99, 321 Covich, A. P. 73
competition 93, 94, 98, 99, 100 Cowles, R. B. 366, 367
competitors 51, 65, 98, 105, 180, 207, 218 Cowman, D. F. 114, 115
Compton, B. W. 350, 351, 352 crab-eating frogs 112
Conant, R. 108 Cramer, V. 507
conceptual models 26–7, 29 Crawford, J. A. 109
Condit, R. 472 crayfish traps 64
conductivity, see water, conductivity crested newts 419
Congdon, J. D. 216, 220 Crinia georgiana 11
connectivity 349–50, 351f19.6, 352, 353, 420 Crinia victoriana 147
Connell, J. 111 Crisman, T. L. 63
conservation 299, 334, 339–40, 364, 388, 420, Crist, T. O. 348
447, 452, 510 critical landscape thresholds 348–9
amphibians 229–30, 248, 379, 448, 458, 459, Crook, A. C. 189, 190, 191
465, 481, 501, 517, 522, 523 Croshaw, D. A. 205
mesocosms 91–2 Crouch, W. B. 144, 150, 152, 153, 154, 155
planning 264, 340 Crump, M. L. 6, 270, 271, 272, 274
strategies 167 cryptic species 407, 417
diets, information 180 Cryptobatrachus 14
toe clipping 130 cryptobranchids 7, 8, 187
conservation biology 156, 203, 274, 321 Cryptobranchus alleganiensis 8, 131, 381
contaminants 15, 100, 105, 114, 222, 380, 487, Cuba 511, 512
489, 497 Cuban treefrogs 511, 512
convenience sampling 432, 435–6, 437, Cuello, M. E. 174
438, 442 Culver, D. A. 75
Converse, K. A. 482, 483, 486, 487 culverts 520, 520f27.2, 521
Conyers, M. K. 314 Cummins, K. W. 76
Cooke, A.S. 145 Curtis, J. T. 310
Cooper, A. G. 147 Cushman, S. A. 356
Cooperman, M. D. 221, 222 cutaneous evaporative water loss 372–3, 375,
Cooperrider, A. Y. 16, 17 389, 390, 392
Cophixalus ornatus 156 see also EWL
core habitat 514–16 CWD (coarse woody debris) 313–14
Cormack, R. M. 452 Cycloramphus stejnegeri 15
Cormack-Jolly-Seber model, see CJS
Corn, P. S. 231, 232, 234, 270, 271, 288, 290, da Rosa, I. 174
291, 292, 294, 433, 436, 437, Dalerum, F. 80
440, 441 Daly, J. W. 178
corridors 519 DAPTF (Declining Amphibian Populations
Corser, J. D. 143, 144, 148, 151, 156 Task Force) 267
Index | 535
Daszak, P. 16, 481, 482, 488 developmental plasticity 366, 381

data: Devineau, O. 459, 476
collection 24, 26, 27, 273–4, 288, 293, 300, DeWoody, J. A. 424
352–3, 401, 431–2, 459, 472, 476 Dexter, R. E. 128
and computers 339 Diamond, J. 90
ecological 333–4 Diamond, S. A. 113
management 267–8, 274 diets 71–2, 74, 79, 168, 169, 205, 334
Daugherty, C. H. 128 assimilatory 78–9
Davey, B. G. 314 isotopic memory 80
Davidson, C. 115, 354, 487 larvae 73
Davidson, C. H. 376 life history 167
Davis, A. K. 483, 486, 488, 522 ontogony 179
Davis, T. M. 129, 133 reptiles 169
Dayton, G. H. 314 see also frogs, diets; gut contents
Dayton, G. S. 65, 66 digestion 74, 76, 167, 171, 173, 366, 371
de la Pena, L. D. 490 see also gut contents
de Solla, S. R. 282, 324 DiGIR (Distributed Generic Information
Declining Amphibian Populations Task Force, Retrieval) 268
see DAPTF dip-nets 56, 57, 60–1, 63, 64, 143
Deegan, L. 131 Discoglossus 11, 45, 151
deformations 496–7; see also malformations diseases 29, 263, 267, 364, 376, 380, 388, 481,
Degraaf, R. M. 237 482, 483, 487–90, 494–5, 497–9,
dehydration 372, 375–6, 377, 379 500–1, 521–3
deMaynadier, P. G. 517 amphibian chytridiomycosis 388
Dendrobates tinctorius 421 detection 273
Dendrobatidae 11, 50, 498 diagnostics 493–4
Dendrocia cerulea 292 eggs/embryos 483t26.1
Denmark 35 infectious 15–16, 31, 482, 488, 497, 498, 500
Denoël, M. 349 larval 284t26.2, 484t26.2
Denton, J. S. 206, 207 notifiable 501
Dermestes maculatus 496 post-metamorphic 285t26.3, 485t26.3
dermestid beetles 496 prevalence 491t26.5
deserts 5, 66, 108, 153, 203, 314, 363, 367, susceptibility 366
376, 511 tadpoles 486f26.1
Desmognathus 9, 44 testing 488–90, 494, 495t26.6, 496
Desmognathus aeneus 9 molecular 495
Desmognathus fuscus 129 treatment 500
Desmognathus monticola 218 zoonotic 490, 498
Desrochers, A. 351 dispersal abilities 21, 25, 26, 28, 341, 342f19.2,
detection: 272, 287 346, 349, 351, 352, 356
ARS 293 dissimilarity 330, 331, 352
imperfect 284–5, 465 dissolved oxygen, see oxygen, dissolved
rates 253, 259 distance sampling 253, 273, 310, 468, 472
and temperature 86 Distributed Generic Information Retrieval, see DiGIR
detection probabilities 323–4, 439–40, 441, 442, distribution 458, 465
448, 456, 457, 458–9, 466f25.1, of organisms 447
467, 468–9, 470, 472–4, 476, 477 diversity, see species, diversity
covariates 154, 282, 284, 451, 474 Dixon, J. R. 136, 137
cumulative 468, 473 DM (dry mass) 75, 76, 77, 81
per visit 468–9, 472–3 DNA analyses 171, 273, 353, 408, 409, 413–14
variation in 474, 475 extraction 409–11
see also frogs, detection probabilities; fingerprints 423
imperfect detection sequences 408, 409, 411, 412, 413, 415, 417,
detritivores 71–2 421, 423
536 | Index
DOC (dissolved organic carbon) 113, 114 timescales 24–5

Dodd, C. K., Jr. 143, 144, 148, 151, 232, 236, ecological guilds 173
253, 326, 327, 331, 439, 440, 509, ecological niche modelling 268–9
521, 522 ecophysiology 365
Dolmen, D. 111 ecosystems 16, 216, 501
Dominican Republic 269 ecotoxicology 100, 115, 116
Donnelly, M. A. 123, 271, 334 ectotherms/y 16, 40, 286, 302, 364, 366, 371,
Doody, J. S. 135, 153 376, 387, 388
Dorazio, R. M. 288, 439, 448, 457, 459, 465, Ecuador 12f1.2
470, 472, 477 Edwards, B. R. 375
Dorcas, M. E. 234, 288, 289, 292, 293 efts 9
Dormann, C. F. 354 red 131
dorsal fins 6, 42, 43–4, 45 salamanders 45
dorsal patterns 32, 124, 135, 220 Egan, R. S. 145, 155, 157
dorsal pouches 13, 14 Egeland, L. M. 231
Downs, F. L. 43f3.2 egg masses 145, 147t9.2, 441, 449
dragonflies 93, 175 coelemic 191
Drake, D. L. 51 counts 143–5, 147, 151–2, 153, 154, 157–8, 440
Drake, J. A. 90 variability in 154–5
drift fences 143, 147, 150, 197, 198, 229, 230–1, marking 155
234, 235, 236, 270, 467, 469–70, egg-laying species, see oviparous species
473, 474, 477, 520f27.2 eggs 3, 5–6, 15, 40, 191, 219, 374
construction 232, 233f13.3 acquatic 15
terrestrial 231f13.2, 233–4 see also caecilians, eggs; frogs, eggs;
dripnets 288 salamanders, eggs;
Droege, S. 63, 285, 286, 439, 440 toads, eggs
Drosophila 179 Ehrlich, A. H. 15
Drost, C. A. 237 Ehrlich, P. R. 15
dry mass, see DM Eigenbrod, F. 438
Duellman, W. E. 5, 6, 40, 44, 49, 55, 73, 168, Elton, C. 447
170, 174, 252, 285, 325, 365 Eleutherodactylus 14, 223, 249–50, 468
Dünker, N. 41, 41f3.1, 73 Eleutherodactylus bransfordii 26
Dumont, H. J. 75 Eleutherodactylus coqui 149, 151, 292
Dunson, W. A. 29, 87, 88, 90, 100, 111, 112, Eleutherodactylus jasperi 14
113, 222 Elliot, S. L. 387
Dupré, R. K. 366 Ellison, A. M. 268
dusky salamanders 129 Elmberg, J. 153
dwarf squeaker 353 embryos 5, 14, 156, 382, 436, 482, 496
Dyer, W. J. 82 counting 156
Dzialowski, E. M. 389, 398, 399, 402 development 41, 111, 155, 517
and pH 111
Eagle, P. 285, 286 see also salamanders, embryos
earthworms 169 Emerson, S. B. 168
Eastern Europe 339 emigration 452–3, 454, 455
Eaton, A. D. 105 temporary 470, 472
ecology 22, 30, 72, 87, 88, 90, 101, 116, Emlen, S. T. 132
167, 223, 321, 330, 356, 363, enclosures, see field enclosures
376, 380, 431, 436, 450, 453, endocrine system 382
458, 459, 481, 510 energetics 365, 366, 375–6, 377
field studies 31–2 energy:
and landscape 339 acquisition 387
and larvae, roles of 82 flow 16, 167, 321
mesocosms 91–2 use 403
processes 25 and water 376
terrestrial 203 Enge, K. M. 231, 232, 233, 234, 236
Index | 537
environment 151, 363 Farrand, L. 189, 192, 193, 195, 197, 199
adaptation to 364 fatty acids 74, 75, 78, 81–2, 177–8
and ARS 288–9 fauna 272, 275
changing 364, 380 Fauth, J. E. 29
conditions 105, 273, 293, 295, 389 fecundity 221–2, 521
stressors 203 Feder, M. E. 105, 106, 363
variables 452 feeding patterns 168, 71–2, 167, 170, 179, 375
Ernst, R. 143, 270, 271 tadpoles 72–3
Escherichia coli 498 Fellers, G. M. 237, 267
Estoup, A. 423 Felton, A. 156
EstimateS 331, 333 Ferner, J. W. 32, 123, 130, 132t8.2, 136t8.7
estivation 5, 8, 108, 374 fertilizers 100, 114, 518
Eterovick, P. C. 174 Fiala, A. C. S. 311, 312
ethanol 74, 137, 170, 171–2, 173, 174, 178, Ficetola, G. F. 144, 273, 344, 349, 418
189, 191, 273, 410, 492 field cages 94, 97, 209
ethics 123 field enclosures 203–4, 207, 215f12.3, 216, 223
and toe clipping 129–30 construction 208–16
use of formalin 174 escape from 216–18
Etxezarreta, J. 522 response metrics 221–3
Euclidean metrics 330, 333, 350, 353 study species 218–19
Eupsophus 11 field experiments 204, 351, 381–2
Eurasia 5, 8 field studies 21–2, 24–6, 32, 91, 101, 123, 124,
Europe 8, 9, 11–12, 15, 130, 144, 145, 153, 180, 230, 274, 365
419, 421, 423, 437, 467, 460t25.2, objectives 30–1
468t25.2, 469, 472 SMART approach 27–9
Eurycea bislineata 133, 135 Finlay, J. C. 79, 81
Eurycea longicauda 133 fish 93, 177, 365, 482, 490, 514, 523
Eurycea lucifuga 133 fish 511
euthanasia of study animals 74, 169, 512 lobe-finned 3
eutrophication 16 Fitch, H. S. 366
Evans, B. J. 421 Fitzgerald, L. E. 314
Evans, C. J. 193 flagship species 26
Evans-White, M. A. 76, 77 Flectonotus 13
evenness 329; see also species, diversity foam nests 11, 12f1.2, 13, 14, 145
evolution 17, 156 Fogarty, J. H. 253, 273
current theory 21 Foltz, K. D. 292
mesocosms 91–2 food 71
evolutionary studies 99 collection of 73
EWL (evaporative water loss) 387, 389, 390–1, ingestion of 74
396, 401 organic carbons 113
physical models 399 food webs 29, 79, 80–1, 82, 93, 167, 178, 321,
see also cutaneous evaporative water loss 380, 487
experiments 29, 29f2.2, 30, 204, 207, 209, 218, aquatic 71, 96, 307
220–2, 223 freshwater 78
area-based surveys 247 natural 90
explosive-breeding species 150–1, 152, 153, 290 stream 79–80
extinction 15–16, 71, 448, 458, 481, 483, foraging activity/patterns 71, 72, 74, 75, 169,
521, 523 171, 172–3, 179, 519
extrapolation 326, 475 Ford, E. D. 22
forest 15, 45, 116, 203, 206, 247, 251, 349
F statistics 419, 424 counts 257, 257f14.1, 258f14.2, 259
Faber, J. 49 cover 347
Faccio, S. D. 189, 192, 199 habitats 341
Fahrig, L. 346, 349, 438 management 91, 100–1
Farnsworth, E. J. 31 tropical 175, 249, 250
538 | Index
forest-floor dwelling species 468 movements 195, 199, 341

forestry 516, 518 physical models 391–2, 393, 396f21.3,
formalin 74, 171, 173, 174, 400, 492, 494 397–8, 402
Forson, D. D. 482 rainforest 198, 388
Fortin, M.-J. 354 reproduction 12f1.2, 13–14, 198, 288
fossil fuels 110, 116 sampling methods 170, 172, 187, 188, 233,
fossorial species 15, 203, 216, 236, 373, 376 254, 379
traps 220 skin 16, 45, 47, 130, 132, 187, 188, 189, 190
underground retreats 314 snail-eating 168
see also toads, fossorial species stream-adapted 11
Fotheringham, A. S. 354 tadpoles 11, 12f1.2, 13, 14, 40, 45–6, 46f3.3,
Fouquet, A. 273 47–50, 72–3, 98, 288
four-toed salamanders 148, 149f9.1, 151 tagging 132–3
France 352 tails 5
Frankham, R. 25 terrestrial 223
Fraser, D. F. 169 toe clipping 125–9, 130, 138
Freda, J. 110, 111, 113 tropical 252, 518
free-living/range amphibians 198, 366, xerophilic 373
375, 377 see also under individual species
diseases 491t26.5 Fronzuto, J. 64
Freed, A. N. 167 fruit flies 179
Freel, K. L. 267 Fry, B. 78, 79
Freidenburg, L. K. 58 FTA cards 410, 411
freshwater habitats 71, 73, 77, 78, 79 Fujiwara, M. 453
Fretey, T. 449 Fukuyama, K. 188, 197, 198, 199
Friedl, T. W. P. 472 fungi 148, 481, 482, 483, 522
Friend, G. R. 230, 231, 236 fungicides 114
frogs 3, 6, 10, 16, 112, 156, 286, 363, 364, Funk, W. C. 33, 251, 253, 273, 420, 468, 469,
401, 469 473, 477
abundance 178, 180, 236 funnel traps, see traps, funnel
arboreal species 10, 12–13, 373, 389
auditory monitoring 281, 283–4, 294 Gabrial, M. W. 134
body temperature 252, 370f20.1, 391, 393–8, Gadgil, M. 269
398f21.5, 402 Gagné, S. 346
branding techniques 130 Gallant, A. L. 16
breeding 143, 152, 237, 284–5, Gamble, L. R. 340
285f16.1, 286 Garcia, T. S. 94
seasonal 153, 293 Gardenfors, U. 447
burrows 10, 314 Garland, H. O. 375
capture methods 237, 239 Garner, T. W. J. 31
calling behaviours 292–4, 467, 470 Garrott, R. A. 193
detection probabilities 285f16.1, 285, 286, Garshelis, D. L. 300, 301
288, 292, 469 Gascon, C. 93, 180, 237
diets 168, 169, 170, 179, 180 gastric pills 188–9, 199
eggs 10, 11, 12f1.2, 13, 14, 145, 146t9.1, Gastrophryne 237
150f9.2, 288 Gastrophryne carolinensis 217f12.4
and EWL 402 Gastrotheca 13, 14
extinction 16 Gates, D. M. 389
forest 251 Gatten, R. E. 106
larvae 6, 11, 39, 40 Gaucher, P. 421
die-offs 482 Gegeneophis ramaswamii 133
metamorphosis 288, 442 gene flow 351, 352, 419
monitoring programs 281–2, 287, 288, 289, generality 31, 204, 208
292, 294 genes coding 412
Index | 539
Genet, K. S. 287 Gray, M. J. 197, 482, 487 489, 498

genetic coding 497 gray treefrogs 100–1, 237
genetic drift 419, 421 Green, D. E. 482, 483, 486, 487
genetics 21, 340f19.1, 407, 408, 419, 420, 423, Green, D. M. 340, 421, 436, 441, 442
425, 522 green frogs 187, 195
diversity 407, 412, 417–18 green salamanders 148–9
see also landscape genetics green toads 112
genomes 413, 414, 416, 424 Greenberg, C. H. 231, 233
genotypes 99, 417, 419, 420, 421, 423 Greenlees, M. J. 218, 221
Gent, T. 123, 507 Greer, A. L. 488
Geocrinia 15 Gregory, R. D. 253
Geocrinia alba 147 Griffis-Kyle, K. L. 94, 307
gestation 14 Grillitsch, B. 115
Ghana 265 Groom, M. 1, 16
Ghioca, D. M. 64, 65 Grosse, W.-R. 134
Gibbons, J. W. 349, 473 growth rates 215f12.3, 219, 222, 387, 453
Gibbs, J. P. 255, 346, 349, 353 Gu, W. 457
Gibson, J. 507 Guisan, A. 346, 347
Gibson, S. 507 Gunzburger, M. S. 60, 97
Gill, D. E. 340, 466 Guppy, M. 375
gills 8, 39, 40–1, 42, 45, 49 gut contents 74–8, 371–2
aquatic aerobic organisms 105–6 see also digestion
Giordano, A. R. 420 Gutleb, A. G. 490
GIS (Geographic Information Systems) 239, Gutzwiller, K. J. 287
264, 269, 306, 339, 340, 341, Guyana 14
342–5, 345t19.1, 351, 352, 353, Gymnophiona, see caecilians
354–5, 356, 433, 438 Gyrinophilus porphyriticus 66
Gittins, S. P. 231, 232
Gladfelter, W. B. 176 habitat matrix 341, 349, 350, 351
Glandt, D. 521 habitat selection 299, 300, 301–2, 344, 351
global warming 388 habitats 25–6, 51, 63, 116, 205, 207, 221, 222,
Goicochea, M. A. 239, 251 232, 314–15, 325, 329, 340f19.1,
Goldberg, C. S. 187, 190, 273 344, 364, 400, 419, 433, 434, 436,
Goldfarb, L. 24 438, 508, 510–11
golf courses 91, 96 alteration 147–8, 348–9, 508, 515–16, 523
Golon, J. 77 arboreal 175
Gomez-Mestre, I. 112 complexity 25, 440, 519
Gonser, R. A. 273 contiguous 517
Gooch, M. M. 284, 287 degradation/destruction 251, 256, 299, 356,
Gordon, M. S. 364 516, 518
Gorley, R. M. 330, 334 fragmentation 100, 347, 351, 352, 420, 517,
Goslee, S. C. 310 519, 523
Gosner, K. L. 111 management 513t27.2, 514
Gotelli, N. J. 268, 275, 324 measurement of 299
Gosner, K. L. 40, 49 modification/destruction 15–16, 167, 247
Gower, D. J. 133, 252 by human activity 21
GPS (global positioning systems) 193, 264, 267, occupancy 438, 439, 456
271, 273, 306, 339 protection 508–9, 522
Grant, B. W. 237 relationships 144, 247, 457
Grant, E. H. C. 66, 135, 136, 154, 155, 449, riparian 80
470, 475 preference 299–300
Grant, E. M. 440 stream 62–3
grasslands 203, 517, 519 structure 157–8, 379
Graves, B. M. 288 terrestrial 203, 230f13.1, 517
540 | Index
habitats (cont.) herbivores 71–2

types 65, 205–6, 207, 221, 223, 266, 274–5, Héron-Royer, L. F. 48f3.4
331, 424 herpesvirus 482, 488, 494
use 5, 185, 197, 199, 218, 299, 200, 300–2, herpetofauna 334
321, 341, 355, 376 herpetology 148, 154, 157, 169, 232, 236, 271,
variables 452 272–3, 364, 381, 465, 475, 481
Haddad, C. F. B. 6 Herrmann, H. L. 346
Hagstrom, T. 135 Hershey, A. E. 79
Haideotriton 8 Hertz, P. E. 371
Hairston, N. G. 31, 87, 204, 247, 271, 272, heterogeneity 327, 333, 451, 455, 473
465, 469, 474 heterozygotes 416, 416f22.2, 417, 418, 521
Halffter, G. 327 Heunisch, G. 134
Hall, J. A. 46 Heyer, W. R. 73, 143, 185, 186, 188, 193, 197,
Hall, R. O. 80 240, 269, 270, 273, 327
Halliday, T. 22, 23, 128 Hickling, G. J. 498
Hammerson, G. A. 387, 388, 399 Higgins, K. F. 310, 313
Hammond, P. 3 Hilborn, R. 431
Handrigan, G. R. 45 Hine, R. L. 231, 233
Hanski, I. 350 Hirai, T. 173
haphazard surveys 247–8 histology 488, 491, 492, 494, 496
haplotypes 417, 421, 422f22.5 Hobbs, N. T. 431
Hardy-Weinberg equilibrium 417–18 Hobel, G. 148
harmonic radar 185, 220 Hocking, D. J. 100, 101
Harp, E. M. 489 hogs 511
Harper, E. B. 205, 207, 209, 217, 218, 222 Holbrook, C. T. 143
Harris, R. N. 58, 65, 143, 151, 156 Homan, R. N. 349
Harrison, R. G. 44 home range 195, 197, 209, 215, 299, 301,
Hart, R. K. 177 302, 514
Hartel, T. 144, 157 homozygotes 416, 416f22.2, 417
Hasegawa, M. 378, 391 Hopkins, W. A. 222
hatchings 7, 40, 41f3.1, 43, 44, 57, 49, 99, Hoplobatrachus occipitalis 197
156–7, 442 Horn 331
Hauer, F. R. 73, 80 Horne, M. T. 112, 113
Hawksely-Lescauly, D. S. 314 Hossack, B. 441
Hayek, L.-A. 24, 55, 265, 431, 432, 437, 439, 440 Hou, P.-C. L. 334
Hayes, J. P. 73 Houlahan, J. E. 157
Hayes, T. B. 512 Houze, C. M. 237, 238
health: Howe, C. M. 307
assessments 488, 493, 497 Howell, K. M. 270
captive amphibians 489–90 Hsu, M. Y. 231
human 497 Huang, C.-Y. 334
Healy, W. R. 135 Huggett, A. J. 355
heat exchange/transfer 365, 367, 370, 376, 377 Huggins, R. M. 450
Hecnar, S. J. 346 Hulbert, S. H. 97, 177, 206, 207, 435
hellbenders 8, 187, 188–9, 381 humans 25
helminthes 487 constraints 509
Hels, T. 340 effects on amphibians 16
Hemidactylium scutatum 9, 148, 149f9.1, humidity 303, 309, 388, 475
155–6, 309 air 474–5, 475f25.2
Hemiphractus 14 Hunter, C. M. 454
Hemisis marmoratum 13 Hunter, M. L. 193, 198, 199, 517
Henderson, I. W. 375 Hurlbert, S. H. 97, 206, 207
Henle, K. 123 Hutchison, V. H. 364, 366
herbicides 114, 514, 518 hybridization 416
Index | 541
hybrids 99, 208f12.2 infectious diseases, see diseases, infectious

Hyde, E. J. 251, 439, 440, 465, 474 Inger, R. F. 250, 254
hydration 232, 364, 372, 377, 379 Inkley, D. B. 282
hydroperiod 510, 511, 514 inoculum 96
hydroregulation 366, 376 insecticides 100, 114–15, 175
Hyla 11, 12 insects 3, 16, 176
Hyla andersoni 111 nets 175
Hyla arborea 286, 437, 467, 472, 469 International Union for the Conservation of
Hyla boans 13 Nature 180
Hyla bokermanni 12f1.2 Internet 339
Hyla chrysoscelis 48f3.4, 237 interpolation 325, 326
Hyla cinerea 291f16.2 interspecific competition 22, 98
Hyla gratiosa 291f6.2 introduced species 15
Hyla labialis 128 integument 364, 372
Hyla leucophyllata 50 inventory 28, 239, 263, 264, 254–6, 272, 283,
Hyla pseudopuma 153 287, 325, 326, 327, 432, 433,
Hyla rosenbergi 13, 158 435, 442
Hyla versicolor 100, 152, 197 ecology 269
Hylidae 14, 46, 47, 50, 71, 132, 145, 168, 233, invertebrates 92, 116, 169, 217, 514
236, 237, 238, 239 Ireland, D. H. 94, 137
Hymenochirus boettgeri 168 Ireland, P. H. 65
Hymenoptera 172 IRI (index of relative importance) 177
Hynobiidae 42, 45 Irwin, L. L. 312
Hynobius nigrescens 44 isotopes:
hyperoliids 13 analyses 81–2, 380
Hyperolius 270 composition 78, 177–8
Hypogeophis rostratus 4f1.1 signatures 74, 74, 178
hypoxia 106 Italy 418, 418f22.3
hypotheses 21, 22, 24, 88, 97, 204, 234, 330, IUCN (International Union for the
334, 365, 377, 380, 431, 435, 442, Conservation of Nature) 15, 263,
438, 451, 455–6, 457–8 265, 275
Ivlev, V. S. 78
Ichthyophiidae 7, 42 Iwasawa, H. I. 44
Ichthyophis banannicus 41f3.1
Ichthyophis glutinosus 41f3.1 Jaccard 330, 333
Ichthyophis kohtaoensis 41f3.1, 169 Jackson, D. C. 308
Ichthyophonus 496, 497, 522 Jacobs, J. F. 176, 273
immigration 439, 450, 452, 453, 455, 511 Jaeger, R. G. 22, 87, 254
immune systems 111, 366, 494 Jamaica 13
imperfect detection 143, 282, 284, 457, 458, James, F. C. 21
465–7, 469, 470, 471–5, 476, 477 James, S. M. 90, 92, 93, 97, 100, 204
see also detection probabilities Japan 8
inbreeding 407, 408, 417, 418 Jared, C. 373, 375, 389, 390
index of relative importance, see IRI Jefferson salamanders 111
Indermaur, L. B. 197, 199 Jehle, R. 188, 231, 235, 414, 419
India 7, 133, 252, 253 Jensen, J. B. 514, 516
individuals 274, 330 Jie, Z. 309
fitness of 407 Jofre, M. B. 307
precise measurements 402 Johannson, M. 433, 440
inference 24, 87, 101, 204, 247, 249, 259, 269, Johnson, B. K. 179
283, 284–5, 300, 365, 413, 423, Johnson, D. H. 31, 286, 299, 439
431–2, 433, 436, 437, 439, 442, Johnson, J. R. 169, 188, 191, 197, 238
447, 448, 451–2, 453, 456, 458, 459, Johnson, N. F. 172
472, 473, 477 Johnson, P. T. J. 105, 487, 489
542 | Index
Johnson, S. A. 64, 232, 234 188, 193, 198, 204, 235, 239, 308,
Johnson, W. S. 176 309, 373, 374, 381–2, 388, 489, 494
Johnston, G. F. 39, 40, 50 equipment 408–9
Jolly, G. M. 452 Lajtha, K. 380
Joly, P. 169, 346, 349, 351, 352 lakes 8, 10, 55, 62, 82, 90, 93, 113, 145
Jorgensen, C. B. 375 LaLiberte, G. D. 76, 77
Juncá, F. A. 174 Lamberti, G. A. 73, 80
Jung, R. E. 63, 65, 66, 440, 441, 449, 451, 455 Lamoureux, V. S. 187, 188, 195, 197, 199
Just, J. J. 382 Lampo, M. 267
Lancia, R. A. 447, 448
Karraker, N. E. 438 land acquisition 275
Kasier, K. 283 landscape 354–5, 519
Kam, Y.-C. 374 configuration 349, 356
Kaplan, H. M. 130, 131 habitats 514
Karasov, W. H. 307 permeability 351–2
Karraker, N. E. 112 perspective 340
Kayanne, H. 81 landscape ecology 339, 340, 340f19.1, 356
Keller, C. M. E. 438 amphibian conservation 341–2, 507,
Kendall, W. L. 453, 454, 455, 469, 470, 508t27.1
471, 474 population 346
Kennedy, A. D. 387, 388, 403 landscape genetics 352–3, 419
Kenya 353 landscape-as-population 340
keratin 44, 47, 49–50, 179, 364, 372, landscape-composition thresholds 349, 355–6
486f26.1, 496 Langton, T. E. S. 521
Kerr, J. T. 275 Lannoo, M. J. 42, 46, 47, 50, 51
Kéry, M. 441 Laposata, M. M. 222
keystone species 26, 98 larvae 5, 6, 9, 12, 39, 40, 50, 56, 59f4.1, 61, 522
Kie, J. G. 193 acquatic 10, 15, 56
Kiesecker, J. M. 87 diets 73
Kilham, S. S. 79 toxins 487
Kimball, K. D. 87, 100 see also caecilians, larvae; frogs, larvae;
Kimmel, C. A. 496 salamanders, larvae
Kingsbury, B. 507 larval development 5, 57, 96, 304, 382, 510, 512
Kinkead, K. E. 471, 472, 474 larval stages 6, 29, 58, 59f4.1, 65, 101, 207, 252,
Kinne, O. 522 306, 288, 435
Kirlin, M. 284, 286, 293 aquatic 5, 8, 203, 374
Klump, G. M. 472 field enclosures 219
Knapp, R. A. 354 growth rates 423–4
Knutson, M. G. 282, 350 larviform 39
Köhl, M. 30 lateral line system, see neuromasts
Kouamé, N. G. 265, 268, 269, 271, 273, 274 Lauber, A. 474
Koussoroplis, A. M. 82 Lauer, A. 380
Kraaijeveld-Smit, F. J. L. 525 Laugen, A. T. 424
Krebs, C. J. 21, 22, 26, 30, 78, 321, 330, Laurance, W. F. 477
333, 447 Lavilla-Pitogo, C. R. 490
Kriger, K. M. 493, 496 Layman, C. A. 81
Kroll, A. J. 433, 442 leaf litter 92f6.3, 95, 100, 174, 175, 217, 218,
Kupfer, A. 169 220, 249, 250, 272, 517
Kupferberg, S. 72 plots 249–50, 253
Kutrup, B. 180 quadrats 250
Kyriakopoulou-Sklavounou, P. 144 species 223, 250
surveys 250–1, 252
labial folds 41, 42 leaf litterbags 59f4.1, 62–3, 237
laboratory experiments 29, 88, 101, 116, 124, Lee, A. K. 364
Index | 543
Legendre, P. 347, 354 long-toed salamander 420, 420f22.4

legislative measures 508 Longcore, J. E. 488
Legler, J. M. 169 lotic conditions/habitats 71, 106, 107
Lehner, P. N. 24 Lotz, A. 287
Lehtinen, R. M. 50, 254 Loveridge, J. P. 373
Leiopelma archeyi 134 Lowe, R. L. 76, 77
Leiopelmatidae 45 Lowe, T. P. 111
Lemckert, F. L. 188, 189, 197 Lowe, W. H. 66
lentic conditions/habitats 71, 106, 107f7.1, Ludwig, P. M. 151
152, 433 Lüddecke, H. 128
leopard frogs 99, 197 Luhring, T. M. 207, 209, 216, 237
northern 108 Luikart, G. 418
southern 153 Lunde, K. B. 487, 489
Lepidobatrachus 45, 50, 71 lungs 106
Leptodactylidae 15, 45, 50, 71, 132, 250 Luscier, J. D. 313
Leptodactylinae 11, 145, 327
Leptodactylus bufonius 11, 12f1.2 McAllister, K. R. 132, 137, 187
Leptopelis 270 MacArthur, R. H. 177
Levin, S. A. 87, 100, 300, 310 McCallum, M. L. 263
Levins, R. 31 McCarthy, M. A. 32, 130, 334
Licht, P. 381 McCauley, D. E. 419
Lieberman, A. 339 McClanahan, L. L. 387, 389, 402
Liebgold, E. B. 423 McCollum, S. A. 106, 115
likelihood models, see models, likelihood M’Closkey, R. T. 346
Lillywhite, H. B. 364, 366, 371, 372, 373, 374, McCulloch, C. E. 21
375, 377, 379, 389, 402 McCullough, J. 265
Lima, A. P. 170, 179 McDiarmid, R. W. 40, 49, 50, 105, 273
Lind, A. J. 300 McDonough, C. 188, 191, 197, 199
Lindenmayer, D. 507 McIntyre, P. B. 106, 115
Linder, G. 115 MacKenzie, D. I. 157, 282, 283, 284, 285, 441,
Link, W. A. 439, 457, 473, 474 448, 456, 457, 458, 469, 474
Linsenmair, K. E. 180, 188, 191, 195, 197, 389 MacNally, R. 174
lipids 6, 81, 82, 221, 378, 389 McShea, W. J. 189, 196
barrier 373 Madagascar 4, 13, 178
storage 178, 222 Madison, D. M. 185, 188, 189, 191, 192, 193,
Lips, K. R. 143, 270, 380, 388, 481 195, 196, 197, 198, 199, 314
Litoria aurea 419 Maerz, J. C. 206, 207, 216, 218, 512
Litoria caerulea 392, 393–4, 394f21.2, 399 Magnuson, J. J. 377
Litoria lesueuri 148, 198, 395–6, 397f21.4, Magnusson, W. E. 170, 174, 179, 267
398f21.5, 401f21.7 Magurran, A. E. 275, 321, 324, 329, 330
Litoria littlejohni 147 Mahan, R. D. 169
Litoria rubella 399 Maher, W. A. 26
Litoria verreauxii alpina 147 Majji, S. 489
litterbags, see leaf litterbags major histocompatibility complex, see MHC
Little, M. 62, 63, 237 malformations 105, 113, 116, 487, 489, 492,
Littlejohn, M. J. 285 493, 496–7
Livia, L. 410 see also deformations
Llorente, B. J. 325 Mallorca 512
Loafman, P. 135 Manel, S. 352, 352
Löfvenhaft, K. 342 Maneyro, R. 174
Lowe, J. 189 Manhattan metrics 330
Loftin, C. S. 144, 148 Manly, B. F. J. 300
Loman, J. 94, 145, 147, 153, 157 Mantellidae 47
Londoño-M, M. C. 334 Mantidactylus lugubris 50
544 | Index
manual calling surveys, see MCS Merritt, R. W. 73

Mao, J. 497 Mertensophryne 49
marbled newts 419 mesocosms 87–91, 92f6.3, 101, 498
marbled salamanders 8, 205 laboratory studies 90–3, 204
marine toad 423 mesomycetozoa 522
mark-recapture 65–6, 130, 205, 222, 235, 241, metabolism 108, 371, 375, 380, 401
251, 253, 255, 256, 259, 288, 326, metals 112, 115
388, 468, 470, 472–3, 476 heavy 309
marking 123, 133, 219–20, 408, 412, 448–9 metamorphosis 39, 40, 41, 42, 44–5, 55, 56,
captured animals 235–6 216, 322, 381, 487, 489
criteria for 123–4, 138 see also frogs, metamorphosis; salamanders,
egg masses 155 metamorphosis
microchip 136 metadata 267–8
sources of materials 124, 124t8.1, 125 metapopulation 341, 349, 350, 408
see also pattern mapping; salamanders, landscape ecology 346
marking; toe clipping; Meteyer, C. U. 489, 496
trailing devices methylated mercury, see MeHg
Marsh, D. M. 143, 206, 216, 222, 223, 239, Mexico 8, 9
251, 254, 283, 340, 341, 351 Meyer, A. H. 24, 145, 153, 157
Marshall, J. L. 109 Meyer, F. 134
Martin, A. A. 147 MHC (major histocompatibility
Martin, M. M. 285 complex) 424–5
Martof, B. S. 125, 282 Michener, R. H. 380
Marzluff, J. M. 193 Michener, W. K. 30
Maser, C. 314 microcosms 88
Mathis, A. 168 microenvironment 58, 192, 366, 367, 372, 377,
Matori, R. 174 379, 387, 388, 390–1, 399, 402, 403
Matsui, M. 173 microhabitats 25, 31, 55, 109, 174, 179,
Matthews, K. R. 263 206f12.1, 207, 209, 215, 234, 236,
Mau, C. X. 333 238, 240, 251, 256, 259, 272, 301,
Mau Tau 333 309, 365, 371, 397, 400, 402
Mauel, M. J. 482 Microhylidae 11, 14, 15, 45, 49, 50, 71, 144–5,
May, J. T. 67 233, 235, 237
May, R. M. 32, 177, 324 microsatellites 414, 415, 416f22.2, 417–18, 419,
Mayhew, W. 379 420, 423, 424
Mazanti, L. E. 114, 115 migration 215, 222, 235, 341, 352, 356, 518
Mazerolle, M. J. 154, 235, 241, 272, 351, barriers to 346, 407–8, 419, 519–20
441, 457, 465, 467, 469, 472, Mikulicek, P. 417
474, 475, 476 Miles, L. 31
MCS (manual calling surveys) 281–5, 285f16.1, Miller, D. L. 487, 488, 489, 490
286–9, 291, 293, 294–5 Mills, L. S. 469, 477
Measey, G. J. 252, 353 Mills, N. E. 93, 106, 115, 307
measurements 138 Millspaugh, J. J. 193
ground (understory) 313–14 Mineau, P. 192
habitat use 299, 200, 301 minerals 265
shrub (midstory) 312–13 minnow traps, see traps, minnow
tree (overstory) 310–12 Miranda, T. 169
vegetation 309–10, 313 Mitchell, J. C. 231, 233, 271, 507
water 308 mites 171
weather 303 mitochondria 412
Megophryidae 45, 50 Mittal, A. K. 372, 373, 374
MeHg (methylated mercury) 113 MNKA, see survival, minimum number known
Menegon, M. 268, 270, 274, 334 alive
Merila, J. 416 models 451–3, 455
Index | 545
hierarchic 457, 458 movements 185, 195, 197, 199, 209, 215, 222–3,
likelihood 450, 451, 453, 454, 457, 458 303, 341, 356, 448, 453
multi-season 458 barriers to migration 346, 407–8, 419, 519
see also agar models; physical models between breeding sites 205, 516, 519
Mohr, J. R. 288 detection probabilities 452
moisture 209, 234, 237, 238, 239, 252, 309, effects of transmitters 198
310, 313, 314, 387, 402, 520, 521 patterns 342f19.2
environments 387, 388 see also frogs, movements; salamanders,
microenvironments 399 movements
soil 250, 251 MS-222 381
molecular ecology 407, 408, 423 mtDNA (mitochondrial DNA) 412–13, 417,
mollusks 167 419, 421, 422f22.5
monitoring programs/schemes 5, 55, 143, mud nests 12f1.2
145, 153, 158, 197, 239–40, 247, Mugglestone, M. A. 354
264, 275, 334, 388, 432, 433, Muller-Navarra, D. C. 82
434f23.1, 435, 473–4, 145, multistate models 453–4
152, 153, 197, 234, 247, 435, Murphey, T. G. 187, 197, 199
441, 481 muscles 106, 190, 366, 379, 496
design 239 mutation 412, 414, 417, 423
devices 368–9t20.1 Muths, E. 151, 292, 441, 456, 457, 459, 482
malformations 489 mycobacteria 482
see also frogs, monitoring programs Myers, N. 15
Monsen, K. J. 419 Mycobacterium lifl andii 482, 498
Moore, G. E. 339 Mycobacterium ulcerans 482
Moore, J. 363 myobatrachids 11, 151
Moore, J. D. 237, 238
morbidity 482, 488, 499–500 NAAMP protocol 283, 284, 287, 437, 438
Moriau, L. 263 Nagai, T. 373
Morin, O. J. 90, 93, 98, 204 Nagendra, H. 269
Morin, P. A. 423 Nagy, K. A. 375, 387
Morisita 330–1 Nanjappa, P. 135
Morisita, M. 311 Napolitano, G. E. 82
Morisita-Horn 333 Narendran, T. C. 27
morphology 168, 365 Narins, P. M. 287
larval 39, 44–5, 51 natterjack toads 353, 415, 415f22.1, 418, 420
tadpoles 40, 45–50 Natrix maura 512
variation 135 natural-cover surveys 250–1
see also caecilians, larval natural selection 156, 408
morphometry 306 Navas, C. A. 310, 370, 372, 373, 378, 388, 391, 400
Morrison, M. L. 25, 300 Naveh, Z. 339
mortality 215–16, 221, 409, 438, 452, 482, Nectophrynoides 14
483, 488, 489, 490–2, 497–8, Nectophrynoides occidentalis 14
499, 501, 522 Nelson, G. L. 288
caused by metals 115 nematode lungworms 500
and embryos 111 neophytes 60
high temperature 209 nested clade analysis 421, 422f22.5
study animals 169 nested designs 249, 257–8
and toe clipping 128 nests 149, 155
Morton, M. L. 145 artificial sites 149
Moseley, K. R. 220, 223 attendance 147–8, 155–6
Mossman, M. J. 282, 283, 437 construction 147, 148
Moulton, C. A. 237, 238 counts 143, 148–9, 153, 154, 157, 158
Mourão, G. 267 embryos 156
mouthparts, see oral apparatus desertion 155–6
546 | Index
nests (cont.) O’Connor, M. P. 378, 390, 400

destruction 155 Odonates 175
rainfall 151 Oedipina 9
terrestrial oviposition 144 OIE (World Animal Health
see also foam nests; salamanders, nests Organization) 487–8
nets 175–6, 467 Okada, S. 187
neuromasts 41, 42, 46–7 Oldham, R. S. 188
New Guinea 14, 15 Oligochaeta 179
New Zealand 134 Ollivier, L. M. 304, 307
Newman, R. A. 419 Olson, D. H. 55, 270, 437
newts 98, 99, 498 omnivores:
alpine 352 freshwater 77
crested 419 tadpoles 72–3
marbled 419 Onychodactylus 44
red-spotted 9, 340 oral apparatus 47, 48f3.4, 49–50, 71, 72–3
Nichols, J. D. 154, 288, 441, 453, 454, 455, 470 O’Reilly, J. C. 169
Nickerson, M. A. 381 Oreophrynella 14
Nieuwkoop, P. D. 49 organic acids 110
Niewiarowski, P. H. 204, 206, 221, 222 organic carbons 105, 113
Nishikawa, K. C. 131 dissolved 308, 309
nitrates 114 organisms 3, 17, 31, 50, 61, 75, 78, 79, 80, 81,
nitrogen 78–8, 79, 308 82, 88, 91, 96, 123, 124, 129, 179,
compounds 114 205, 207, 209, 215, 301–2, 303, 341,
cycling 80 349, 354, 355, 365, 372, 380, 387,
Noble, G. K. 363 442, 447–8, 448, 481, 486, 523
nocturnal sites 398 acquatic 71, 105–6, 110, 112, 113, 114, 191
Noel, S. 420 Orser, P. N. 129
non-calling amphibians 288, 294 Orthopterans 173
Noon, B. R. 253 Orton, F. 307
Noonan, B. P. 421 Oseen, K. L. 286, 289, 293
normoxic conditions 106 Osteopilus brunneus 13
Norris, J. D. 451 Osteopilus septentrionalis 511
North America 5, 8, 10, 13, 15, 108, 116, 144, Otis, D. L. 450, 451, 452, 455, 471, 472, 474
145, 150f9.2, 152, 153, 282, 286, Ott, J. A. 137
419, 421, 424, 482, 501 Ouellet, M. 487
northern green frogs 113, 194f11.2 Ovaska, K. 129, 133
northern leopard frogs 108 Overton, E. 372
Northern Hemisphere 25 oviparous species 6, 7, 8, 9, 10, 11, 13, 15, 143
Noss, R. F. 16, 17 explosive-breeding 150–1
Notophthalmus viridescens 9, 45, 98, 131 oviposition:
pattern mapping 135 sites 147, 148, 512
nuclear DNA 412, 413, 423 spatial distribution 152
nucleotides 413 strategies 144–5, 151, 153, 164–5ap9.1
single nucleotide polymorphisms (SNPs) 417 terrestrial 144
Nussbaum, R. A. 6 oxygen:
nutrients 16, 80, 90, 93, 95, 96, 105, 308, dissolved 105, 106–7, 107f7.1, 109, 308
309, 380 meters 107–8; water temperature 109f7.2
cycling 221, 223, 321 free (O2) 105–6
Nuttle, T. 311 ponds 107
Occidozyga 50 Pacific treefrogs 109

occupancy, see species, occupancy paedomorphism 7, 8, 9
occupany models, vital rates 457–8 PAHs (polycyclic aromatic hydrocarbons)
occurence, see species, occurrence 115–16, 309
Index | 547
Palen, W. J. 436 Peterjohn, B. G. 437, 438

Palis, J. G. 234 Peterman, W. E. 129, 188, 189, 190, 191
Pallister, J. 495 Peterson, C. G. 78, 79
Papua-Australia 4, 5 Peterson, C. R. 288, 289, 292, 378, 389, 390,
parasites 105, 267, 376, 381, 482, 487, 495, 522 401, 402
Paris, M. J. 99, 482, 489, 495, 497 Peterson, M. G. 155
Parmelee, J. R. 151 Peterson, S. 501
Parmesan, C. 388 Petranka, J. W. 143, 145, 151, 219, 250, 251,
Parris, K. M. 32, 130, 231, 334 256, 489, 514
Parry, D. 389, 392 Pfennig, D. W. 65
passive integrated transponders, see PITs Pfingsten, R. A. 43f3.2
passive traps, see traps, passive pH 105, 109, 110–12, 115, 308, 314, 511, 512,
pasture 247 516, 518
counts 257, 257f14.1, 258f14.2, 259 Phaeognathus hubrichti 4f1.1, 253
Paszkowski, C. A. 231, 288 pharmaceuticals 116
Patarnello, T. 416 Philautus 250
pathogens 91, 133, 267, 380, 424, 481, 482, Phillips, C. A. 235
483, 486, 487–8, 489, 490, Phillips, K. 15
495t26.6, 496, 497, 499, 500–1, Phillott, A. D. 32, 130
522, 523 photography 134–5, 220, 266, 339, 413
die-offs 491 ground (understory) measurements 313
host interactions 387, 403 tree canopies 311–12
human 498 photosynthesis 107, 309
sample size 491t26.4 Phrynobatrachus 47
Paton, P. W. C. 144, 145, 150, 152, 153, 154, Phyllomedusa 12
155, 157, 188, 191, 197, 199 Phyllomedusine 145
Patrick, D. A. 216, 221, 223 phylogenetic species 364, 374, 422f22.5
pattern mapping 134–5 phylogeography 26, 421, 423
Pauley, T. K. 62, 63, 237 Physalaemus 11
Pavajeau, L. 487, 521 Physalaemus lisei 179
Pawar, S. S. 269, 334 physical models 377–8, 388–9, 390–2,
PCBs (polychlorinated biphenyls) 116, 309 392f21.1, 393, 395–6, 400, 402–3
PCQ (point-center-quarter) 310–11 and EWL 399
PCR (polymerase chain reaction) 352–3, 408, spectral properties 399
409, 410–11, 412–4, 424, 493, see also agar models; frogs, physical models
495–6 physiological ecology 363–4, 365, 374, 382
Pearman, P. B. 251, 254, 270 physiology 56, 349, 352, 356, 363, 365, 371,
Pearson, P. G. 222 374, 379, 363, 365, 371, 381, 387
Pechmann, J. H. K. 24, 143, 203, 205, 207, 219, and ecology 380
220, 221, 222, 230, 232, 235, 473 and environment 379
pedomorphs 39, 42, 44–5 monitoring devices 368–9t20.1
coloration 44 phytoplankton 96
pedotypes 39, 44–5 phytotelmons 50
Peek, J. M. 312 Pialek, J. 417
Pellet, J. 24, 283, 286, 346, 350, 437, 438, 457, Pianka, E. R. 177, 326
467, 468, 469, 470, 472, 474, 477 Picco, A. M. 487, 489, 501
Pelobatidae 47 Pidancier, N. 409
Penman, T. D. 294 Pierce, B. A. 287
pens 88, 207, 208–9, 216 Pieterson, E. C. 334
Percsy, C. 521 pigment analyses 78
permits 267, 273 pilot studies 26–7
Perry, J. N. 354 Pinder, A. W. 108, 374
Peru 179–80 pine barren’s treefrogs 111
pesticides 93, 114–15, 115–16, 205, 309, 354 Pineda, E. 327
548 | Index
Piotrow, P. T. 27 temporary 11, 13, 28, 29

Pipa 10, 168 vernal 60, 61, 105, 514
pipe sampling 59f4.1 pool frogs 423
Pipidae 49 Poole, V. A. 500
diets 71, 168 pools:
tadpoles 45 breeding 145, 353
PITs (passive integrated transponders) 123, vernal 144–5
125, 130, 136, 136t8.7, 137, 138, wading 89f6.2, 93, 94, 95, 100–1
220, 372 Pope, S. E. 340, 346, 350
pitfall traps, see traps, pitfall population 288, 340f19.1, 341, 374, 408, 433
Pittman, S. E. 237 connectivity 349–51, 420
plankton 92, 93, 100 decline 380, 436, 482, 483, 488
plants 206, 309, 310, 312, 313, 512 defining 408
Plat, J. R. 431 density 61, 62, 63, 157, 203, 204–5, 207,
Platypelis 13 215f12.3, 216, 217f12.4, 218, 235,
Platz, J. E. 290 241, 247, 248, 253
Pledger, S. 451 die-offs 489, 499
Pleistocene–Holocene recolonizations 25 fluctuations 143, 167, 292
Plethodon cinereus 155, 237, 353, 420, 423 genetics 26, 352
Plethodon glutinosus 131 growth 157
Plethodon jordoni 131 isolation 91, 101, 349
Plethodon punctatus 148 landscape ecology 346
Plethodon vehiculum 129, 133 monitoring 143
Plethodonthyla 13 persistence 21, 347, 355, 356
Plethodontidae 42, 44, 106, 129, 133, 439–40, processes 339, 340
441, 472, 490 size 148–9, 154, 156, 272–3, 420–1, 440,
Pleurodema 11 441, 466f25.1, 469
Pleurodema borelli 12f1.2 structures 288, 366, 412, 413, 414, 417, 419,
plot, see area-based surveys, plot 420, 420f22.4
Plötner, J. 408 trends 157, 285, 467–8, 473
poikilotherms 108 Post, D. M. 79
point-center-quarter, see PCQ Pough, F. H. 4, 5, 16, 149
pint counts 472 Pounds, J. A. 376, 388, 401, 436
poison frogs (poison dart frogs) 178, 498 power analysis 30, 73, 239, 255
Pollock, K. H. 30, 447, 451, 455, 466, 470 Prado, C. P. A. 6
pollutants 105, 113, 114, 307, 510 Prathapan, K. D. 27
organic 115–16 precipitation 151, 152, 240, 286, 292, 293,
polychlorinated biphenyls, see PCBs 303, 514
polycyclic aromatic hydrocarbons, see PAHs precision 31, 76, 117, 204, 207, 208
polymerase chain reaction, see PCR predation 93, 94, 98, 99, 100, 366, 489
polymorphic DNA, see RAPD predator-prey relationships 321, 387
ponds 8, 9, 10, 11–12, 40, 45, 62, 64, 67, predators 40, 51, 65, 98, 105, 106, 156, 167,
91, 100, 105, 107, 155, 219, 341, 170, 199, 207, 217, 232, 234, 511
343f19.3, 347f19.4, 351–2, 435, densities 218
440, 471 temperature 109
breeding 145, 151, 152, 153, 154, 195, 198, prey 73, 75–6, 78, 80, 81, 93, 105, 207
203, 220, 232, 270, 303, 340, 408, abundance 78, 175, 176
438, 441–2, 467, 472, 515f27.1, availability 174–6, 179, 180
520f27.2 densities 218, 221
creation of 512, 513t27.2, 514 electivity index 176
communities 93, 96 invertebrates 217
drying 93, 99 sampling methods 167, 168, 169–70, 171,
natural 88, 90, 95, 96, 100 172–5, 178–9
restoration 514 Price, J. E. 207, 215, 216
Index | 549
primers 413–14, 415–16, 495 Rana cancrivora 112

Primmer, C. R. 416, 433 Rana capito 286, 288
probiotics 191 Rana cascadae 419
proteins 410, 424 Rana catesbeiana 46f3.3, 106, 132, 151, 152,
Proteus anguinus 8 154, 284, 285, 344, 498, 511, 523
protozoa 487 Rana clamitans 113, 125, 152, 197, 218, 284,
Pseudacris 11, 98, 99, 151 285f16.1, 288, 291f16.2
Pseudacris crucifer 98, 285, 293 Rana dalmatina 153, 157
Pseudacris feriarum 293 Rana latastei 418, 418f22.3
Pseudacris maculata 188 Rana lessonae 152, 423
Pseudis paradoxa 1 Rana luteiventris 419–20
Pseudacris regilla 109 Rana mucosa 134, 354, 482
Pseudobranchus 8 Rana palmipes 4f1.1
pseudoreplicates/pseudoreplication 97, 206, Rana pipiens 108, 128, 198, 208f12.2, 400
347, 435 Rana porosa 152
Puerto Rico 14, 149 Rana pretiosa 128, 132
Pyke, G. H. 137 Rana ridibunda 180
Rana sevosa 145, 153, 290, 293
Qadri, S. U. 76 Rana sphenocephala 98–9, 100, 153, 291f16.2,
quadrats, see area-based surveys, quadrats 293
questions 22, 27, 31, 204, 205, 260 Rana subaquavocalis 290
field enclosures 207 Rana sylvatica 26, 66, 108, 133, 145, 150f9.2,
construction 208–18 152, 153, 154, 155, 198, 286, 419,
habitats 205–7, 300, 523 470, 475
Rana temporaria 145, 147, 150, 153, 157, 423,
RA (Rapid biodiversity-assessment) 263–4, 424, 440
265–75, 321 Ranavirus 481, 482, 483, 486f26.1, 488, 489,
and DNA 273 490, 493, 494, 496, 497, 499,
radiation 113, 377 500, 501
ultraviolet 308, 436 random drift 424
radio-isotopes 185 random sampling 253, 434, 435, 448
radiography 496, 497 Ranidae 11, 46, 47, 49, 50, 71, 106, 144, 153,
radiotelemetry 185, 193, 199–200, 371–2, 216, 233, 250, 372, 389
379–80, 395 Ranvestel, A. W. 72, 76, 82
equipment 186–9 RAPD (polymorphic DNA) 413, 414, 416, 417
validation 197–9 Raper, S. J. 144
radiotracking 191, 194f11.2, 198, 199 Rapid biodiversity assessment, see RA
animal release 191–2 rare species, see species, rare/threatened
signal output 192–3 Rathbun, G. B. 187, 197, 199
radiotransmitters 187–8, 197, 220, 449 Raven, P. H. 15
antenna/receivers 186 Raxwothy, C. J. 269
implanted units 188, 197–9 Ray, C. 364
transmitters 186–7, 189 Ray, N. 346, 351, 352
external 187–8, 199; internal 188–9, realism 31, 87, 90, 94, 204, 207, 209,
190f11.1; subcutaneous 189; 110–14t12.1, 215, 216, 217–18
waist bands 187–8, 198, 199 Reaser, J. K. 128
rainfall 110, 193, 240, 270, 284, 286, 290, 303 recapture 125, 130, 132, 134, 135, 137, 189,
movement patterns 303 204, 209, 219, 220, 235
see also acid rain red efts 131
rainforest 198, 250, 325, 388, 401, 517 Red List of Threatened Species (IUCN) 15
Rana 46f3.3, 99 red-leg disease 482–3, 486f26.1
Rana arvalis 147, 153, 421 red-spotted newts 9, 340
Rana aurora 197, 482 redback salamanders 155, 237, 353, 420
Rana blairi 99 Reed, J. M. 353
550 | Index
reed frogs 270 Robertson, J. G. M. 132

refugia 95, 232, 237, 238 Rocha, C. F. D. 171, 250
artificial 240, 241 Rodda, G. H. 271
glacial 412 Rödel, M.-O. 143, 270, 271
Regester, K. J. 77, 154, 155 Rodgers, D. W. 76
Regosin, J. V. 205, 207, 220, 222, 223 Roelants, K. 263
Reilly, S. M. 39 Rogers, L. E. 75
relocation, repatriation, translocation, see RRT Román, A. 512, 522
Relyea, R. A. 40, 51, 94, 115, 307, 487 Rome, L. C. 108, 365, 366
removal sampling 449, 455, 472 Rosi-Marshall, E. J. 76
replicates/replication 97, 155, 260, 435, 475 Rosovsky, J. 31
reproduction 5–6, 197, 198, 371, 375, 380, 381, Rostand, J. 489
382, 387, 448, 452, 455 Rothermel, B. B. 204, 205, 206, 207, 209, 216,
endocrine system 382 221, 222, 223, 351
maturity 222 Routman, E. J. 61
temperature effects 366 Rowe, C. L. 29, 87, 88, 90, 100, 110
see also frogs, reproduction; oviparous species; Rowe, G. 412, 418, 420
oviposition; Rowley, J. J. L. 198, 367, 388, 395
salamanders, reproduction Royle, J. A. 288, 441, 448, 457, 459, 465, 470,
reptiles 168, 238, 328f18.1, 365, 389, 398, 402 472, 477
Resetarits, W. J. 237 RRT (relocation, repatriation,
respiration 308, 365, 374 translocation) 521–2
Rexstad, E. 450 Rubbo, M. J. 95
Reyer, H.-U. 219, 222 Rubio, X. 522
Rhabdias 500 Rueda-Almonacid, J. V. 270, 271
Rhacoporidae 13, 47, 50 Ruibal, R. 45, 375
Rheobatrachus 10 Ruiz-Jean, M. C. 334
Rhyacotriton 44 Ryan, T. J. 231
Rhyacotritonidae 44
Rhinoderma darwinii 12f1.2, 14 Sadinski, W. J. 111
Rhinoderma rufum 11 Saenz, D. 286, 292
Rhinophrynidae 49, 71 salamanders 3, 4, 6, 17, 42, 44, 294, 350f19.6,
rhinophrynids, tadpoles 45 424, 439, 472, 493
Rhinophrynus dorsalis 13 abundance 239, 251, 253, 449–50
Rhinotrematidae 7, 42 adults 62
Rhodes, M. 482 aquatic 8
Rhyacotritonidae 42 branding techniques 131
Ribeiroia metacecariae 495 burrows 7, 8, 9, 253
Ribeiroia ondatrae 105, 496, 522 combination aquatic/terrestrial 8–9
Riberon, A. 414 counts 465
Rice, A. N. 232 diets 168, 169
Rich, P. M. 311 egg masses 150f9.2
Richards, S. J. 25, 143, 148, 157, 436, 473, 477 eggs 7, 8, 9, 10
Richter, S. C. 145, 153 embryos 106, 111, 113
Richter-Boix, A. 340 feeding 71, 72
riparian species 80, 203 habitats 5, 112
Ritchie, M. E. 94 larvae 7, 8–9, 39, 41–2, 43, 43f3.2, 50, 62, 74
Rittenhouse, T. A. G. 133, 187, 188, 189, 191, coloration 44; size 73
193, 197, 198, 199 life styles 7
RNA 412 marking 131, 133, 250
roads: metamorphosis 39, 42, 44–5
habitats 438, 519 movements 199, 471
surveys 282, 284, 437–8 nests 148, 149, 155
Roberts, L. 264 reproduction 7–9, 151
Index | 551
sampling methods 197–8, 236, 238f13.4, satellite imagery 269, 339, 355433
239, 254, 450 Sauer, J. R. 474
seepage 9 Sayre, R. 264, 265, 266, 268, 269
skin 8, 131, 187, 189 Scallan, J. T. 438
stream 249 Scaphiopodidae 46, 47, 50
surgery 190 Scaphiopus 13, 98, 99, 237, 286, 290, 293
tails 5, 7, 131 Scaphiophus couchii 108, 152
terrestrial species 9, 169, 237, 249, 250–1, Scaphiopus holbrookii 98
439, 440, 441, 442, 449–50, Schabetsberger, R. 188, 189, 199
515, 518 Schaub, M. 255, 453, 454
toe clipping 129–30, 131, 133 Schindler, D. W. 87, 101
see also under individual species Schismaderma 49
Salamandra atra 9, 414 Schmidt, B. 272
air humidity 475, 475f25.2 Schmidt, B. R. 283, 286, 350, 439, 440, 441,
detection probabilities 474–5 449, 455, 457, 465, 466, 468,
Salamandra salamandra 9 469, 474
detection probabilities 449 Schmuck, R. 389
Salamandridae 498 Schneider, D. C. 24
Salek, L. 312 Schoener, T. W. 174
salinity 105, 112 Schotthoefer, A. M. 496
Salthe, S. N. 6 Schoutedenella xenodactyloides 353
Salmonella 498 Schwarz, C. J. 453
Salvidio, S. 268, 270, 274, 334 Schwarzkopf, L. 388, 390, 400
Sammouri, R. 41 scientific investigation 21, 24
sampling methods 26–7, 30, 31, 56–8, 205, cycle 23f2.1, 24
256, 323, 325, 374, 381–2, 388, goals 22–3
433, 456, 458, 466 applied 23; non-applied 23
amphibian larvae 55–6, 56t4.1, 60–1 objectives 23, 27–8
availability 470 effects and causes 23; relevance 27–8;
design 432, 435, 447–8, 458–9, 472, 477 time 28
food sources 73–8 Scinax boulengeri 153
larvae 252 Scolemorphidae 7
non-exposure 470–1, 472, 476 Scott, C. T. 30
plans 268–74 Scott, D. E. 94, 137, 205, 232
prey 167, 168, 169–70, 171, 172–4, 179–80 Scott, J. M. 321
sites 269, 271, 275, 283, 347, 353, 420, 436 Scott, N. J. 270, 271, 272, 274
size 433 Scott, W. A. 283
surgery 189–90, 190f11.1, 191 Seale, D. 65, 152
techniques 58, 59f4.1, 60–8, 80–1, 178–9, seasons 400, 401
229, 255, 260, 265, 267, 269, Seber, G. A. F. 253, 448, 452
270–2, 324, 327, 331, 407, 440 Seebacher, F. 388, 400
time (seasonal) 270, 324 Seigel, R. A. 511, 522
see also adaptive cluster sampling; area-based seine 61–2, 64
surveys; ARS; coverboards; selection 424
distance sampling; frogs, sampling Semlitsch, R. D. 93, 100, 101, 109, 129, 187,
methods; MCS; photography; 188, 189, 191, 193, 197, 198, 199,
RA; radiotelemetry; radiotracking; 205, 206, 207, 216, 217, 221, 222,
salamanders, sampling methods; 223, 230, 235, 340, 341, 349, 351,
traps 356, 487, 507, 510, 514, 516
Sanzone, D. M. 80 Senar, J. C. 22
Saprolegnia 131, 486, 522 Serenidae 7, 42–3
Sarasin, P. 41, 41f3.1 larvae 43, 44
Sarasin, S. 41, 41f3.1 Service, P. M. 131
Sargent, L. G. 287 Seutin, G. 273
552 | Index
Sexton, O. J. 235 Snyder, G. K. 387, 388, 399

sexual maturity, age at 215f12.3 Soberón, M. J. 325
sexual selection 408, 423 Sofianidou, T. S. 144
Seychelles 7 Sokol, O. M. 168
Seymour, C. L. 334 Solé, M. 74, 169
Seymour, R. S. 374 Solomon, C. T. 72
Shaffer, G. 419 South America 6–7, 9, 10, 11, 14, 145, 168, 421
Shaffer, H. B. 55, 62, 67, 231 South Asia 250
Shannon, C. E. 176 Southeastern Asia 7
Shannon-Weaver diversity index 176 Southerland, M. T. 207, 218, 223
Sherman, C. K. 145 Southwood, T. R. E. 300
Shields, J. A. S. 207, 215, 216 Spadefoot toads 13
Shine, D. J. 129 Couch’s 108
Shirose, L. J. 287 Sparling, D. W. 110, 111, 114, 115, 116
Shoemaker, V. H. 112, 372, 373, 387, 389, 399 spatial autocorrelation 206, 310, 347, 353, 354
Shoop, C. R. 185 spatial data 267, 339, 342, 344, 345, 346–53,
shrub (midstory) measurements 312–13 355, 356, 447
Sih, A. 219 spatial scales 24–5, 340, 471
Silva, H. R. 168 habitats 300, 341, 417
silviculture 517–18 objectives 25
similarity: pilot studies 26
indices 321, 322, 323, 329–31, 332f18.2, spatial statistics 353–4
333, 334 spatial variation 469, 470
measures of 353 spawn clumps, see egg masses
see also dissimilarity Spea 11, 13, 50, 286, 290, 293
Simon, M. P. 167 Spear, S. F. 353
Simons, T. R. 251, 439, 440, 465, 474 species 263, 328f18.1
Simpson, E. H. 177 diversity 247, 251, 264, 270, 271, 275, 321,
single nucleotide polymorphisms, see SNPs 323–5, 326–8, 328, 330, 331, 334,
Sinsch, U. 471, 472 363, 364, 440, 517, 518
Siren 8, 137, 168, 234 accumulation curves 325–7, 328f18.1,
Sjögren-Gulve, P. 340, 346 331–4
Skei, J. K. 111 distribution 3–5, 25, 363
Skelly, D. K. 58, 60, 77, 87, 90, 143, 311 indices 321–2, 322t18.1, 330, 333, 334
Skerratt, L. F. 16, 481, 496 lists 271, 274
Skidds, D. E. 157, 306 geographic distribution 299
skin 4, 41, 199, 401, 496, 498 models 25–6
attributes 364, 372, 373, 380 occupancy 282, 283, 284, 288, 295, 325, 327,
permeability 3, 5, 191, 374, 375, 376 334, 350, 441, 442–3, 448, 455–6,
resistance 378, 389, 390 456–7, 458
surgery 191, 199 occurence 41, 143, 176, 177, 198, 310, 346,
see also skin, frogs; salamanders, skin 434f23.1, 437, 441, 448, 457
SMART approach 27–8, 30 population fluctuations 25
Smith, A. M. 340 rare/threatened 26, 144, 275, 322, 324,
Smith, C. K. 250, 251, 256 326–7, 331, 441
Smith, D. G. 73 richness 324–5, 327, 330, 331, 333, 346, 438,
Smith, E. P. 324, 326 465, 517
Smith, K. F. 16 survival 363
Smith, K. G. 512 see also abundance
Smith, L. M. 64, 65 spiders 168, 179
Smith, M. A. 273, 421 Spieler, M. 188, 191, 195, 197
Smith-Vaniz, W. F. 334 spiracles 45, 46
Snodgrass, J. W. 321 Spotila, J. R. 365, 367, 376, 391
SNPs (single nucleotide polymorphisms) 417, 423 spotted frogs 419–20
Index | 553
spotted salamanders 8, 150f9.2, 197, 348f19.5, Taber, C. A. 131

348f19.5, 352, 353 tadpoles 5, 10, 13–14, 15, 16, 40, 50–1, 409,
see also Ambystoma maculatum 441, 449, 498, 500, 512
spring salamanders 66 consumption of 16
Squire, T. 419 diets 50
stable-isotope analysis 78–81 habitats 40
Stanley, T. R. 451 morphology 40
state variables 448, 453, 458 tails 45
Stefania 14 see also frogs, tadpoles
Steidl, R. J. 73 tagging/banding 131–2, 132t8.2, 133–4
Steinberg, C. E. W. 309 radiotelemetry 185
Steinman, A. D. 76, 78 see also VIE tagging
Stereochilus marginatus 44 tailed frogs 10, 106, 421, 422f22.5
Stevens, C. E. 231, 288 tails 45
Stevens, V. M. 350, 353 clipping 129
Stewart, M. M. 149, 151, 252 prehensile 7
stomach flushing 169–70, 175–6, 179 Taiwan 231
materials 170, 170f10.1, 171–2, 172f10.2, 173 Talley, B. L. 63
Storfer, A. 16, 143, 352, 482, 489, 495, 497, 501 tannins 113
Storm, R. L. 128 Tanzania 252
Storrs, S. I. 487 Taricha 498
streams 105, 107, 113, 116, 252, 436–7, 442, 519 Tattersall, G. J. 108
egg masses 145 Taub, F. B. 303
habitats 175–6, 304, 306, 518 Taylor, E. J. 300, 301
Streich, W. J. 134 Taylor, J. 131
Strumpel, A. H. P. 346 taxonomy 355
Stuart, S. N. 15, 23, 71 TDS (total dissolved solids) 112, 113–14
Stübing, D. 82 telemetry, see radiotelemetry
sub-Saharan Africa 5, 13 temperature 105, 252, 286, 293, 303,
substrates 307, 367, 377 365, 388
suction-feeding 41, 42, 71, 168 changes 365–6, 387
Sullivan, L. J. 169 and predators 109
Sun, J. W. C. 287 regulations 364
surface-dwelling species 50, 203, 272 seasonal 303, 365
surgery 189–91 water 108, 109
unit implantation 371–2 temporal scales 24–5, 26, 300, 340, 455, 471
surveys: temporal variation 465, 469, 470
testing techniques 260 termites 169
see area-based surveys; haphazard surveys; terrestrial amphibians 518
time-constrained terrestrial density 205, 215
surveys; VESs terrestrial ecology 203
survival 197, 199, 215f12.3, 219, 222, 375, terrestrial enclosures 204–5, 208f12.2,
382, 448, 449, 452 217f12.4, 218, 223
estimation 470 terrestrial habitats 55, 203, 218, 233–4, 304,
minimum number known alive 309, 351, 517
(MNKA) 221–2 terrestrial species 218, 249, 270, 376–7
Sus 511 habitat use 304, 309
Sushchik, N. N. 82 see also aquatic species; salamanders,
Swan, M. J. S. 188 terrestrial species
Sweden 147, 153 Teunis, S. F. M. 135
Sweeney, M. 135 Thailand 169
Swihart, R. K. 457 Thakur, K. A. 251, 254
Switzerland 286, 352, 437 thermal relations 365, 387, 388
Synapturanus salseri 15 thermal stress 399
554 | Index
thermometers 109, 303, 314, 366–7, 370, 371, transformation see abundance, transformation
373–4 transients 470
thermoregulation 366, 376, 377, 388 transmitters, see radiotransmitters
Thibaudeau, D. G. 39 traps 64–5, 174–5, 217, 229, 232, 235, 236,
Thomas, C. 389, 402 238, 288, 325–6, 447, 467
Thomas, C. D. 388, 403 aquatic 234
Thomas, D. L. 300, 301 active 229, 236, 237, 238, 239, 241
Thomas, E. 45 capture 235
Thomas, L. 253, 273 funnel 220, 229, 230, 230f13.1, 231, 232,
Thompson, G. G. 326 233f13.2, 234, 470–1
Thompson, S. K. 253, 432, 435 minnow 64, 234, 270, 470–1
Thoms, C. 60 passive 229–33, 233f13.3, 234, 235–6, 239
Thorpe, J. H. 73 pitfall 220, 229, 230, 231, 231f13.2, 232,
Thorson, T. B. 364, 372 233f13.3, 236, 270, 352
tiger salamanders 8, 106, 108, 198, 353 Trauth, J. B. 511
time-constrained surveys 248 Travis, J. 65, 88, 93
Timm, R. M. 263 tree (overstory) measurements 310–12
Tischendorf, L. 349 treefrogs 100–1, 237, 238, 270, 419, 437, 467
toads 5, 341, 498, 498 Cuban 511
acquatic larvae 12 detection probabilities 469, 472
burrows 376 gray 237
egg masses 151 pine barren’s 111
eggs 12, 13–14 trees 421, 514, 518
fossorial species 13–14, 237 trematodes 105, 487, 522
larvae 40 Trenham, P. C. 283, 340, 341, 351
physical models 378f20.2, 400 Triplehorn, C. A. 172
radiotransmitters 197 Triturus alpestris 352
reproduction 11–12 Triturus cristatus 135, 236, 419
sampling methods 187, 233, 435 Triturus marmoratus 419
tadpoles 13–14, 50 Triturus vulgaris 135
see also under individual species trophic niche/status:
Tobler, W. R. 353 breadth 176–7
Todd, B. D. 204, 206, 223, 231, 233, 235, 236 larval amphibians 71–2, 74, 77, 78
Todd, M. J. 284, 286, 288, 292 overlap 177
toe clipping 32, 123, 124, 125, 220, 273 prey availability 174
alphanumeric code for 126–7t8.2 relationships 167
and digit–regeneration 128, 128t8.3, 129 tadpoles 72–3
and ethics 129–30 stable-isotope data 81
see also markings; frogs, toe clipping; structure 176–7
salamanders, toe clipping tropical amphibians 145, 223, 421
Toft, C. A. 167, 168, 174, 1798, 180, 247 tropical frogs 252, 518
Toledo, R. C. 373, 375, 389 Trueb, L. 6, 40, 44, 49, 55, 73, 168, 170, 174,
Tonge, S. 522 285, 365
total dissolved solids, see TDS TSS (total suspended solids) 113
total suspended solids, see TSS Tucker, G. 31
tooth rows 47, 49, 50 tunnels 520–1
Townsend, D. S. 149, 151, 252 turbidity 114
toxicology 100 Turner, M. G. 339
toxins 188, 487, 497, 498 turtles 106
tracking 388, 395; see also radiotracking Typhlomolge 8
Tracy, C. R. 365, 366, 367, 372, 376, 377, 379, Typhlonectidae 6–7
387, 388, 389, 390, 400, 401, 402
trailing devices 133–4, 134t8.6 Uganda 250, 270
Trammell, C. 496 Ultsch, G. R. 106, 108, 112, 374
transect surveys, see area-based surveys, transect umbrella species 26
Index | 555
Underwood, A. J. 30, 435 flawed assessments 287

United Nations General Assembly 17 seasonal 286
uraeotyphlids 7 male frogs 281–2, 288
Urban, M. C. 344 Vogt, R. C. 231, 233
urbanization 420, 437, 438 Voituron, Y. 108
Urbina-Cardona, J. N. 334 Vonesh, J. R. 94, 250, 270, 272
uricotelism 373 Vos, C. C. 346
urination 374, 375 voucher specimens 273
Urodela 3, 4, 4f1.1 Vredenburg, V. T. 483
urodeles 112
egg masses 145, 147t9.2 Wahl, G. W. 148, 309
terrestrial oviposition 144 Waichman, A. V. 125, 129
US Environmental Protection Agency Wake, D. B. 15, 363, 408, 483, 487
(USEPA) 110 Wake, M. H. 6, 73
USA 8, 10, 13, 26, 93, 108, 129–30, 145, Waldron, J. L. 59f4.1, 63, 237
148–9, 154, 231, 237, 238, 282, Walker, S. F. 488
284, 287, 291f16.2, 293, 328f18.1, Walker, S. J. 282
331, 332f18.2, 350f19.6, 352, 353, Wallace, J. B. 76
419, 420, 434f23.1, 436, 438, 471, Walls, S. C. 87
490, 491t26.5, 492 Walsh, P. S. 410
Warburg, M. R. 364, 372
vagility 352 Waringer-Löschenkohl, A. 144
Vallitanios, M. 282 Warnick, R. M. 321, 330, 334
Valls, J. H. 106 washing equipment 498–9
Van Bambeke, C. 48f3.4 water 377
van Belle, G. 324, 326 anions 112
Van Buskirk, J. 437 balance 365, 373, 376
van Dam, H. 517 brackish 112
Vander Zanden, M. J. 79 cations 112
Vasconcelos, D. 198 conditions 308
vegetation 25, 60, 62, 95, 116, 145, 148, 155, conductivity 105, 112, 113, 308
187, 195, 199, 206, 217, 236, 251, depth 116
254, 259, 309–10, 313, 510, 512, exchange 366, 374–5, 391
516, 517, 518 loss 303, 401f21.7, 402
Veith, M. 123 permeability 373–4
Vences, M. 273, 408 quality 105, 516–17
Venezuela 14, 47 relations 364, 372–4, 388
Verburg, P. 79 temperature 108, 109, 286, 293
Verrell, P. 64 dissolved oxygen 109f7.2
vertebrates 3, 45, 106, 116, 167, 364, 372, 380, transfer 365
412, 414, 424, 497, 511 stress 400, 401
VESs (visual encounter surveys) 67, 158, 270–5, velocity 79, 81
288, 301, 447, 468 see also EWL
VIE tagging 65–6, 133, 138, 155 water conservation 400
see also tagging water snakes 512
Vieites, D. R. 423 watermolds 486
Vierling, K. T. 312 Watson, J. W. 132, 193, 195
Vilella, F. J. 253, 273 weather conditions 256, 324, 331, 440, 469
Villanueva-Rivera, L. J. 289, 292 and MCS 284
virus 481, 482, 494 measurement equipment 304
visual surveys, see VESs seasonal 303
Vitt, L. J. 174 Weaver, W. 176
viviparous species 6, 7, 9, 14 Weddeling, K. 231, 232, 235
vocalizations 281, 284, 285–6, 288, 290, Weeber, R. C. 282
292–3, 294 Weick, S. E. 187, 190, 197, 199, 200
556 | Index
Weir, L. A. 26, 143, 152, 282, 283, 286, 287, 437 Woolley, H. P. 131, 133
Wellborn, G. A. 306 Woosley, L. B. 154, 155
Wells, K. D. 5, 6, 144, 145, 151, 152, 153 World Animal Health Organization, see OIE
Welsh, H. H. 300, 304, 307, 439, 440 World Charter for Nature 17
Wengert, G. M. 134 worms 179
Werner, E. E. 58, 60, 61, 143, 174, 340, 436 Wright, A. A. 282
Werneria 11 Wright, A. H. 282
Western Alps 26 Wright, K. M. 482, 487, 492, 500
wetlands 55, 58, 60, 65, 88, 93, 96, 105, 110, Wu, Z. J. 169
113, 143, 144–5, 152, 153, 185, Wuycheck, J. C. 76
232, 233–4, 236, 237, 239, 341, Wygoda, M. L. 387, 389, 390, 402
433, 438, 510–15, 515f27.1, 516, Wyman, R. L. 314
517–18
Weyrauth, S. L. 514 xenobiotics 364
Whiles, M. R. 72, 79, 82, 263, 501 Xenohyla truncata 168
Whitaker, B. R. 482, 487, 492, 500 Xenopus 10, 49, 168, 421
White, G. C. 193, 453, 454, 455, 457, 476 Xenopus laevis 112, 375
Whiteman, H. H. 189, 190, 191 xerophilic frogs 373
Whitfield, S. M. 334
Whitford, W. G. 364 Yamamuro, M. 81
Whitlock, M. C. 419 Yamasaki, M. 237
Whunam, J. 31 Yamashita, K. 44
Wieczorek, A. M. 352, 353 yellow-legged frog 354
Wiens, J. A. 300, 339, 341 Yersinia 498
Wilbur, H. M. 29, 88, 90, 93, 98, 216, 222, 237 Yilmaz, Z. C. 180
Wiley, R. H. 465, 469, 474 Yip, M. J. 482
Williams, A. K. 216 Yoccoz, N. G. 30, 466, 473, 476
Williams, B. K. 95, 143, 154, 155, 269, 270, Yosemite toads 145, 419
272, 431, 432, 433, 434, 435, Young, B. 274, 275
447, 448, 450, 451, 453, 470, Young, C. A. 237
472, 477 Young, J. E. 153, 387, 389
Williams, T. 497 Yuan, J. S. 495
Wilson, E. O. 15 Yurewicz, K. L. 143, 216, 222
wind 303–4
Windmiller, B. S. 132, 154 Zahradnik, D. 312
Winkler, C. 134 Zamboni, D. S. 173
With, K. A. 348 Zamudio, K. R. 352, 353
Withers, P. C. 326, 375, 387, 399, 402 Zane, L. 416
Wolff, J. O. 21, 22, 26 Zanini, F. 346
Wood, K. V. 453 Zar, J. H. 179
wood frogs 66, 108, 111, 150f9.2, 198, 419 Zeisset, I. 423
wooden boards 237 Zimmermann, N. E. 346, 347
Woodhams, D. C. 387, 388 zooplankton 93, 96, 512
Woolbright, L.L. 254, 282 Zucchini, W. 476

Amphibians Ecology and Conservation OK

Uploaded by

Copyright:

Available Formats

Amphibians Ecology and Conservation OK

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Amphibians Ecology and Conservation OK

Uploaded by

Copyright:

Available Formats

Amphibian Ecology and Conservation

Techniques in Ecology and Conservation Series

Bird Ecology and Conservation: A Handbook of Techniques

Conservation Education and Outreach Techniques

Forest Ecology and Conservation: A Handbook of Techniques

Habitat Management for Conservation: A Handbook of Techniques

Conservation and Sustainable Use: A Handbook of Techniques

Invasive Species Management: A Handbook of Techniques

Amphibian Ecology and Conservation: A Handbook of Techniques

ISBN 978–0–19–954118–8 (Hbk.)

in these disciplines: understanding ecology leads to conservation options (see

List of contributors xxv

2 Setting objectives in ﬁeld studies 21

2.2 Steps required for a successful study 24

5 Dietary assessments of larval amphibians 71

7 Water-quality criteria for amphibians 105

Part 3 Juveniles and adults 121

8.2.1 Anurans 125

9 Egg mass and nest counts 143

10 Dietary assessments of adult amphibians 167

11 Movement patterns and radiotelemetry 185

11.7 Conclusions 199

12 Field enclosures and terrestrial cages 203

Part 4 Amphibian populations 227

13.2.3 What can passive traps tell you?

14 Area-based surveys 247

15 Rapid assessments of amphibian diversity 263

15.2.5 Data management 267

16 Auditory monitoring of anuran populations 281

17 Measuring habitat 299

17.6 What to measure 302

Part 5 Amphibian communities 319

19 Landscape ecology and GIS methods 339

19.2.4 Landscape permeability 351

Part 6 Physiological ecology and genetics 361

21 Models in ﬁeld studies of temperature and

21.2.2.3 Example of model construction and

22 Genetics in ﬁeld ecology and conservation 407

Part 7 Monitoring, status, and trends 429

24 Capture–mark–recapture, removal sampling,

25 Quantifying abundance: counts, detection

25.4.1 Estimation of abundance 471

26 Disease monitoring and biosecurity 481

27 Conservation and management 507

27.4.3 Restoring degraded lands 518

Viorel D. Popescu Department of Wildlife Ecology, University of Maine, Orono,

1.2 Amphibian species richness and distribution

1.3 Amphibian lifestyles and life history diversity

1.3.1.2 Combination aquatic and terrestrial

1.3.1.3 Terrestrial and/or fossorial

As a group salamanders exhibit diverse reproductive modes. Some salaman-

1.3.2.2 Combination of aquatic and terrestrial

In contrast, female marbled salamanders (Ambystoma opacum) lay their eggs

Crinia georgiana from Australia oviposit in small water-ﬁ lled depressions,

1.4 Amphibian declines and why they matter

Scientists have hypothesized six major threats to amphibians: habitat modi-

1.4.2 Ecosystem function

(Pough et al. 2004). If amphibians disappeared, would the world be overrun

Wake, M. H. (1993). Evolution of oviductal gestation in amphibians. Journal of

2.1 Basic concepts for a good start