A Future Beyond the Fence – an Assessment of Prey Predator Recognition in Australian Threatened Species

A Future Beyond the Fence – An Assessment of Prey Predator Recognition in Australian Threatened Species

Lisa Anna Steindler

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Biological, Earth and Environmental Sciences

Faculty of Science

February 2019

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………………......

Date……………………………………………......

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent

of the final officially approved version of my thesis. No emendation

of content has occurred and if there are any minor variations in

formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

Date ……………………………………………......

Thesis/Dissertation Sheet

Surname/Family Name : STEINDLER Given Name/s : LISA ANNA Abbreviation for degree as give in the : PhD University calendar Faculty : SCIENCE School : BIOLOGICAL, EARTH and ENVIRONMENTAL SCIENCES A FUTURE BEYOND THE FENCE – AN ASSESSMENT OF PREY Thesis Title : PREDATOR RECOGNITION IN AUSTRALIAN THREATENED SPECIES Abstract 350 words maximum: (PLEASE TYPE)

Predator-prey interactions have played a strong selective factor in the evolution of predator avoidance behaviour of prey. At an individual level predator recognition involves the understanding and inspection of a predator and discrimination between whether it poses a threat or not. Recognition is than followed by a defence response such as fight, flight or avoidance. However, in order for prey to appropriately and successfully respond and avoid predation, it is essential that prey species recognise a predatory threat in the first place. The isolation of prey species on predator free islands, geographically isolated continents (such as Australia) and predator free fenced reserves means that prey are increasingly isolated from predator driven natural selection processes. Australia has the highest contemporary mammal extinction rate in the world. Substantial evidence indicates that predation from introduced mammalian predators (feral cats and red foxes) are the main cause for these extinctions. Predator proof fences are now considered an important resource for ongoing conservation efforts, however, their long term effectiveness in terms of practical viability are still being evaluated. There are a number of theories that attempt to predict the responses of prey to their predators, particularly in environments where prey are isolated from predators and/or predators are introduced and may be considered ‘novel’ to the native fauna within that environment. In Chapter 1, I provide an overview of the current status and knowledge gaps in prey predator recognition, with particular focus on three native, threatened marsupials. Within Chapters 2, 3, 4 and 5 I carried out a number of olfactory and/or visual predator recognition experiments in order to explore: (1) predator recognition skills and abilities of threatened species within predator free fenced reserves, (2) whether seemingly ‘naïve’ prey species can develop predator recognition skills, and (3) the influence of evolutionary and ontogenetic experience on predator recognition. Results from these chapters provide support for the idea that one rule does not fit all when it comes to prey predator recognition. Rather prey predator recognition is species specific, highlighting the need for a paradigm shift if threatened species are to have a ‘future beyond the fence’.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………………………………… ……………………………………..……………… ……….……………………...…….…

Signature Witness Signature Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE Date of completion of requirements for ONLY Award:

INCLUSION OF PUBLICATIONS STATEMENT

UNSW is supportive of candidates publishing their research results during their candidature as detailed in the UNSW Thesis Examination Procedure.

Publications can be used in their thesis in lieu of a Chapter if:  The student contributed greater than 50% of the content in the publication and is the “primary author”, ie. the student was responsible primarily for the planning, execution and preparation of the work for publication  The student has approval to include the publication in their thesis in lieu of a Chapter from their supervisor and Postgraduate Coordinator.  The publication is not subject to any obligations or contractual agreements with a third party that would constrain its inclusion in the thesis

This thesis has publications (either published or submitted for publication) ☒ incorporated into it in lieu of a chapter and the details are presented below

CANDIDATE’S DECLARATION I declare that:  I have complied with the Thesis Examination Procedure  where I have used a publication in lieu of a Chapter, the listed publication(s) below meet(s) the requirements to be included in the thesis. Name Signature Date (dd/mm/yy)

Lisa Anna Steindler

Postgraduate Coordinator’s Declaration (to be filled in where publications are used in lieu of Chapters) I declare that:  the information below is accurate  where listed publication(s) have been used in lieu of Chapter(s), their use complies with the Thesis Examination Procedure  the minimum requirements for the format of the thesis have been met. PGC’s Name PGC’s Signature Date (dd/mm/yy)

For each publication incorporated into the thesis in lieu of a Chapter, provide all of the requested details and signatures required

Details of publication #1:

Full title: Discrimination of introduced predators by ontogenetically naïve prey scales with duration of shared evolutionary history

Authors: Lisa A. Steindler, Daniel T. Blumstein, Rebecca West, Katherine E. Moseby, Mike Letnic

Journal name: Animal Behaviour

Volume/page numbers: 137/133 - 139

Date accepted/ published: 6 Nov. 2018

Status Published X Accepted and In In progress press (submitted)

The Candidate’s Contribution to the Work

L.S. performed the field work, statistical analysis, experimental design and writing; D.T.B, R.B. and K.M. conceived the study and helped draft the manuscript; M.L. conceived the study, participated in the design of the study, assisted with statistical analysis and assisted writing. All authors gave final approval for publication. Location of the work in the thesis and/or how the work is incorporated in the thesis:

Chapter 2 Primary Supervisor’s Declaration

I declare that:

• the information above is accurate

• this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter

• All of the co-authors of the publication have reviewed the above information and have agreed to its veracity by signing a ‘Co-Author Authorisation’ form.

Supervisor’s name Supervisor’s signature Date (dd/mm/yy)

Mike Letnic

Preface

This thesis is comprised of six chapters: an introductory chapter providing a background and context for the research, four data chapters providing detailed original research completed as part of this doctoral thesis, and a concluding discussion chapter which summaries the research presented within the thesis, the key findings and how this research has broadened knowledge within the field of prey-predator recognition. I composed the research chapters as separate articles, which I have submitted for consideration for publication in scientific journals. Each chapter is presented with its own reference list and supplementary material (when applicable).

Research was funded by the Australian Research Council (ARC-Linkage Grant

(#LP130100173) to M.L., K.M. and D.T.B.) and ESA Holsworth Wildlife Research

Endowment (to L.S and M.L, #RG152215 and RG171896).

Work was conducted under animal ethics APEC Approval Number 1/2014

Tackling Prey Naivety in Australia’s Threatened Mammals, in accordance with South

Australian Wildlife Ethics Committee, and ACEC Approval Number 15/19A, in accordance with The Australian Code of Practice for the Care and Use of Animals for

Scientific Purposes (1997).

Acknowledgements

I would like to thank my family and friends for all the support they have provided me throughout the past five years while I have undertaken this PhD. Without the continuous support of my mum, Elisabeth Steindler, and my brother, Andrew

Steindler, I would have not been able to accomplish what I have. I would like to thank my friends (particularly Mimi Zivadinovic, Sophie Reid, Astrid Quarry, Zip Cass, Ugur

Ozer, Elizabeth Wemyss and Felicity L’Hotellier) and housemates (Tanya Harley,

Hannah Waterhouse and Stephanie Courtney Jones) for listening to my endless PhD stories and providing emotional support when I needed it the most.

Words cannot express the depth of my gratitude to my primary supervisor Mike Letnic.

Without your unbroken support throughout out the years this PhD would not have been possible. Even when I lost motivation, you believed that I could complete my PhD and for that I will always be grateful. I would also like to thank my co-supervisors

Daniel Blumstein, Katherine Moseby and Rebecca West for contributions to the ideas and funding behind the projects. Thank you to all my supervisors for all the feedback and support provided throughout my thesis and on the chapters contained within this thesis.

I would like to thank the wonderful volunteers that assisted me during my field work at

Arid Recovery, Roxby Downs – South Australia, Scotia Wildlife Sanctuary, AWC – New

South Wales, and Astrebla Downs National Park – Queensland. All volunteers brought enthusiasm and a desire to learn and work hard. Thank you for working during the hot summer days and throughout the long cold nights catching bilbies and setting up experiments.

I would also like to thank my manager, Jennifer Pierson and Pete Costell (from ACT

Parks and Conservation Service) for allowing me to take leave and study leave to complete my PhD thesis, whilst working full time. Without your support I would not have been able to complete writing my thesis.

Finally, I would like to thank the relevant land managers who allowed me to work on property. Thank you to all Arid Recovery Reserve staff for supporting my field work and allowing to me study the unique species reintroduced onto the property. A huge thank you to the staff from Scotia Wildlife Sanctuary, Australian Wildlife Conservancy for supporting my research of brush-tailed bettongs. Thank you Chris Mitchell, ranger in charge at Astrebla Downs National Park, for all your support during the planning and execution of my field trip to Astrebla, where I saw my first wild bilby!!

I would also like to acknowledge the Traditional Custodians of the land on which I worked and lived, and recognise their continuing connection to land, water and community. I pay respect to Elders past, present and emerging.

iii

Table of Contents

Preface ………………...... i

Acknowledgements ...... ii

List of Tables ...... ix

List of Figures ...... xiv

Abstract …………… ...... xix

CHAPTER 1: Introduction...... 1

1.1 An introduction to the theories on prey-predator recognition ...... 2

1.2 Prey naïveté ...... 2

1.3 Theories of predator recognition ...... 4

1.3.1 Evolutionary History – Innate recognition ...... 4

1.3.2 Ontogenetic Experience - Learned recognition hypothesis ...... 6

1.3.3 Generalisation ...... 9

1.3.3.1 The multi-predator hypothesis ...... 9

1.3.3.2 Predator archetype hypothesis ...... 12

1.3.3.3 Common constituents hypothesis ...... 14

1.3.4 Theories of predator recognition...... 16

1.4 Fencing for Conservation ...... 20

1.4.1 Financial cost of fencing for conservation ...... 21

1.4.2 Conversation value of fencing for conservation ...... 22

1.4.3 Ecological cost of fencing for conservation ...... 23

1.5 The extinction crisis in Australia ...... 28

1.6 Project aims and objectives ...... 30

1.7 References ...... 33

CHAPTER 2: Discrimination of introduced predators by ontogenetically naïve prey, scales with duration of shared evolutionary history ...... 50

2.1 ABSTRACT ...... 51

2.2 INTRODUCTION ...... 52

2.3 MATERIALS AND METHODS ...... 58

2.3.1 Study Area ...... 58

2.3.2 Study species ...... 59

2.3.3 Sources and storage of treatment odours ...... 60

2.3.4 Bilby burrow emergence behaviour ...... 60

2.3.5 Behavioural scoring ...... 63

2.3.6 Analysis of behavioural data ...... 65

2.3.7 Ethical Note ...... 66

2.4 RESULTS ...... 67

2.5 DISCUSSION ...... 72

2.6 DATA ACCESSIBILITY ...... 76

2.7 ACKNOWLEDGEMENTS ...... 76

2.8 FUNDING ...... 76

2.9 REFERENCES ...... 77

2.10 SUPPLEMENTARY MATERIAL ...... 87

CHAPTER 3: The influence of evolutionary history with predators on olfactory predator recognition in endangered marsupials ...... 88

3. 1 ABSTRACT ...... 89

3.2 INTRODUCTION ...... 91

3.3 MATERIAL AND METHODS ...... 96

3.3.1 Study species and study sites ...... 96

3.3.1.1 Burrowing bettongs (Bettongia lesueur) at Arid Recovery Reserve ...... 96 3.3.1.2 Brush-tailed bettongs (Bettongia penicillata) at Scotia Sanctuary 98 3.3.2 Experimental rationale ...... 100

3.3.3 Sources and storage of treatment odours ...... 100

3.3.4 Population level vigilance behaviour field methods ...... 101

3.3.4.1 Burrowing bettongs (Bettongia lesueur) ...... 101 3.3.4.2 Brush-tailed bettongs (Bettongia penicillata) ...... 104 3.3.5 Behavioural scoring ...... 106

3.3.6 Analysis of behavioural data ...... 109

3.4 RESULTS ...... 112

3.4.1 Visits to Station ...... 112

3.4.2 Behavioural response to odour treatment ...... 112

3.5 DISCUSSION ...... 121

3.6 ACKNOWLEDGEMENTS ...... 129

3.7 REFERENCES ...... 129

3.8 SUPPLEMENTARY MATERIAL ...... 141

CHAPTER 4: Exposure to a novel predator induces visual predator recognition by naïve prey ...... 151

4.1 ABSTRACT ...... 152

4.2 INTRODUCTION ...... 154

4.3 MATERIALS AND METHODS ...... 157

4.3.1 Study species ...... 157

4.3.2 Study site ...... 159

4.3.3 Experimental rationale ...... 161

4.3.4 Population level vigilance behaviour field methods ...... 162

4.3.5 Behavioural scoring ...... 165

4.3.6 Analysis of behavioural data ...... 167

4.3.7 Ethical Note ...... 170

4.4 RESULTS ...... 171

4.4.1 Visits to Station ...... 171

4.4.2 Behavioural Response to Model Type ...... 171

4.4.2.1 No predator treatment ...... 171

4.4.2.2 Cat enclosure treatment ...... 176

4.5 DISCUSSION ...... 181

4.6 ACKNOWLEDGEMENTS ...... 188

4.7 REFERENCES ...... 188

4.8 SUPPLEMENTARY MATERIAL ...... 199

CHAPTER 5: Not so naïve: Endangered mammal responds to olfactory cues of an introduced predator after less than 150 years of co-existence ...... 211

5. 1 ABSTRACT ...... 212

5.2 INTRODUCTION ...... 213

5.3 MATERIALS AND METHODS ...... 217

5.3.1 Study Area ...... 217

5.3.2 Sources and storage of treatment odours ...... 219

5.3.3 Bilby Behaviour ...... 220

5.3.4 Behavioural scoring ...... 223

5.3.5 Analysis of behavioural data ...... 225

5.4 RESULTS ...... 226

5.5 DISCUSSION ...... 232

5.6 ACKNOWLEDGMENTS ...... 237

vii

5.7 REFERENCES ...... 237

CHAPTER 6: Conclusion ...... 251

6.1 Introduction ...... 252

6.2 Summary of findings...... 252

6.3 Prey-predator recognition theories ...... 256

6.4 A future beyond the fence? ...... 258

6.5 A paradigm shift ...... 265

6.6 Future directions for research...... 269

6.7 Conclusions ...... 271

6.8 References ...... 273

viii

List of Tables

CHAPTER 1: Introduction ...... 1

Table 1.1 List of exclusion fences in Australia (≤10 ha), including location, area enclosed and managing agency...... 25

CHAPTER 2: Discrimination of introduced predators by ontogenetically naïve prey, scales with duration of shared evolutionary history ...... 50

Table 2.1 Ethogram of greater bilby (Macrotis lagotis) burrow emergence behaviour ...... 64

Table 2.2 Results from linear mixed-effects models to test bilby (Macrotis lagotis) burrow emergence behaviour in response to treatment (cat, dog, rabbit, and control – no odour), including order effects 70

Table S2.1 Ethogram of greater bilby (Macrotis lagotis) behaviour ...... 87

CHAPTER 3: The influence of evolutionary history with predators on olfactory predator recognition in endangered marsupials ...... 88

Table 3.1 The number of foraging stations (n) by treatment type deployed during the 2015 and 2016 sampling periods ...... 101

Table 3.2 Ethogram of burrowing bettong (Bettongia lesueur) and brush- tailed bettong (Bettongia penicillata) behaviour ...... 104

Table 3.3 The number of foraging stations (n) by treatment type deployed across night 2 and 3 ...... 108

Table 3.4 The number of foraging stations visited by bettongs by odour treatment, based upon track data. The number in parentheses indicates the total number of feeding stations, deployed with each odour treatment, from which videos were analysed for bettong behavioural responses ...... 112

Table 3.5 Results from generalized estimating equations model testing for differences between odour treatments (cat, dingo, rabbit, and control – no odour) on the mean proportion of time spent on each behaviour by burrowing bettongs (Bettongia lesueur) and brush- tailed bettongs (Bettongia penicillata) ...... 119

Table S3.1 Spatial autocorrelation of sample sites in the residuals of the fitted values for analysed behaviours for burrowing and brush-tailed bettong across the two study sites (Arid Recovery Reserve and Scotia Wildlife Sanctuary) ...... 141

Table S3.2 Brush-tailed bettongs: Results of Fisher’s Least Significant Difference (LSD) post-hoc comparisons testing for differences between model types: cat, rabbit, bucket and control, in the mean proportion of time spent on each behaviour by brush-tailed bettongs ...... 142

Table S3.3 Repeated measures results from generalized estimating equations model testing for differences between odour treatments (cat, dingo, rabbit and control – no odour) on the mean proportion of time spent on each behaviour by burrowing bettongs (Bettongia lesueur) and brush-tailed bettong (Bettongia penicillata) ...... 143

Table S3.4 List of odour treatment (cat, dingo, rabbit and control – no odour) deployed across sites (ID) per year (2015 and 2016) for burrowing bettongs (Bettongia lesueur) ...... 145

Table S3.5 List of odour treatment (cat, dingo, rabbit and control – no odour) deployed across sites (ID) per night (2 and 3) for brush-tailed bettong (Bettongia penicillata) ...... 147

Table S3.6 Ethogram of burrowing bettong (Bettongia lesueur) and brush-tailed bettong (Bettongia penicillata) behaviour ...... 150

CHAPTER 4: Exposure to a novel predator induces visual predator recognition by naïve prey ...... 151

Table 4.1 The number of foraging stations (n) by treatment type deployed during the 2015 and 2016 sampling periods ...... 163

Table 4.2 Ethogram of burrowing bettong (Bettongia lesueur) behaviour . 166

Table 4.3 The number of foraging stations visited by bettongs by model type. The number in parentheses indicates the total number of feeding stations deployed with each model type from which videos were analysed for bettong behavioural responses ...... 171

Table 4.4 Results from generalized estimating equations model testing for differences between model types (cat, rabbit, bucket and control – no visual) on the mean proportion of time spent on each behaviour by burrowing bettongs (Bettongia lesueur) in the “no predator” and “cat enclosure” ...... 179

Table S4.1 Spatial autocorrelation of sample sites in the residuals of the fitted values for analysed behaviours within the “no predator” and “cat enclosure” …...... 199

Table S4.2 “No Predator”: Results of Bonferroni pairwise comparisons testing for differences between model types: cat, rabbit, bucket and control, in the mean proportion of time spent on each behaviour by burrowing bettongs (Bettongia lesueur) ...... 200

Table S4.3 “Cat enclosure”: Results of Bonferroni pairwise comparisons testing for differences between model types: cat, rabbit, bucket and control, in the mean proportion of time spent on each behaviour by burrowing bettongs (Bettongia lesueur) ...... 201

Table S4.4 Repeated measures results from generalized estimating equations model testing for differences between model types (cat, rabbit, bucket and control – no visual) on the mean proportion of time

spent on each behaviour by burrowing bettongs (Bettongia lesueur) in the “no predator” and “cat enclosure”...... 202

Table S4.5 List of model types (bucket, cat, rabbit and control – no visual) deployed in the “no predator” enclosure across sites (ID) per year (2015 and 2016) ...... 204

Table S4.6 List of model types (bucket, cat, rabbit and control – no visual) deployed in the “Cat enclosure” across sites (ID) per year (2015 and 2016) ...... 207

Table S4.7 Ethogram of burrowing bettong (Bettongia lesueur) behaviour .. 210

CHAPTER 5: Not so naïve: Endangered mammal responds to olfactory cues of an introduced predator after less than 150 years of co-existence ...... 211

Table 5.1 Ethogram of greater bilby (Macrotis lagotis) behaviour ...... 224

Table 5.2 Results from a series of linear mixed effects models testing for differences between odour treatments (cat, dog, rabbit, and control – no odour) on the mean log proportion of time spent (log10[behaviour +1]) on each behaviour by wild greater bilbies (Macrotis lagotis)...... 230

Table S5.1 Spatial autocorrelation of sample sites in the residuals of the fitted values for analysed behaviours for greater bilbies (Macrotis lagotis) across Astrebla Downs National Park ...... 247

Table S5.2 Results of Fisher’s Least Significant Difference (LSD) post-hoc comparisons testing for differences between odour treatments: cat, dog, rabbit and control, in the mean log proportion of time spent (log10[behaviour +1]) on each behaviour by wild greater bilbies (Macrotis lagotis) ...... 248

Table S5.3 Ethogram of greater bilby (Macrotis lagotis) behaviour ...... 250

xii

CHAPTER 6: Conclusion ...... 251

Table 6.1 Summary of studies on olfactory predator recognition in Australian mammals ...... 260

Table 6.2 Summary of studies on visual predator recognition in Australian

mammals ...... 261

xiii

List of Figures

CHAPTER 1: Introduction ...... 1

Figure 1.1 Flow chart of prey-predator recognition based on evolutionary and/or ontogenetic experience...... 19

CHAPTER 2: Discrimination of introduced predators by ontogenetically naïve prey, scales with duration of shared evolutionary history ...... 50

Figure 2.1 Experimental set up for bilby predator odour discrimination study. Infrared motion sensor video camera mounted on a metal post outside the burrow entrance of a radio tracked bilby ...... 62

Figure 2.2 Principal component analysis of bilby burrow emergence behaviour (PCA) ...... 68

Figure 2.3 The mean (± 1 SEMs) proportion of time in sight (PIS) that bilbies allocated to burrow emergence behaviours (a) out of sight in burrow, (b) walk – slow locomotion, (c) run – fast locomotion, (d) partially emerged and (e) fully emerged. Similar letters (e.g. a or b) above bars identify pairwise differences that are not statistically distinguishable (P > 0.05) ...... 71

CHAPTER 3: The influence of evolutionary history with predators on olfactory predator recognition in endangered marsupials ...... 88

Figure 3.1 Map of Arid Recovery Reserve where burrowing bettong (Bettongia lesueur) odour recognition studies were conducted. Burrowing bettongs were reintroduced into the 22 km2 predator free fenced area between 1999 and 2000...... 97

Figure 3.2 Map of Scotia Wildlife Sanctuary where brush-tailed bettong (Bettongia penicillata) odour recognition studies were conducted.

xiv

Brush-tailed bettongs were reintroduced into the 40 km2 predator free fenced in 2008 ...... 99

Figure 3.3 Experimental set up for burrowing bettong (Bettongia lesueur) population level predator odour recognition experiment at Arid Recovery Reserve (a) 10 x 10 cm odour impregnated towel (cat, dingo, rabbit or control – no odour), positioned at foraging site, 5 m off the road, and (b) an infrared motion sensor video camera mounted on a metal stake 2 m from the foraging site ...... 103

Figure 3.4 Experimental set up for brush-tailed bettong (Bettongia penicillata) population level predator odour recognition experiment at Scotia Wildlife Sanctuary (a) 10 x 10 cm odour impregnated towel (cat, dingo, rabbit or control – no odour), positioned at foraging site, 5 m off the road, and (b) an infrared motion sensor video camera mounted on a metal stake 2 m from the foraging site ...... 106

Figure 3.5 Relationship between component scores derived (diamonds) from a principal component analysis of burrowing bettong behaviour (PCA) and the mean factor scores (circles ± SEM) representing the behaviours displayed by burrowing bettongs that were filmed at foraging trays for each of the odour treatments ...... 114

Figure 3.6 The mean (± SEMs) proportion of time in sight (PIS) that burrowing bettongs (Bettongia lesueur) allocated to the behaviours in response to odour treatment (cat n = 8, dingo n = 17, rabbit n = 14 and a procedural control n = 8) in the first minute of visitation to the foraging station. No letters indicate there are no significant differences ...... 115

Figure 3.7 Relationship between component scores derived (diamonds) from a principal component analysis of brush-tailed bettong behaviour (PCA) and the mean factor scores (circles ± SEM) representing the behaviours displayed by brush-tailed bettongs that were filmed at foraging trays for each of the odour treatments ...... 117

Figure 3.8 The mean (± SEMs) proportion of time in sight (PIS) that brush- tailed bettongs (Bettongia penicillata) allocated to the behaviours in response to odour treatment (cat n = 21, dingo n = 27, rabbit n = 31 and a procedural control n = 22) in the first minute of visitation to the foraging station. Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences .... 120

CHAPTER 4: Exposure to a novel predator induces visual predator recognition by naïve prey ...... 151

Figure 4.1 Burrowing bettong (Bettongia lesueur) ...... 158

Figure 4.2 Map of Arid Recovery Reserve showing areas of where the model presentation studies were conducted. The “no predator” exclosure was free of placental predators, with bettongs reintroduced between 1999 and 2000. Within the “cat enclosure” one cat of unknown sex was present in 2014, with five additional cats added between 6 and 8 months after the initial release of bettongs in 2014 ...... 160

Figure 4.3 Three dimensional models used to represent (a) cat, (b) rabbit, (c) novel object (plastic bucket), and (d) camera set up at the control, which had no physical model present. Cats are an introduced ‘novel’ predator, rabbits are a non-threatening herbivore and buckets are a non-threatening novel object. A scale is provided in the pictures, with each square measuring 100 cm2 ...... 164

Figure 4.4 Relationship between component scores derived (diamonds) from a principal component analysis of burrowing bettong behaviour (PCA) and the mean factor scores (circles ± SEM) representing the behaviours displayed by burrowing bettongs in the no predator exclosure that were filmed at foraging trays for each of the odour treatments ...... 173

xvi

Figure 4.5 The mean (± SEMs) proportion of time in sight (PIS) that burrowing bettongs (Bettongia lesueur) allocated to the behaviours in response to model types, in the “no predator” study area (bucket n = 16, cat n = 13, control n = 14 and rabbit n = 18). Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences ...... 175

Figure 4.6 Relationship between component scores derived (diamonds) from a principal component analysis of burrowing bettong behaviour (PCA) and the mean factor scores (circles ± SEM) representing the behaviours displayed by burrowing bettongs in the cat exclosure that were filmed at foraging trays for each of the odour treatments ...... 177

Figure 4.7 The mean (± SEMs) proportion of time in sight (PIS) that burrowing bettongs (Bettongia lesueur) allocated to the behaviours in response to model types, in the “cat enclosure” study area (bucket n = 11, cat n = 12, control n = 13 and rabbit n = 8). Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences ...... 180

CHAPTER 5: Not so naïve: Endangered mammal responds to olfactory cues of an introduced predator after less than 150 years of co-existence ...... 211

Figure 5.1 (a) Map of Australia showing the approximate location of Astrebla Downs National Park. (b) Map of Queensland with the exact location of Astrebla Downs National Park (1,740 km²). (c) Map of Astrebla Downs National Park. The green circles indicate the locations of the burrows where odour recognition studies on wild greater bilbies (Macrotis lagotis) were conducted ...... 218

Figure 5.2 Experimental set-up for predator odour discrimination study of wild greater bilbies at Astrebla Downs National Park. Infrared motion xvii

sensor video camera mounted on a metal post outside the burrow entrance of a wild bilby, where odour treatments (cat, dog and rabbit faeces and procedural control – no odour) were presented ...... 222

Figure 5.3 Relationship between component scores derived (diamonds) from a principal component analysis of bilby behaviour (PCA) and the mean factor scores (circles ± SEM) representing the behaviours displayed by bilbies that were filmed at burrows for each of the odour treatments ...... 227 Figure 5.4 The mean (±SEMs) proportion of time in sight (PIS) that wild greater bilbies (Macrotis lagotis) allocated to the behaviours in response to faecal odour treatments (cat, n = 44, procedural control, n = 35, dog, n = 38 and rabbit n = 36) outside bilby burrows. Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences ...... 231

CHAPTER 6: Conclusion ...... 251

Figure 6.1 Flow chart of prey-predator recognition based on evolutionary and/or ontogenetic experience ...... 268

xviii

Abstract

Australia has the highest contemporary mammal extinction rate in the world.

Substantial evidence indicates that predation from introduced mammalian predators

(feral cats and red foxes) are the main cause for these extinctions. Predator proof fences are now considered an important resource for ongoing conservation efforts, however, their long term effectiveness in terms of practical viability are still being evaluated.

There are a number of theories that attempt to predict the responses of prey to their predators, particularly in environments where prey are isolated from predators and/or predators are introduced and may be considered ‘novel’ to the native fauna within that environment.

xix

In Chapter 1, I provide an overview of the current status and knowledge gaps in prey predator recognition, with particular focus on three native, threatened marsupials. Within Chapters 2, 3, 4 and 5 I carried out a number of olfactory and/or visual predator recognition experiments in order to explore: (1) predator recognition skills and abilities of threatened species within predator free fenced reserves, (2) whether seemingly ‘naïve’ prey species can develop predator recognition skills, and (3) the influence of evolutionary and ontogenetic experience on predator recognition.

Results from these chapters provide support for the idea that one rule does not fit all when it comes to prey predator recognition. Rather prey predator recognition is species specific, highlighting the need for a paradigm shift if threatened species are to have a ‘future beyond the fence’.

CHAPTER 1:

Introduction

1.1 An introduction to the theories on prey-predator recognition

There are a number of theories that attempt to predict the responses of prey to their predators (Carthey and Blumstein, 2017), particularly in environments where prey are isolated from predators and/or predators are introduced and may be considered

‘novel’ to the native fauna within that environment. Theories on prey-predator discrimination and recognition are divided as to whether some prey species may be considered ‘naïve’, due to a lack of co-evolutionary and ontogenetic experience with predators, or whether a prey’s ability is proportionate to the duration of co-evolution

(Blumstein, 2002; Banks and Dickman, 2007). Alternatively, prey-predator recognition may be a result of ontogenetic experience (Berger, 1998) with specific predators. Or prey simply generalise their response to all predators based on shared characteristics among predators (Cox and Lima, 2006; Apfelbach et al., 2015).

1.2 Prey naïveté

Predator-prey interactions have played a strong selective factor in the evolution of predator avoidance behaviour of prey (Csányi and Dóka, 1993). At an individual level predator recognition involves the understanding and inspection of a predator and discrimination between whether it poses a threat or not. Recognition is than followed by a defence response such as fight, flight or avoidance (Csányi and Dóka, 1993).

However, in order for prey to appropriately and successfully respond and avoid predation, it is essential that a prey species recognise a predatory threat in the first place (Lima and Dill, 1990; Mitchell et al., 2015).

The prey naïveté hypothesis posits that in situations where the risk of predation is low or non-existent, that the benefits of expressing anti-predator behaviours may be outweighed by the costs of missed opportunities, such as foraging, mate choice and territorial defence (Blumstein and Daniel, 2005). Consequently, relaxed selection by predators on both ontogenetic and evolutionary time-scales may result in ‘prey naïveté’, whereby species may have diminished anti-predator behaviour and/or fail to recognize and/or mount effective responses against novel predators

(Banks, 1998; Blumstein, 2006; Blumstein, Daniel, & Springett, 2004; Goldthwaite,

Coss, & Owings, 1990).

The isolation of prey species on predator free islands, geographically isolated continents (such as Australia) and predator free fenced reserves, means that prey are isolated from predator driven natural selection processes (Jolly et al., 2018). In order for prey to recognise and respond appropriately to a potential predator, prey must perceive a cue and match it against knowledge developed through evolutionary and/or ontogenetic experience (Reed, 2004). The degree to which a perceived cue is matched to an internal template of recognition determines whether predator recognition occurs

(Blumstein and Bouskila, 1996). Due to a lack of co-evolutionary and ontogenetic experience with novel predators, or even all predators as a result of lifetime isolation, naïve species are simply unable to mount effective antipredator responses.

In environments where prey are naïve, introduced predators succeed by taking advantage of the native population that are unable to recognise and/or respond appropriately to the presence of novel predators (Ferrari et al., 2015). A number of studies have explored the concept of prey naïveté in mammals (Banks, 1998;

Blumstein et al., 2002b), fish (Ferrari et al., 2015), birds (Salo et al., 2007) and reptiles

(Gérard et al., 2016). Since the distribution of predators and prey are not constant over time nor space, as a result of range expansions, extinctions and introductions

(accidental or deliberate) (Carthey and Blumstein, 2017), we need to develop a better understanding of how long is long enough for ‘naïve’ prey to develop appropriate antipredator recognition of novel predators. We also need to develop a better understanding of the consequences of isolation from predators and whether this compounds the issue of prey naïveté.

1.3 Theories of predator recognition

1.3.1 Evolutionary History – Innate recognition

Predation is a strong selective force that can lead to behavioural modifications in prey species (McEvoy et al., 2008). Innate or ‘hard-wired’ predator recognition typically only occurs when there is an evolutionary history between predators and prey (Gall and

Mathis, 2010). If predator recognition is ‘hard-wired’, it is assumed that prey animals will be able to perform antipredator behaviour more or less correctly upon their first exposure to a predator (Blumstein, 2002). Wisenden (2003) argues that innate predator recognition should only evolve in stable habitats with constant predatory regimes and where prey species have coexisted with predators over evolutionary time

(Kats and Ferrer, 2003).

The ‘ghosts of predators past’ hypothesis posits that prey species subject to past selection for appropriate and effective antipredator behaviour will retain these behaviours if they are not too costly to do so (Peckarsky and Penton, 1988; Byers,

1997). This implies that if predator recognition is innate some species will be able to recognise and respond appropriately to the presence and threat of predation, despite

isolation from all predators within their lifetime. Innate predator recognition has been shown in mammals (Owings and Owings, 1979; Li et al., 2011), amphibians (Griffiths et al., 1998), reptiles (van Damme and Castilla, 1996) and fish (Hawkins et al., 2004; Gall and Mathis, 2010).

Despite a life time, decade, hundreds or even thousands of years of isolation from predators, some prey species have retained appropriate predator recognition abilities and responses to co-evolved, historical predators (Blumstein et al., 2008; Li et al., 2011). Coss (1999) suggested that antipredator behaviour may be retained if they are functional in other contexts. Coss (1999) found that ground squirrels isolated from predators for 70,000 – 300,000 years have retained predator recognition abilities.

Predator naïve and experienced salamanders (Eurycea nana) significantly lowered activity in response to a native predator (Epp and Gabor, 2008). Minnows (Phoxinus phoxinus) isolated from predatory pike (Esox lucius) for thousands of years are capable of performing synchronized anti-predator behaviour when first exposed to their historical predators (Pitcher et al., 1986). A study of European rabbits (Oryctolagus cuniculus) to predator odour found that there was a clear antipredator response to historical, co-evolved predators, although the rabbits were naïve to predators in their lifetime (Monclús et al., 2005). The Tasmanian swamp rat (Rattus lutreolus velutinus) responded to odours of a native, co-evolved predator (spotted-tail quoll, Dasyurus maculatus), but not introduced predators (McEvoy et al., 2008).

All of these studies suggest that there is an innate component to prey predator recognition when there has been a history of co-evolution, which is independent of experience. Despite this, it has been suggested that isolation from all predators, over a life time and/or several generations, may still lead to the loss and/or erosion of

predator recognition of historical predators for some prey species. Coss (1999) found that Artic squirrels (Spermophilus parryii), who have been isolated from predators for over 3 million years have lost their ability to recognise predatory snakes. While Jolly et al. (2018) found that northern quolls (Dasyurus hallucatus) isolated from predators for only 13 generations, showed no recognition and avoidance of predator scents.

Blumstein (2002) similarly found that although tammar wallabies (Macropus eugenii) retained visual recognition of historical predators, they lacked olfactory and acoustic predator recognition after 9,500 years of relaxed selection. It was suggested that these skills and abilities may have to be learnt through experience, rather than being innate or ‘hard-wired’. From a conservation and reintroduction perspective it is critical that we understand the time frames over which prey species can lose innate predator recognition and response abilities, which species may be affected and most importantly if they can learn to recognise historical and novel predators.

1.3.2 Ontogenetic Experience - Learned recognition hypothesis

Although innate predator recognition can be advantageous in the way naïve prey do not need to have an encounter with a predator in order to respond adaptively to potential predatory threats (Epp and Gabor, 2008), it can also be costly. Innate predator recognition implies that prey will only recognise and respond to predators they have co-evolved with, potentially limiting the extent of predators that they identify. This leaves prey vulnerable to predation by introduced, novel predators

(Wisenden, 2003). Prey species that use both innate predator recognition and experience in response to predators may be considered more efficient at predator recognition and avoidance.

The development of appropriate learnt anti-predator responses is most likely to occur if there is some genetic variation and/or behavioural plasticity within the population (Schlaepfer et al., 2002; Schlaepfer et al., 2005; Carthey and Banks, 2014).

As predator avoidance is costly in terms of reduced time and energy available for foraging, mating and territorial defence (Brown and Chivers, 2005), naïve prey are at a selective advantage if they are able to learn to assess local predation risk appropriately and respond according to the level of perceived risk (Chivers et al., 2001; Brown and

Chivers, 2005). Learning also has fitness costs associated with the development and maintenance of neural structures involved in learning and memory (Mery and Kawecki,

2003). The ability for prey to acquire recognition of novel predators can be beneficial, especially in diverse or fluctuating predatory environments (Wisenden, 2003; Epp and

Gabor, 2008; Anson and Dickman, 2013; Mitchell et al., 2015). As part of the learning experience prey must survive the initial encounter with a predator (Kelley and

Magurran, 2003).

The ‘learned recognition’ hypothesis posits that a naïve species ability to recognise and respond to introduced predators can be induced through experience

(Turner et al., 2006). Failure to recognise and appropriately respond to a predation threat increases the risk of prey capture (Chivers and Smith, 1995). As such, prey that are able to alter their behaviour patterns in accordance with learned information are expected to have a greater degree of flexibility in their response to potential predation

(Brown, 2003; Brown and Chivers, 2005).

Many species do not have innate predator recognition abilities, relying on learning and experience to properly recognise and perform antipredator behaviours

(Griffin et al., 2000). The ability to develop learnt antipredator recognition skills

towards previously evolutionary and ontogenetically unfamiliar predators has been shown in fish (Ferrari et al., 2005; Holmes and McCormick, 2010; Ferrari, 2014), birds

(Maloney and McLean, 1995) and mammals (Mineka and Cook, 1988; Griffin et al.,

2000; Webb et al., 2008). For example Martin (2014) found that naïve red swamp crayfish (Procambarus clarkia) did not recognise predators, however wild prey exposed to predators throughout their lifetime responded strongly to predator cues. Based on these results, Martin (2014) suggested that learning plays a key role in the development of anti-predator behaviour. Similarly Maloney and McLean (1995) found that after 120 years of association, that New Zealand robins (Petroica australis) learnt to recognise an introduced predator, the stoat (Mustela erminea), and that this learning enabled their survival. Similarly, it has been found that predator experienced ringtail possums (Pseudocheirus peregrinus) respond to the smells of both introduced and native predators and modify their foraging and space use patterns to avoid predation (Anson and Dickman, 2013). Experience with predators will ultimately increase the likelihood of surviving in the natural world. Only prey that recognise the risk associated with specific predators have the ability to fine-tune their behaviour to optimise survival, foraging and reproductive success, whilst under the pressure of predation (Lönnstedt et al., 2012).

If prey species are isolated from all predators, than a lack of experience may lead to a failure to recognise and respond appropriately to a novel predation threat

(Blumstein, 2002; Cox and Lima, 2006). Experience dependent behaviours may change rapidly following isolation from predators, while more hard-wired behaviours persist for many generations (Blumstein, 2002; Jolly et al., 2018). Again this has potential implications for ‘naïve’ prey species that are isolated from predators across

evolutionary and ontogenetic timescales. It is essential to have an understanding of the mechanisms leading to predator recognition in order to successfully reintroduce potentially naïve species into areas where they formerly existed.

1.3.3 Generalisation

Direct observation and/or physical encounters with a predator are extremely dangerous and represent a high risk of imminent predation threat for a prey species

(Powell and Banks, 2004). Prey may exploit cues deposited by a predator within an area, such as fur, urine or faeces, in order to avoid predator activity, with the assumption that the more concentrated a cue is, the higher the perceived predation risk (Kats and Dill, 1998; Powell and Banks, 2004; Monclús et al., 2005). It has been suggested that prey may generalise their response to predator cues, such as odour, in order to avoid a potential predator encounter. There are a number of prey-predator recognition theories that are based on the concept of a generalised response to predators, including the ‘multi-predator’ hypothesis (Blumstein et al., 2004), the

‘predator archetype’ hypothesis (Cox and Lima 2006) and ‘common constituents’ hypothesis (Kats and Dill, 1998).

1.3.3.1 The multi-predator hypothesis

The ‘multi-predator’ hypothesis suggests that prey species may retain predator recognition skills and responses towards predators they no longer coexist with if the prey species is subject to persistent pressure from other predator types (Blumstein et al., 2004; Blumstein, 2006). The ‘multi-predator’ hypothesis is restricted to prey species with multiple predators and aims to predict the condition under which

antipredator behaviour will persist following isolation from historical, co-evolved predators (Blumstein, 2006). The hypothesis proposes that antipredator behaviours are genetically linked and makes two main predictions; (1) the genes responsible for antipredator behaviour are found close together on the same chromosome(s), rather than scattered throughout the genome, and (2) the presence of any predators may be sufficient to maintain antipredator behaviour for ‘missing’ predators (Blumstein,

2006).

There are a number of studies that have tested the validity of the ‘multi- predator’ hypothesis including a study on the visual recognition of yellow-bellied marmots (Marmota flaviventris) to extant and extinct predators (Blumstein et al.,

2009). The result of the study suggested that visual predator discrimination for ontogenetically and evolutionarily novel predators may be maintained by the presence of extant predators (Blumstein et al., 2009). Similarly, tammar wallabies (Macropus eugenii) isolated from native mammalian predators for thousands of years, but exposed to avian predation, modified time allocated to vigilance and foraging in response to group size effects (Blumstein and Daniel, 2002). Group size effects are an effective antipredator behaviour against mammalian predators, with more eyes and ears increasing the chances of detecting predators and diluting the chances of being preyed upon (Blumstein et al., 2004). Group size effects should be sensitive to the overall level of predation risk, with costly antipredator behaviour selected against when there is no benefit (Blumstein et al., 2004). In this study tammar wallabies decreased the percentage of time allocated to vigilance behaviour and increased the percentage of time allocated to foraging as group size (defined by the number of conspecifics within 10m) increased (Blumstein and Daniel, 2002). These results

suggested that tammar wallabies were sensitive to the risk of predation and the persistance of some antipredator behaviour (group-size effects) may persist despite isolation from some predators. A study of European rabbits, introduced to Australia, tested for the response of rabbits to co-evolved and novel predators (cat, Felis catus,

European red fox, Vulpes vulpes, ferret, Mustela furo and spotted-tailed quoll,

Dasyurus maculatus). It was found that rabbits in Australia responded to the odour of predators they had co-evolved with in Europe (fox, cat and ferret), regardless of ontogenetic isolation from some of their historical predators (only foxes and cats were found in the study area) (Barrio et al., 2010). It was found that rabbits responded to the odour of natural co-evolved predators by decreasing their use of scented plots. In contrast, no significant avoidance of rabbits was detected when exposed to the novel allopatric predator, the spotted-tailed quoll (Barrio et al., 2010). Based on the ‘multi- predator’ hypothesis, these results suggest that the continued presence of a predator may be sufficient to retain antipredator behaviours towards a former predator.

A key prediction of the ‘multi-predator’ hypothesis is that complete isolation from all predators may lead to the rapid loss of antipredator behaviours (Blumstein et al., 2004). If antipredator behaviour is redundant as a result of isolation from all predators than antipredator traits should be rapidly lost in order to inhibit fitness costs

(Blumstein et al., 2009). This has been seen in tammar wallabies in New Zealand, who after 130 years of isolation from all predators have displayed a loss in group size effects and a rapid breakdown in visual recognition abilities (Blumstein et al., 2004).

One key, but reasonable assumption of this study was that prior to the introduction of these wallabies to New Zealand, the animals from mainland South Australia had group size effects and visual predator recognition abilities (Blumstein et al., 2004). These

assumptions were based on previous studies, which found that tammar wallabies in

New Zealand responded to the sight of both evolutionary and ontogenetically novel predators (feral cat, red fox and thylacine) (Blumstein et al., 2000), suggesting the wallabies have visual recognition abilities. From a conservation perspective, understanding the rate of decay for behavioural traits following the complete isolation from all predators is critical to ensure the successful translocation of threatened and endangered species back into areas where they formerly existed. It is also essential to understand whether the presence of a predator is sufficient for prey species to display appropriate antipredator responses to ‘novel’ predators, or whether the proper performance of antipredator behaviour may require both a heritable predisposition

(Riechert and Hedrick, 1990), as well as experience (Magurran, 1990).

1.3.3.2 Predator archetype hypothesis

The ‘predator archetype’ hypothesis argues that terrestrial species from contiguous continents should rarely be naïve towards introduced, novel predators due to their evolutionary history with functionally similar predators (Cox and Lima, 2006). A predator archetype refers to a set of predator species that use similar morphological and behavioural adaptations in obtaining prey. It may also be defined as the set of predators against which a given suite of antipredator adaptations are effective (Cox and Lima, 2006).

In terrestrial systems where continents have been joined over evolutionary time (Africa, Eurasia and the Americas), faunal exchanges have occurred (Vermeij,

1991) that have ensured the geographical distribution of predator archetypes (Cox and

Lima, 2006). For example the blue breasted quail (Coturnix chinensis); introduced to

Guam, continue to persist on the island, despite the introduction of a predatory brown tree snake (Boiga irregularis). Although the quail have no evolutionary history with the brown tree snake, they have evolved with snake archetypes (Engbring and Fritts,

1988). In a study of visual predator recognition of predator ‘naïve’ burrowing bettongs

(Bettongia lesueur), it was suggested that bettongs responded to the visual cue of a dingo (Canis familiaris), based on their evolutionary history with thylacines (Atkins et al., 2016). Burrowing bettongs used in this study were originally sourced from islands that have been isolated from all predators for 8000 years, as a result of rising sea levels

(Short et al., 1997). Prior to their isolation, burrowing bettongs would have been exposed to marsupial predators, such as thylacines (Letnic et al., 2012). According to the predator archetype hypothesis, the co-evolution of bettongs with morphologically convergent and similarly sized thylacines (Letnic et al., 2012), prior to their isolation on islands, may have been sufficient for them to develop appropriate predator recognition of dingos (Atkins et al., 2016).

The ‘predator archetype’ hypothesis posits that prey species do not need to have evolutionary and/or ontogenetic experience with a specific predator; rather experience with any predator archetype is sufficient for prey to recognise and display appropriate antipredator responses. Based on similar morphological adaptations of predators to capture prey, seemingly naïve prey may display appropriate antipredator responses (Cox and Lima, 2006).

However, in Australia there is a lack of marsupial predator archetypes, which use similar morphological and behavioural adaptations in obtaining prey, comparable to introduced eutherian predators, such as the European red fox and cat. The

Australian continent has been geographically isolated from the rest of the contiguous

continents for over 175 million years. Dingoes are thought to have been present on mainland Australia for between 3,000 – 4,000 years (Letnic et al., 2014), whilst cats have been present for over 200 years (Dickman, 2012). Although both species are members of the same order, Carnivora, these predators differentiate in their preference of food sources and hunting styles (Bradshaw, 2006). Dingos utilise a conservative feeding strategy (Corbett and Newsome, 1987), preferring large, infrequent meals, which is a reflection of the competitive feeding behaviour of their pack hunting ancestors, the wolf (Canis lupus) (Bradshaw, 2006). In contrast, cats are exclusively solitary hunters, and as such hunt prey that are typically smaller in body mass than themselves, resulting in cats having to eat several small meals per day

(Bradshaw, 2006). As such despite some morphological similarities between native marsupial and introduced eutherian predators, such as the red fox and cat, these predators may differ in some key behavioural traits, which render the defences of

Australian native prey species ineffective against introduced predators (Cox and Lima,

2006).

1.3.3.3 Common constituents hypothesis

In order to minimise and avoid the chances of being predated upon, prey species may exploit the presence of odours that indicate recent predator activity (Banks et al.,

2014). The ‘common constituents’ hypothesis proposes that predator odours share sulphurous compounds resulting from the digestion of meat. These compounds than induce a generalised recognition and response by prey, even if the predator is ‘novel’ and unfamiliar (Dickman and Doncaster, 1984; Nolte et al., 1994). This implies that prey should respond equally to odour from familiar and ‘novel’ carnivores (Banks et al.,

2014).

The detection and recognition of olfactory cues plays a key role in the evasion of predators (Banks et al., 2014). Some odours are deliberately produced to communicate with conspecifics regarding the presence, identity, reproductive and social status of the donor (Banks et al., 2014). Other odours are unavoidably produced as part of bodily functions, such as excretion and digestion (Banks et al., 2014). Odour signals are often exploited by prey species, as they provide instant information regarding predator presence and allow prey to avoid areas of high risk (McEvoy et al.,

2008; Spencer et al., 2014). Generalised predator recognition of chemical cues has been shown in reptiles (Webb et al., 2009), amphibians (Ferrari and Chivers, 2009) and fish (Ferrari et al., 2007). The ability to generalise a predatory response from known to an unknown threat is critical to prey survival in the face of novel predation threats. By generalising their response based on the recognition of common constituents in predator odour cues, prey are able to develop appropriate recognition abilities of unknown predators, prior to encountering them (Ferrari et al., 2016). If prey species are able to take information from local threats and use them to respond to novel threats, than they might be less susceptible to invasions and would protect themselves against a greater variety of predators (Ferrari et al., 2016).

In co-evolved predator-prey systems, prey use predator odours to detect risk and respond accordingly, but such recognition may be lacking where predators are alien (Banks et al., 2014). Although the ‘common constituents’ hypothesis suggests that prey will respond to the odour cues of all carnivores, regardless of their evolutionary and/or ontogenetic experience with those predators, it has been suggested that recognition may be dependent on the phylogenetic relationships of the predators, with only closely related species being recognised as threatening (Mitchell

et al., 2015). In a study of the volatile chemical profiles of urine, scats and bedding from marsupial versus placental carnivores it was found that there was little overlap between placental and marsupial carnivores across all odour types (Carthey et al.,

2017). As such the ability to detect and avoid predators may still depend on the evolutionary and/or ontogenetic history of the prey species (McEvoy et al., 2008).

1.3.4 Theories of predator recognition

Understanding of the factors that dictate prey species' abilities to recognize and respond to predators is an important theoretical issue (Cox and Lima, 2006; Ferrari et. al, 2008; Parsons et al., 2017), as well as an important applied topic. Evolutionary prey naïveté towards introduced predators has been hypothesized to be a major factor contributing to population declines of native prey species and failed attempts to reintroduce locally extinct species (Moseby et al., 2011; Salo et. al., 2007). However, our understanding of the evolutionary timeframes necessary for prey species to maintain or acquire appropriate responses to introduced predators is poorly known.

Hypotheses on the discrimination and recognition of predators by prey are divided as to whether a prey's ability is proportionate to the duration of co-evolution

(Banks and Dickman, 2007; Blumstein, 2002) or a result of ontogenetic experience

(Berger, 1998) with specific predators, whether it is a combination of co-evolution and ontogenetic experience (learning) or whether prey simply generalize their response to all predators based on characteristics shared among predators (Apfelbach et. al., 2015;

Cox & Lima, 2006). In order to determine the drivers of prey-predator recognition (i.e. evolutionary, ontogenetic experience or both), we need to understand whether these theories generate mutually exclusive predictions that would allow us to tease them

apart empirically (Table 1.1). From Table 1.1, it can be seen that based on specific evolutionary and ontogenetic experience there can be severe consequences for the survival prospectus of prey, with the potential for extinction when they lack the appropriate predator recognition and response abilities and behaviours.

Prey may exhibit innate abilities to recognize and respond to the scents and images of coevolved predators (Apfelbach et. al., 2005; Blumstein et. al, 2002c;

Monclús et. al., 2005). In contrast, prey species that have not been evolutionarily exposed to predators may learn through ontogenetic experience to recognize and respond to predators' olfactory cues (Anson and Dickman, 2013; Berger et. al., 2001) or to their visual cues (Atkins et al., 2016).

There is a theoretical overlap in predator recognition theories when prey species have both evolutionary and ontogenetic experience with predators (Table 1.1, i.e. generalisation, innate predator recognition and learned experience). Many studies that have investigated the evolved abilities of wild prey to recognize cues associated with coevolved and novel predator species did not control for variation in ontogenetic exposure to predators (Anson and Dickman, 2013; Carthey and Banks, 2012, 2016).

Thus, it remains possible that responses of prey species to predator cues reported in many studies were, to some extent, shaped by generalization (Dickman and Doncaster,

1984) or reflect a result of both an individual's lifetime experience and the history of evolutionary exposure to predators (Blumstein, 2006; Hettena et. al., 2014). In order to test the determinants of predator recognition, signals from a number of predators (e.g. cat, dingo and/or fox) and control, would need to be tested at the same time. The duration of co-evolutionary and ontogenetic experience between prey and the predators being tested would also need to vary, in order to determine the mechanisms

behind recognition (i.e. innate vs learned vs generalisation). To further determine the influence of ontogenetic and evolutionary history it is critical to know the historical and life history traits of the prey population being studied. The experiments discussed in the thesis aim to establish whether prey-predator recognition studies follow the theoretical framework outline in Table 1.1.

1. Specific evolutionary 2. Applicable 3. Applicable 4. Hypothesised prey- 5. Predicted 6. Conservation – ontogenetic experience: experience: ontogenetic predator recognition antipredator response prognosis experience evolutionary timescale timescale hypothesis

No experience Failure to Poor (with novel/recently Prey Naïveté recognise/mount (No discrimination, introduced predators, hypothesis effective response no response) = isolation from (relaxed selection) (against predators – death when historical predators) novel/historical) encounter ALL predators Evolutionary Cue discrimination – Poor – Unable to Evolutionary ‘Ghosts of recognise novel experience (prey recognise/mount experience ONLY predators past’ predators = death recognise and effective response (with historical hypothesis Good – despite respond to co-evolved (against historical predators only) (innate/hardwired) ontogenetic isolation predators) predators) respond to historical

Ontogenetic predators Ontogenetic Learned Cue discrimination – experience (period experience ONLY recognition recognise/mount of lifetime experience (period of coexistence hypothesis (acquire effective response Good with any predator, with introduced (Discrimination and including ‘novel’ recognition through (against learned predators) response) = survive introduced) experience) predators)

Generalisation: Evolutionary and Cue discrimination – Excellent - Multi-predator Ontogenetic recognise/mount (Discrimination and - Predator archetype experience (with effective response response to NOVEL historical and/or novel - Common constituents (against historical AND AND HISTORICAL predators) Learned recognition learned predators) predators) = Innate recognition survive Figure 1.1 Flow chart of prey-predator recognition based on evolutionary and/or ontogenetic experience.

1.4 Fencing for Conservation

There is a global crisis when it comes to protecting threatened and/or endangered species against threats caused by habitat loss and destruction, hunting, changes in fire regimes and the introduction of predators. Exclusion fencing, also known as conservation fencing and predator proof exclosures, have become increasingly used as a means to protect areas of high conservation value, as well as create ‘islands’ of protected habitat for native threatened and/or endangered fauna (Short and Smith,

1994; Moseby and O' Donnell, 2003; Burns et al., 2012). There are typically two types of exclosure fences; (1) expensive fences that exclude all mammalian pests and (2) cheaper, semi-permeable or ‘leaky’ fences that leak some mammalian pests. Leaky fences stop larger predators, such as cats and foxes, but allow small mammals such as mice to move through the area (Norbury et al., 2014). Exclusion fences are constructed from wire mesh and/or netting, with a buried apron running along the base of the fence on both sides, with or without an overhanging top, and sometimes carrying a sufficient electric charge to shock and repel animals attempting to climb over and across (Dickman, 2012). Typically once construction is complete; all predators are eradicated from within the fenced reserve (Dickman, 2012).

The major objective of exclusion fences is conservation. This is typically achieved through conservation of in-situ species, the reintroduction of extinct (from the area), endangered and threatened species, education/ecotourism, education/research and captive breeding (Dickman, 2012). However the value and effectiveness of conservation fences is dependent on the benefits outweighing the ecological, financial and social costs (Hayward et al., 2014). The cost effectiveness of

exclusion fences are not straightforward as it often depends on the scale, setup costs, maintenance costs and efficacy of the method (Norbury et al., 2014). It has been suggested that exclusion fencing may be the most expensive and least cost-effective conservation method (Norbury et al., 2014). Based on baseline costs and predator control efficacies, Norbury et al. (2014) found that exclusion fences in New Zealand were the cheapest and most cost-effective when the fenced area was below 1 ha.

When the area being protected was greater, a ‘leaky fence’ was most cost effective for

1 – 219 ha and trapping for areas above 219 ha (Norbury et al., 2014). It was suggested that the high capital cost needed for fenced areas greater than 1 ha was not as cost effective compared to predator trapping and ‘leaky fences’ (Norbury et al., 2014).

1.4.1 Financial cost of fencing for conservation

In between 1999 and 2009, there were 28 conservation areas in New Zealand, covering a total of 8,396 ha (Burns et al., 2012). With over 110 km of ‘predator proof’ fencing the overall capital cost was in excess of AU$22 million (Scofield et al., 2011). In South

Africa in 2012, the cost of predator proof fencing was AU$7,300 – $10,000/km, with a maintenance cost of AU$445/km per year (Lindsey et al., 2012). In Australia the cost of excluding cats and foxes is between AU$7,000 – $12,500/km (Moseby and Read, 2006;

Bode and Wintle, 2010), however this cost does not include transport of materials, clearing of lines for fence placement and labour (Pickard, 2007). It is conservatively estimated that it would cost approximately AU$1,000/km per year to maintain a predator proof fence in Australia. The high financial cost of exclusion or predator free fences means they are probably a stop gap measure until more cost-effective and long term conservation solutions are found (Burns et al., 2012).

1.4.2 Conversation value of fencing for conservation

Despite the high financial cost associated with exclusion fences they do provide a valuable tool in the conservation, protection and breeding of threatened and endangered species. Fences provide the opportunity for the eradication of introduced, threatening predators and pests, rather than having to control for fluctuations in predator and pest densities over time and through space (Burns et al., 2012). By eradicating predators within exclusion fences, prey populations are able to rapidly grow (Reardon et al., 2012). A review by Dickman (2012) found that the release of vertebrate prey species into predator free areas had a success rate of approximately

80%, compared to 0 – 60% success rate into unprotected areas. The success of releases outside of exclusion fences was dependent on the intensity of predator control, the release protocol and the identity, number, sex ratio and quality of animals released

(Dickman, 2012).

Exclusion or predator proof fences are now considered an important resource for ongoing conservation efforts, however, their long term effectiveness in terms of practical and financial viability are still being evaluated (Burns et al., 2012). The first conservation exclusion fence in Australia was established in 1975 (Dickman, 2012).

Today there are now at least 39 fenced reserves in Australia (Table 1.1) exceeding 10 ha, totalling 132,100 ha of conservation land. Exclusion or predator proof areas provide a unique opportunity to reconstruct species composition and ecological processes (Dickman, 2012) that prevailed prior to the introduction of predators, such as the European red fox and feral cat.

1.4.3 Ecological cost of fencing for conservation

Despite all the conservation benefits of exclusion fences, there are a number of known ecological costs associated with fencing. By limiting the movement of species within fenced reserves they become vulnerable to problems associated with islands and small populations, which render them more susceptible to environmental, demographic and genetic stochasticity (Hayward and Kerley, 2009; Lindsey et al., 2012). Fenced reserves prevent the immigration, migration and emigration of species, which has the potential to interrupt gene flow between populations, increasing the risk of inbreeding and enhancing the prevalence and impacts of founder effects and genetic drift (Hayward et al., 2007). Fencing also prevents the movement of species to utilise patches of primary productivity, potentially reducing the ecological capacity of the landscape (Lindsey et al., 2012). The construction of fenced reserves or exclosures fail to take into account changes in climate and vegetation structure (Reardon et al., 2012) and are unable to be moved as these changes occur.

But what about the unknown ecological cost associated with isolating prey species from all mammalian predators within exclusion fences? The isolation of prey species on predator free islands, geographically isolated continents (such as Australia) and predator free fenced reserves, means that prey are isolated from predator driven natural selection processes (Jolly et al., 2018). The possible consequences of this are the loss of appropriate antipredator behaviour and the hindrance of their development (Blumstein et al., 2006). In order to determine if and when exclusion fences should be incorporated into a conservation management plan, and which species should be held within them, we first need to weigh up the costs of restricting

species evolutionary potential in the face of change, with the benefits of conserving them in exclusion fences (Hayward and Kerley, 2009).

Table 1.1 List of exclusion fences in Australia (≥10 ha), including location, area enclosed and managing agency.

Fence name/location Fence Area (ha) Managing agency State

Mulligans Flat Woodland Sanctuary 484 Environment A.C.T A.C.T. **

Tidbinbilla Nature Reserve 102 Environment ACT A.C.T.

Australian Wildlife Walkabout Park 32 Privately owned N.S.W *

Living Desert Flora and Fauna Sanctuary 180 Broken Hill City Council N.S.W *

Scotia Wildlife Sanctuary 8,000 Australian Wildlife Conservancy N.S.W *

Australian Wildlife Conservancy and NSW Pilliga 5, 900 N.S.W. National Parks

Australian Wildlife Conservancy and NSW Mallee Cliffs 8, 000 N.S.W. National Parks

Watarrka National Park 120 Parks and Wildlife Service N.T. *

Newhaven 65, 000 Australian Wildlife Conservancy N.T.

Watarrka National Park 100 Parks and Wildlife Service N.T. **

Currawinya National Park 2,800 Queensland Parks and Wildlife Service QLD. *

Epping Forest National Park 2, 500 Queensland Parks and Wildlife Service QLD. **

Cleland Wildlife Park 35 National Parks and Wildlife S.A. *

Venus Bay Conservation Park 2,000 National Parks and Wildlife Service S.A. *

BHP, Department of Environment, Water and Arid Recovery Reserve 6,000 Natural Resources, S.A. and the University of S.A. * Adelaide

Yookamurra Sanctuary 1,100 Australian Wildlife Conservancy S.A. *

Banrock Station 1, 600 Privately owned S.A. **

Warrawong Sanctuary 34 Privately owned S.A. **

Royal Botanic Gardens Cranbourne 370 The Royal Botanic Gardens Board VIC. *

Woodlands Historic Park 300 Parks Victoria VIC. *

Hamilton Community Parklands 100 Parks Victoria VIC. *

Little Desert Nature Lodge 140 Privately owned VIC. *

Moonlit Sanctuary 10 Privately owned VIC. **

La Trobe Melbourne Wildlife Sanctuary 30 La Trobe University VIC. **

Mount Rothwell Biodiversity and 400 Privately owned VIC. ** Interpretive Centre

Useless Loop Community Biosphere Project Heirisson Prong 1,200 W.A. * Group

Department of Biodiversity, Conservation and Ellen Brook Sanctuary 30 W.A. * Attractions

Department of Biodiversity, Conservation and Twin Swamps 150 W.A. * Attractions

Department of Biodiversity, Conservation and Peron Peninsula 105,000 W.A. * Attractions

Paruna Sanctuary 2,000 Australian Wildlife Conservancy W.A. *

Karakamia Sanctuary 275 Australian Wildlife Conservancy W.A. *

Mt Gibson 7, 800 Australian Wildlife Conservancy W.A.

Department of Biodiversity, Conservation and Lorna Glen 1, 100 W.A. ** Attractions

Wadderin Sanctuary 430 Community Project Group W.A. **

Department of Biodiversity, Conservation and Woodland Reserve, Whiteman Park 200 W.A. ** Attractions

Yelverton Brook Conservation Sanctuary 40 Privately owned W.A. **

Harry Waring Marsupial Reserve 254 Privately owned W.A. **

Department of Biodiversity, Conservation and Twin Swamps Nature Reserve 150 W.A. ** Attractions

Department of Biodiversity, Conservation and Ellen Brook Nature Reserve 34 W.A. ** Attractions * (Long and Robley, 2004) ** (Dickman, 2012)

1.5 The extinction crisis in Australia

Australia has the highest contemporary mammal extinction rate in the world, with 22 species becoming extinct over the last two centuries (Johnson, 2006; Hayward et al.,

2014). Substantial evidence indicates that predation from introduced; exotic mammalian predators (feral cats and European red foxes) are the main cause for these extinctions (Short and Smith, 1994; Abbott, 2002; Salo et al., 2007; Banks et al., 2014).

Cats were first introduced to Australia in the late 18th Century and today occur in all terrestrial habitats across continental Australia, as well as over 40 offshore islands

(Dickman, 2012). Predation by feral cats is listed by the Australian government as a key threatening process under the Environment Protection and Biodiversity Conservation

(EPBC) Act 1999 (Dickman, 2012).

The European red fox was introduced to Australia on several occasions in the mid-19th Century, however it did not become established until the early 1870’s following the spread of the European rabbit. Foxes now occupy the southern three quarters of Australia and have been implicated in the decline of small to medium sized mammals (35 – 5500g) mammals, birds and reptiles (Burbidge and McKenzie, 1989;

Dickman, 2012). The red fox has also been listed as a key threatening process under the EPBC Act 1999 (Dickman, 2012). Despite over 150 years of coexistence with these introduced predators, many Australian native animals are yet to evolve appropriate risk assessment abilities and responses (Hayes et al., 2006; McEvoy et al., 2008; Mella et al., 2010).

Current methods employed to reduce the detrimental effects of feral cats and foxes include shooting, trapping and baiting. However these methods are often costly,

labour intensive and potentially effective in only small areas (Dickman, 2012). The use of exclusion fences provides an important resource for conservation, with reintroduction strategies in Australia typically focusing on re-introducing mammals into fenced reserves and/or islands from which predators have been removed and excluded. Unfortunately most translocations and reintroductions of species from captivity, islands or fenced reserves, to areas where predators are present, often fail as a result of predation (Beck et al., 1991; Miller et al., 1994; Moseby et al., 2011).

‘Beyond the fence’ releases of burrowing bettongs, greater bilbies (Macrotis lagotis) and bridled nailtail wallabies (Onychogalea fraenata) sourced from fenced reserves all failed as a result of predation (Moseby et al., 2011; Hayward et al., 2012).

Although exclusion fences provide a safe haven for ‘naïve’ native species to flourish, it may be at the detriment of the long term survival of these species, especially if they are ever to have a ‘future beyond the fence’. A lack of exposure to predators over ontogenetic and evolutionary time has the potential to lead to prey naïveté. As a result prey may not be able to display the adequate responses necessary to survive predation from certain predators (Banks, 1998; Blumstein, 2002; Russell,

2005; Cox and Lima, 2006; McEvoy et al., 2008). When prey have not co-evolved alongside predators, prey may not be adapted to the cues, appearance or behaviour of predators, resulting in an ecological mismatch that leads to a failure to recognise and/or defend against a novel threat (Carthey and Banks, 2014). In order for native, naïve species to evolve strategies to avoid predation from historical and novel, introduced predators, they may need to persist alongside low densities of introduced predators (Hayward et al., 2014; Moseby et al., 2016).

1.6 Project aims and objectives

The main objective of predator proof fences is conservation. Hayward et al. (2014) suggested that predator exclosures are a safety net that allows threatened species to persist until they have time to evolve strategies to cope with predation by introduced predators.

However, we do not know the impact and influence of experience on predator recognition. To ensure a ‘future beyond the fence’ we need to develop a greater understanding of the potential implications of living in a predator free environment, such as a fenced reserve. If a species has an evolutionary history with predators will they retain innate predator recognition skills, or will this be lost once isolated from all predators? Is it possible for ‘naïve’ prey species to develop predator recognition of introduced predators, such as the feral cat, through life time experience? These are some of the key questions that I intend to answer throughout my thesis, with the hope that one day our native mammals will have a future beyond the fence.

The species studied in this thesis have all suffered severe range contractions and population declines since European settlement, with the introduction of feral cats and European red foxes identified as the key drivers for these declines (Short and

Turner, 2000; Cardillo and Bromham, 2001; Moseby et al., 2011). The greater bilby now occupies only 20% of its former range (Southgate, 1990). The brush-tailed bettong

(Bettongia penicillata) has declined by 99% and are restricted to three small relict populations in Western Australia (Wheeler and Priddel, 2009; Marlow et al., 2015), while the burrowing bettong has been extinct on mainland Australia since the mid-20th century, with sub-populations found on islands off the Western Australia coast (Short

and Turner, 1993; Short and Turner, 1999; 2000). All three species have been successfully reintroduced to some areas and islands within their former range, where feral cats and foxes are absent, intensively controlled or eradicated (Moseby and O'

Donnell, 2003; Finlayson and Moseby, 2004; Pizzuto et al., 2007).

I designed this thesis to elucidate: (1) predator recognition skills and abilities of threatened species within predator free fenced reserves, (2) whether seemingly ‘naïve’ prey species can develop predator recognition skills, and (3) the influence of evolutionary and ontogenetic experience on predator recognition.

In order to answer these questions I carried out a number of olfactory and/or visual predator recognition experiments. I examined whether a population of ontogenetically predator naïve greater bilbies living within a large (60 km2) predator- free exclosure modified their burrow-emergence behaviour in response to olfactory stimuli from historical (dogs) and introduced predators (cats) and whether their response scaled according to the period of co-existence. I further examined the influence of evolutionary history on predator recognition, by comparatively studying the behavioural response of two species of predator naïve bettongs (burrowing bettongs and brush-tailed bettongs) living within large predator-free exclosures, to

‘whole body’ odour samples from two predators (dogs and cats). I also wished to understand whether it was possible for these species to retain predator recognition abilities, despite isolation from all terrestrial, mammalian predators over their lifetime.

In order to test whether ‘naïve’ prey had the ability to develop predator recognition of introduced predators, I assessed whether burrowing bettongs living in the presence (for 8 - 15 months) and absence of an introduced predator (feral cats),

were able to visually distinguish between a predator and harmless objects. In my final experiment I wished to understand whether a population of wild-living bilbies were naïve to the scent of an introduced predator or had developed recognition through evolutionary and ontogenetic experience.

Through this thesis I hope to answer some of these questions, which will in turn help inform conservation management decision making in Australia to ensure a ‘future beyond the fence’.

1.7 References

Abbott, I. 2002. Origin and spread of the cat, Felis catus, on mainland Australia, with a

discussion of the magnitude of its early impact on native fauna. Wildlife

Research 29(1):51-74. doi: 10.1071/WR01011

Anson, J. R., and C. R. Dickman. 2013. Behavioral responses of native prey to disparate

predators: naiveté and predator recognition. Oecologia 171(2):367-377. doi:

10.1007/s00442-012-2424-7

Apfelbach, R., C. D. Blanchard, R. J. Blanchard, R. A. Hayes, and I. S. McGregor. 2005.

The effects of predator odors in mammalian prey species: A review of field and

laboratory studies. Neuroscience & Biobehavioral Reviews 29(8):1123–1144.

doi.org/10.1016/j.neubiorev.2005.05.005.

Apfelbach, R., M. H. Parsons, H. A. Soini, and M. V. Novotny. 2015. Are single odorous

components of a predator sufficient to elicit defensive behaviors in prey

species? Frontiers in neuroscience 9doi: 10.3389/fnins.2015.00263

Atkins, R., D. T. Blumstein, K. E. Moseby, R. West, M. Hyatt, and M. Letnic. 2016. Deep

evolutionary experience explains mammalian responses to predators.

Behavioral Ecology and Sociobiology 70(10):1-9. doi: 10.1007/s00265-016-

2181-4

Banks, P. B. 1998. Responses of Australian Bush Rats, Rattus fuscipes, to the odor of

introduced Vulpes vulpes. Journal of Mammalogy 79(4):1260-1264. doi:

10.2307/1383017

Banks, P. B., Bytheway, J. P., Carthey, A. J., Hughes, N. K., & Price, C. J. (2014). Olfaction

and predator-prey interactions amongst mammals in Australia. In A. Glen, & C. R.

Dickman (Eds.), Carnivores of Australia: past, present and future (pp. 389–412).

Collingwood, Australia: CSIRO Publishing.

Banks, P. B., and C. R. Dickman. 2007. Alien predation and the effects of multiple levels

of prey naiveté. Trends in Ecology & Evolution 22(5):229-230. doi:

10.1016/j.tree.2007.02.006

Barrio, I. C., C. G. Bueno, P. B. Banks, and F. S. Tortosa. 2010. Prey naiveté in an

introduced prey species: the wild rabbit in Australia. Behavioral Ecology

21(5):986-991. doi: 10.1093/beheco/arq103

Beck, B., D. Kleiman, J. Dietz, I. Castro, C. Carvalho, A. Martins, and B. Rettberg-Beck.

1991. Losses and reproduction in reintroduced golden lion tamarins,

Leontopithecus rosalia, Dodo, Journal of Jersey Wildlife Preservation Trust

27:50-61.

Berger, J. 1998. Future prey: some consequences of the loss and restoration of large

carnivores, in: Caro, T. (Ed), Behavioral ecology and conservation biology,

Oxford University Press, New York, pp. 80-100.

Berger, J., J.E. Swenson, and I. L. Persson. 2001. Recolonizing carnivores and naïve

prey: Conservation lessons from pleistocene extinctions. Science

291(5506):1036–1039. doi.org/10.1126/science.1056466.

Blumstein, D. T. 2002. Moving to suburbia: ontogenetic and evolutionary

consequences of life on predator-free islands. Journal of Biogeography 29(5-

6):685-692. doi: 10.1046/j.1365-2699.2002.00717.x

Blumstein, D. T. 2006. The Multipredator Hypothesis and the evolutionary persistence

of antipredator behavior. Ethology 112(3):209-217. doi: 10.1111/j.1439-

0310.2006.01209.x

Blumstein, D. T., L. Barrow, and M. Luterra. 2008. Olfactory predator discrimination in

yellow-bellied marmots. Ethology 114(11):1135-1143. doi: 10.1111/j.1439-

0310.2008.01563.x

Blumstein, D. T., A. Bitton, and J. DaVeiga. 2006. How does the presence of predators

influence the persistence of antipredator behavior? Journal of Theoretical

Biology 239(4):460-468. doi: 10.1016/j.jtbi.2005.08.011

Blumstein, D. T., and A. Bouskila. 1996. Assessment and decision making in animals: a

mechanistic model underlying behavioral flexibility can prevent ambiguity.

Oikos 77(3):569-576. doi: 10.2307/3545948

Blumstein, D. T., and J. C. Daniel. 2002. Isolation from mammalian predators

differentially affects two congeners. Behavioral Ecology 13(5):657-663. doi:

10.1093/beheco/13.5.657

Blumstein, D. T., and J. C. Daniel. 2005. The loss of anti-predator behaviour following

isolation on islands. Proceedings of the Royal Society B: Biological Sciences

272(1573):1663-1668. doi: 10.1098/rspb.2005.3147

Blumstein, D. T., J. C. Daniel, A. S. Griffin, and C. S. Evans. 2000. Insular tammar

wallabies (Macropus eugenii) respond to visual but not acoustic cues from

predators. Behavioral Ecology 11(5):528-535. doi: 10.1093/beheco/11.5.528

Blumstein, D. T., J. C. Daniel, M. R. Schnell, J. G.Ardron, and C. S. Evans. 2002c.

Antipredator behaviour of red-necked pademelons: A factor contributing to

species survival? Animal Conservation 5(4):325–331.

doi.org/10.1017/S1367943002004080.

Blumstein, D. T., J. C. Daniel, and B. P. Springett. 2004. A Test of the Multi-Predator

Hypothesis: Rapid loss of antipredator behavior after 130 years of isolation.

Ethology 110(11):919-934. doi: 10.1111/j.1439-0310.2004.01033.x

Blumstein, D. T., E. Ferando, and T. Stankowich. 2009. A test of the multipredator

hypothesis: yellow-bellied marmots respond fearfully to the sight of novel and

extinct predators. Animal Behaviour 78(4):873-878. doi:

10.1016/j.anbehav.2009.07.010

Blumstein, D. T., M. Mari, J. C. Daniel, J. G. Ardron, A. S. Griffin, and C. S. Evans. 2002b.

Olfactory predator recognition: wallabies may have to learn to be wary. Animal

Conservation 5(2):87-93. doi: 10.1017/s1367943002002123

Bode, M., and B. Wintle. 2010. How to build an efficient conservation fence.

Conservation Biology 24(1):182-188. doi: 10.1111/j.1523-1739.2009.01291.x

Bradshaw, J. W. S. 2006. The Evolutionary Basis for the Feeding Behavior of Domestic

Dogs (Canis familiaris) and Cats (Felis catus). The Journal of Nutrition

136(7):1927S-1931S. doi: 10.1093/jn/136.7.1927S

Brown, G. E. 2003. Learning about danger: chemical alarm cues and local risk

assessment in prey fishes. Fish and Fisheries 4(3):227-234. doi: 10.1046/j.1467-

2979.2003.00132.x

Brown, G. E., and D. P. Chivers. 2005. Learning as an adaptive response to predation.

In: P. Barbosa and I. Castellanos, editors, Ecology of predator/prey interactions.

Oxford University Press, Oxford. p. 34-54.

Burbidge, A. A., and N. McKenzie. 1989. Patterns in the modern decline of Western

Australia's vertebrate fauna: causes and conservation implications. Biological

conservation 50(1):143-198. doi: 10.1016/0006-3207(89)90009-8

Burns, B., J. Innes, and T. Day. 2012. The use and potential of pest-proof fencing for

ecosystem restoration and fauna conservation in New Zealand, Fencing for

conservation. Springer. p. 65-90.

Byers, J. A. 1997. American pronghorn: social adaptations and the ghosts of predators

past. Chicago University Press, Chicago.

Cardillo, M., and L. Bromham. 2001. Body Size and Risk of Extinction in Australian

Mammals. Conservation Biology 15(5):1435-1440. doi: 10.2307/3061499

Carthey, A. J., and P. B. Banks. 2012. When does an alien become a native species? A

vulnerable native mammal recognizes and responds to its long-term alien

predator. PLoS One 7(2): e31804. doi.org/10.1371/journal.pone.0031804.

Carthey, A. J., and P. B. Banks. 2014. Naïveté in novel ecological interactions: lessons

from theory and experimental evidence. Biological Reviews 89(4):932-949. doi:

10.1111/brv.12087

Carthey, A. J., and P. B. Banks. 2016. Naiveté is not forever: Responses of a vulnerable

native rodent to its long term alien predators. Oikos 125(7).

doi.org/10.1111/oik.02723.

Carthey, A. J., and D. T. Blumstein. 2017. Predicting predator recognition in a changing

world. Trends in Ecology & Evolution 33(2):106-115. doi:

10.1016/j.tree.2017.10.009

Carthey, A. J. R., M. P. Bucknall, K. Wierucka, and P. B. Banks. 2017. Novel predators

emit novel cues: A mechanism for prey naivety towards alien predators.

Scientific Reports 7(1):16377 doi: 10.1038/s41598-017-16656-z

Chivers, D. P., R. S. Mirza, P. J. Bryer, and J. M. Kiesecker. 2001. Threat-sensitive

predator avoidance by slimy sculpins: understanding the importance of visual

versus chemical information. Canadian Journal of Zoology 79(5):867-873. doi:

10.1139/z01-049

Chivers, D. P., and R. J. F. Smith. 1995. Free—living fathead minnows rapidly learn to

recognize pike as predators. Journal of Fish Biology 46(6):949-954. doi:

10.1111/j.1095-8649.1995.tb01399.x

Corbett, L. K., and A. E. Newsome. 1987. The feeding ecology of the dingo. Oecologia

74(2):215-227. doi: 10.1007/bf00379362

Coss, R. G. 1999. Effects of relaxed natural selection on the evolution of behavior. In:

S.A. Foster and J.A. Endler, editors, Geographic Variation in Behavior:

Perspectives on Evolutionary Mechanisms. Oxford University Press. p. 180-208.

Cox, J. G., and S. L. Lima. 2006. Naiveté and an aquatic–terrestrial dichotomy in the

effects of introduced predators. Trends in Ecology & Evolution 21(12):674-680.

doi: 10.1016/j.tree.2006.07.011

Csányi, V., and A. Dóka. 1993. Learning interactions between prey and predator fish.

Marine Behaviour and Physiology 23(1-4):63-78. doi:

10.1080/10236249309378857

Dickman, C. R. 2012. Fences or ferals? Benefits and costs of conservation fencing in

Australia. In M. Somers and M. Hayward, editors, Fencing for Conservation.

Springer, New York. p. 43-63.

Dickman, C. R., and C. P. Doncaster. 1984. Responses of small mammals to Red fox

(Vulpes vulpes) odour. Journal of Zoology 204(4):521-531. doi: 10.1111/j.1469-

7998.1984.tb02384.x

Engbring, J., and T. H. Fritts. 1988. Demise of an insular avifauna: the brown tree snake

on Guam. Transactions of the Western Section of the Wildlife Society 24:31-37.

Epp, K. J., and C. R. Gabor. 2008. Innate and learned predator recognition mediated by

chemical signals in Eurycea nana. Ethology 114(6):607-615. doi:

10.1111/j.1439-0310.2008.01494.x

Ferrari, M. C., and D. P. Chivers. 2009. Temporal variability, threat sensitivity and

conflicting information about the nature of risk: understanding the dynamics of

tadpole antipredator behaviour. Animal Behaviour 78(1):11-16. doi:

10.1016/j.anbehav.2009.03.016

Ferrari, M. C., A. L. Crane, G. E. Brown, and D. P. Chivers. 2015. Getting ready for

invasions: can background level of risk predict the ability of naïve prey to

survive novel predators? Scientific reports 5:8309. doi: 10.1038/srep08309

Ferrari, M. C., A. L. Crane, and D. P. Chivers. 2016. Certainty and the cognitive ecology

of generalization of predator recognition. Animal Behaviour 111:207-211. doi:

10.1016/j.anbehav.2015.10.026

Ferrari, M. C., A. Gonzalo, F. Messier, and D. P. Chivers. 2007. Generalization of learned

predator recognition: an experimental test and framework for future studies.

Proceedings of the Royal Society of London B: Biological Sciences

274(1620):1853-1859. doi: 10.1098/rspb.2007.0297

Ferrari, M. C., F. Messier, and D. P. Chivers. 2008. Can prey exhibit threat-sensitive

generalization of predator recognition? Extending the predator recognition

continuum hypothesis. Proceedings of the Royal Society of London B: Biological

Sciences 275(1644):1811–1816. doi.org/10.1098/rspb.2008.0305.

Ferrari, M. C., J. J. Trowell, G. E. Brown, and D. P. Chivers. 2005. The role of learning in

the development of threat-sensitive predator avoidance by fathead minnows.

Animal Behaviour 70(4):777-784. doi: 10.1016/j.anbehav.2005.01.009

Ferrari, M. C. O. 2014. Short-term environmental variation in predation risk leads to

differential performance in predation-related cognitive function. Animal

Behaviour 95:9-14. doi: 10.1016/j.anbehav.2014.06.001

Finlayson, G. R., and K. E. Moseby. 2004. Managing confined populations: the influence

of density on the home range and habitat use of reintroduced burrowing

bettongs (Bettongia lesueur). Wildlife Research 31(4):457-463. doi:

10.1071/WR03035

Gall, B. G., and A. Mathis. 2010. Innate predator recognition and the problem of

introduced trout. Ethology 116(1):47-58. doi: 10.1111/j.1439-

0310.2009.01718.x

Gérard, A., H. Jourdan, A. Millon, and E. Vidal. 2016. Knocking on Heaven's Door: Are

novel invaders necessarily facing naïve native species on islands? PLoS ONE

11(3):e0151545. doi: 10.1371/journal.pone.0151545

Griffin, A. S., D. T. Blumstein, and C. S. Evans. 2000. Training captive‐bred or

translocated animals to avoid predators. Conservation biology 14(5):1317-

1326. doi: 10.1046/j.1523-1739.2000.99326.x

Griffiths, R. A., L. Schley, P. E. Sharp, J. L. Dennis, and A. Román. 1998. Behavioural

responses of Mallorcan midwife toad tadpoles to natural and unnatural snake

predators. Animal Behaviour 55(1):207-214. doi: 10.1006/anbe.1997.0596

Hawkins, L., A. Magurran, and J. Armstrong. 2004. Innate predator recognition in

newly-hatched Atlantic salmon. Behaviour 141(10):1249-1262. doi:

10.1163/1568539042729694

Hayes, R. A., H. F. Nahrung, and J. C. Wilson. 2006. The response of native Australian

rodents to predator odours varies seasonally: a by-product of life history

variation? Animal Behaviour 71(6):1307-1314. doi:

10.1016/j.anbehav.2005.08.017

Hayward, M. W., J. Adendorff, J. O’Brien, A. Sholto-Douglas, C. Bissett, L. C. Moolman,

P. Bean, A. Fogarty, D. Howarth, R. Slater, and G. I. H. Kerley. 2007. Practical

considerations for the reintroduction of large, terrestrial, mammalian predators

based on reintroductions to South Africa’s Eastern Cape province. The Open

Conservation Biology Journal 7:1 - 11. doi: 10.2174/1874839200701010001

Hayward, M. W., and G. I. H. Kerley. 2009. Fencing for conservation: Restriction of

evolutionary potential or a riposte to threatening processes? Biological

Conservation 142(1):1-13. doi: 10.1016/j.biocon.2008.09.022

Hayward, M. W., F. L'Hotellier, T. O'Connor, G. Ward-Fear, J. Cathcart, T. Cathcart, J.

Sephens, J. Stephens, K. Herman, and S. Legge. 2012. Reintroduction of bridled

nailtail wallabies beyond fences at Scotia Sanctuary-Phase 1. In: Proceedings of

the Linnean Society of New South Wales. p A27.

Hayward, M. W., K. Moseby, and J. L. Read. 2014. The role of predator exclosures in

the conservation of Australian fauna. In A. Glen and C. Dickman, editors,

Carnivores of Australia: past, present and future. CSIRO Publishing Sydney. p.

353-371.

Hettena, A. M., N. Munoz, and D. T. Blumstein. 2014. Prey responses to predator's

sounds: A review and empirical study. Ethology 120(5):427–452.

doi.org/10.1111/eth.12219.

Holmes, T. H., and M. I. McCormick. 2010. Smell, learn and live: the role of chemical

alarm cues in predator learning during early life history in a marine fish.

Behavioural Processes 83(3):299-305. doi: 10.1016/j.beproc.2010.01.013

Johnson, C. 2006. Australia's Mammal Extinctions: A 50,000 year history. Cambridge

University Press, Port Melbourne. p. 1-278.

Jolly, C. J., J. K. Webb, and B. L. Phillips. 2018. The perils of paradise: an endangered

species conserved on an island loses antipredator behaviours within 13

generations. Biology Letters 14(6): 20180222. doi: 10.1098/rsbl.2018.0222

Kats, L. B., and L. M. Dill. 1998. The scent of death: Chemosensory assessment of

predation risk by prey animals. Écoscience 5(3):361-394. doi:

10.1080/11956860.1998.11682468

Kats, L. B., and R. P. Ferrer. 2003. Alien predators and amphibian declines: review of

two decades of science and the transition to conservation. Diversity and

Distributions 9(2):99-110. doi: 10.1046/j.1472-4642.2003.00013.x

Kelley, J. L., and A. E. Magurran. 2003. Learned predator recognition and antipredator

responses in fishes. Fish & Fisheries 4(3):216-226. doi: 10.1046/j.1467-

2979.2003.00126.x

Letnic, M., M. Fillios, and M. S. Crowther. 2012. Could direct killing by larger dingoes

have caused the extinction of the thylacine from mainland Australia? PLoS One

7(5):e34877. doi: 10.1371/journal.pone.0034877

Li, C., X. Yang, Y. Ding, L. Zhang, H. Fang, S. Tang, and Z. Jiang. 2011. Do Père David's

Deer Lose Memories of Their Ancestral Predators? PLoS One 6(8):e23623. doi:

10.1371/journal.pone.0023623

Lima, S. L., and L. M. Dill. 1990. Behavioral decisions made under the risk of predation:

a review and prospectus. Canadian Journal of Zoology 68(4):619-640. doi:

10.1139/z90-092

Lindsey, P. A., C. L. Masterson, A. L. Beck, and S. Romañach. 2012. Ecological, social and

financial issues related to fencing as a conservation tool in Africa. In M. Somers

and M. Hayward, editors, Fencing for Conservation. Springer, New York. p. 215-

234.

Long, K., and A. Robley. 2004. Cost effective feral animal exlusion fencing for areas of

high conservation value in Australia - Part 2: Catalogue of fence designs. In: T.

D. o. t. E. a. Heritage (ed.). p C-12. Australian Government,

http://www.environment.gov.au/biodiversity/invasive-

species/publications/cost-effective-feral-animal-exclusion-fencing.

Lönnstedt, O. M., M. I. McCormick, M. G. Meekan, M. C. Ferrari, and D. P. Chivers.

2012. Learn and live: predator experience and feeding history determines prey

behaviour and survival. Proceedings of the Royal Society of London B: Biological

Sciences 279(1736):2091-2098. doi: 10.1098/rspb.2011.2516.

Magurran, A. E. 1990. The inheritance and development of minnow anti-predator

behaviour. Animal Behaviour 39(5):834-842. doi: 10.1016/S0003-

3472(05)80947-9

Maloney, R. F., and I. G. McLean. 1995. Historical and experimental learned predator

recognition in free-living New-Zealand robins. Animal Behaviour 50(5):1193-

1201. doi: 10.1016/0003-3472(95)80036-0

Marlow, N. J., N. D. Thomas, A. A. Williams, B. Macmahon, J. Lawson, Y. Hitchen, J.

Angus, and O. Berry. 2015. Cats (Felis catus) are more abundant and are the

dominant predator of woylies (Bettongia penicillata) after sustained fox (Vulpes

vulpes) control. Australian Journal of Zoology 63(1):18-27. doi:

10.1071/ZO14024

Martin, C. W. 2014. Naive prey exhibit reduced antipredator behavior and survivorship.

PeerJ 2:e665. doi: 10.7717/peerj.665

McEvoy, J., D. L. Sinn, and E. Wapstra. 2008. Know thy enemy: Behavioural response of

a native mammal (Rattus lutreolus velutinus) to predators of different

coexistence histories. Austral Ecology 33(7):922-931. doi: 10.1111/j.1442-

9993.2008.01863.x

Mella, V. S. A., C. E. Cooper, and S. J. J. F. Davies. 2010. Predator odour does not

influence trappability of southern brown bandicoots (Isoodon obesulus) and

common brushtail possums (Trichosurus vulpecula). Australian Journal of

Zoology 58(5):267-272. doi: 10.1071/ZO10049

Mery, F., and T. J. Kawecki. 2003. A fitness cost of learning ability in Drosophila

melanogaster. Proceedings: Biological Sciences 270(1532):2465-2469. doi:

10.1098/rspb.2003.2548

Miller, B., D. Biggins, L. Hanebury, and A. Vargas. 1994. Reintroduction of the black-

footed ferret (Mustela nigripes). In: P. J. S. Olney, G. M. Mace and A. T. C.

Feistner, editors, Creative Conservation: Interactive management of wild and

captive animals. Springer Netherlands, Dordrecht. p. 455-464.

Mineka, S., and M. Cook. 1988. Social learning and the acquisition of snake fear in

monkeys. In: T. R. Zentall and B. G. Galef, Jr., editors, Social learning:

Psychological and biological perspectives. Erlbaum, New Jersey. p. 51-73.

Mitchell, M. D., D. P. Chivers, M. I. McCormick, and M. C. Ferrari. 2015. Learning to

distinguish between predators and non-predators: understanding the critical

role of diet cues and predator odours in generalisation. Scientific reports 5:

13918. doi: 10.1038/srep13918

Monclús, R., H. G. Rödel, D. Von Holst, and J. De Miguel. 2005. Behavioural and

physiological responses of naïve European rabbits to predator odour. Animal

Behaviour 70(4):753-761. doi: 10.1016/j.anbehav.2004.12.019

Moseby, K., and J. Read. 2006. The efficacy of feral cat, fox and rabbit exclusion fence

designs for threatened species protection. Biological Conservation 127(4):429-

437. doi: 10.1016/j.biocon.2005.09.002

Moseby, K. E., D. T. Blumstein, and M. Letnic. 2016. Harnessing natural selection to

tackle the problem of prey naïveté. Evolutionary applications 9(2):334-343. doi:

10.1111/eva.12332

Moseby, K. E., and E. O' Donnell. 2003. Reintroduction of the greater bilby, Macrotis

lagotis (Reid) (Marsupialia : Thylacomyidae), to northern South Australia:

survival, ecology and notes on reintroduction protocols. Wildlife Research

30(1):15-27. doi: 10.1071/WR02012

Moseby, K. E., J. L. Read, D. C. Paton, P. Copley, B. M. Hill, and H. A. Crisp. 2011.

Predation determines the outcome of 10 reintroduction attempts in arid South

Australia. Biological Conservation 144(12):2863-2872. doi:

10.1016/j.biocon.2011.08.003

Nolte, D. L., J. R. Mason, G. Epple, E. Aronov, and D. L. Campbell. 1994. Why are

predator urines aversive to prey? Journal of Chemical Ecology 20(7):1505-1516.

doi: 10.1007/bf02059876

Norbury, G., A. Hutcheon, J. Reardon, and A. Daigneault. 2014. Pest fencing or pest

trapping: a bio‐economic analysis of cost‐effectiveness. Austral Ecology

39(7):795-807. doi: 10.1111/aec.12147

Owings, D. H., and S. C. Owings. 1979. Snake‐directed behavior by black‐tailed prairie

dogs (Cynomys ludovicianus). Zeitschrift für Tierpsychologie 49(1):35-54. doi:

10.1111/j.1439-0310.1979.tb00272.x

Parsons, M. H., R. Apfelbach, P. B. Banks, E. Z. Cameron, C. R. Dickman, A. S. Frank et

al. 2018. Biologically meaningful scents: A framework for understanding

predator–prey research across disciplines. Biological Reviews 93(1):98–114.

doi.org/10.1111/brv.12334.

Peckarsky, B. L., and M. A. Penton. 1988. Why do Ephemerella nymphs scorpion

posture: a "ghost of predation past"? Oikos 53(2):185-193. doi:

10.2307/3566061

Pickard, J. 2007. Predator-proof fences for biodiversity conservation: some lessons

from dingo fences. . In: C. Dickman, D. Lunney and S. Burgin, editors, Animals of

Arid Australia: out on their Own? Royal Zoological Society of New South Wales,

Sydney, Australia. p. 197–207.

Pitcher, T. J., D. A. Green, and A. E. Magurran. 1986. Dicing with death: predator

inspection behaviour in minnow shoals. Journal of Fish Biology 28(4):439-448.

doi: 10.1111/j.1095-8649.1986.tb05181.x

Pizzuto, T. A., G. R. Finlayson, M. S. Crowther, and C. R. Dickman. 2007. Microhabitat

use by the brush-tailed bettong (Bettongia penicillata) and burrowing bettong

(B. lesueur) in semiarid New South Wales: implications for reintroduction

programs. Wildlife Research 34(4):271-279. doi: 10.1071/WR06127

Powell, F., and P. B. Banks. 2004. Do house mice modify their foraging behaviour in

response to predator odours and habitat? Animal Behaviour 67(4):753-759.

doi: 10.1016/j.anbehav.2003.08.016

Reardon, J. T., N. Whitmore, K. M. Holmes, L. M. Judd, A. D. Hutcheon, G. Norbury, and

D. I. Mackenzie. 2012. Predator control allows critically endangered lizards to

recover on mainland New Zealand. New Zealand Journal of Ecology 36(2):141-

150.

Reed, J. M. 2004. Recognition behavior based problems in species conservation.

Annales Zoologici Fennici 41(6): 859-877.

Riechert, S. E., and A. V. Hedrick. 1990. Levels of predation and genetically based anti-

predator behaviour in the spider, Agelenopsis aperta. Animal Behaviour

40(4):679-687. doi: 10.1016/S0003-3472(05)80697-9

Russell, B. G. 2005. The role of odour in Australia mammalian predator/prey

interactions. Ph.D. Dissertation, University of New South Wales, Australia.

Salo, P., E. Korpimäki, P. B. Banks, M. Nordström, and C. R. Dickman. 2007. Alien

predators are more dangerous than native predators to prey populations.

Proceedings of the Royal Society B: Biological Sciences 274(1615):1237-1243.

doi: 10.1098/rspb.2006.0444

Schlaepfer, M. A., M. C. Runge, and P. W. Sherman. 2002. Ecological and evolutionary

traps. Trends in Ecology & Evolution 17(10):474-480. doi: 10.1016/S0169-

5347(02)02580-6

Schlaepfer, M. A., P. W. Sherman, B. Blossey, and M. C. Runge. 2005. Introduced

species as evolutionary traps. Ecology Letters 8(3):241-246. doi:

10.1111/j.1461-0248.2005.00730.x

Scofield, R. P., R. Cullen, and M. Wang. 2011. Are predator-proof fences the answer to

New Zealand's terrestrial faunal biodiversity crisis? New Zealand Journal of

Ecology 35(3):312-317.

Short, J., and A. Smith. 1994. Mammal Decline and Recovery in Australia. Journal of

Mammalogy 75(2):288-297. doi: 10.2307/1382547

Short, J., and B. Turner. 1993. The distribution and abundance of the burrowing

bettong (Marsupialia : Macropoidea). Wildlife Research 20(4):525-533. doi:

10.1071/WR9930525

Short, J., and B. Turner. 1999. Ecology of burrowing bettongs, Bettongia lesueur

(Marsupialia: Potoroidae), on Dorre and Bernier Islands, Western Australia.

Wildlife Research 26(5):651-669. doi: 10.1071/WR98039

Short, J., and B. Turner. 2000. Reintroduction of the burrowing bettong, Bettongia

lesueur (Marsupialia: Potoroidae) to mainland Australia. Biological

Conservation 96(2):185-196. doi: 10.1016/S0006-3207(00)00067-7

Short J., B. Turner, C. Majors and J. Leone. 1997. The fluctuating abundance of

endangered mammals on Bernier and Dorre Islands, Western Australia -

conservation implications. Australian Mammalogy. 20(1):53-61

Southgate, R. (1990). Distribution and abundance of the greater bilby Macrotis lagotis

Reid (Marsupialia: Peramelidae). In J. H. Seeback, P. R. Brown, R. I. Wallis, & C. M.

Kemper (Eds.), Bandicoots and bilbies (pp. 293–302). Sydney, Australia: Surrey

Beatty.

Spencer, E. E., M. S. Crowther, and C. R. Dickman. 2014. Risky business: do native

rodents use habitat and odor cues to manage predation risk in Australian

deserts? PLoS One 9(2):e90566. doi: 10.1371/journal.pone.0090566

Turner, A. M., S. E. Turner, and H. M. Lappi. 2006. Learning, memory and predator

avoidance by freshwater snails: effects of experience on predator recognition

and defensive strategy. Animal Behaviour 72(6):1443-1450. doi:

10.1016/j.anbehav.2006.05.010 van Damme, R., and A. M. Castilla. 1996. Chemosensory predator recognition in the

lizard Podarcis hispanica: Effects of predation pressure relaxation. Journal of

Chemical Ecology 22(1):13-22. doi: 10.1007/BF02040196

Vermeij, G. 1991. When biotas meet: understanding biotic interchange. Science

253(5024):1099-1104. doi: 10.1126/science.253.5024.1099

Webb, J. K., G. P. Brown, T. Child, M. J. Greenlees, B. L. Phillips, and R. Shine. 2008. A

native dasyurid predator (common planigale, Planigale maculata) rapidly learns

to avoid a toxic invader. Austral Ecology 33(7):821-829. doi: 10.1111/j.1442-

9993.2008.01847.x

Webb, J. K., W. G. Du, D. A. Pike, and R. Shine. 2009. Chemical cues from both

dangerous and nondangerous snakes elicit antipredator behaviours from a

nocturnal lizard. Animal Behaviour 77(6):1471-1478. doi:

10.1016/j.anbehav.2009.02.020

Wheeler, R., and D. Priddel. 2009. The impact of introduced predators on two

threatened prey species: A case study from western New South Wales.

Ecological Management & Restoration 10:S117-S123. doi: 10.1111/j.1442-

8903.2009.00457.x

Wisenden, B. D. 2003. Chemically mediated strategies to counter predation. In S. P.

Collins and N. J. Marshall, editors. Sensory processing in aquatic environments.

Springer, New York. p. 236-251.

CHAPTER 2:

Discrimination of introduced predators by ontogenetically

naïve prey, scales with duration of shared evolutionary history

Lisa Steindler1, Daniel T. Blumstein2, Rebecca West1, 3, Katherine E. Moseby1, 3, Mike

Letnic1*

1Centre for Ecosystem Science, School of Biological, Earth and Environmental Sciences,

University of New South Wales, 2052, Sydney, Australia

2Department of Ecology and Evolutionary Biology, University of California, 621 Young

Drive South, Los Angeles, CA 90095-1606, USA

3Arid Recovery Ltd., P.O. Box 147, Roxby Downs 5725, Australia

*Corresponding Author: Mike Letnic

Published in: Animal Behaviour, 137(133 – 139)

Date accepted/ published: 6 Nov. 2018

2.1 ABSTRACT

Hypotheses on the discrimination and recognition of predators by prey are divided as to whether the prey species' ability to recognize and avoid predators is proportionate to the duration of evolutionary exposure to specific predators or is a result of more generalized discrimination processes. Moreover, understanding of the timeframes necessary for prey species to maintain or acquire appropriate responses to introduced predators is poorly understood. We studied a population of wild, ontogenetically predator naïve greater bilbies, Macrotis lagotis, living within a large (60 km2) predator- free exclosure, to determine whether they modified their burrow-emergence behaviour in response to olfactory stimuli from introduced predators, dogs, Canis familiaris, and cats, Felis catus. Greater bilbies have shared over 3000 years of co- evolutionary history with dogs but less than 200 years with cats. Bilbies spent more time only partially emerged (with at most head and shoulders out) as opposed to fully emerged (standing quadrupedally or bipedally) from their burrows when dog faeces were present, in comparison to faeces of cats, rabbits and an unscented control. Our results were consistent with the ‘ghosts of predator past’ hypothesis, which postulates that prey species' abilities to respond to the odours of predators scales with their period of coexistence. Our study supports the notion that introduced predators should be regarded as naturalized if prey possess an innate ability to detect their cues and respond accordingly.

Keywords: Anti-predator behaviour, Evolutionary history, Greater bilby, Ontogenetic naïveté, Prey naiveté hypothesis, Predator odour discrimination

2.2 INTRODUCTION

Hypotheses on the discrimination and recognition of predators by prey are divided as to whether a prey's ability is proportionate to the duration of co-evolution (Banks &

Dickman, 2007; Blumstein, 2002) or a result of ontogenetic experience (Berger, 1998) with specific predators, or whether prey simply generalize their response to all predators based on characteristics shared among predators (Apfelbach, Parsons, Soini,

& Novotny, 2015; Cox & Lima, 2006). The ‘ghosts of predators past’ hypothesis

(Peckarsky & Penton, 1988) suggests that species that have had a long period of co- evolution with a predator may possess ‘hard-wired’ antipredator recognition. Prey may exhibit innate abilities to recognize and respond to the scents and images of coevolved predators (Apfelbach, Blanchard, Blanchard, Hayes, & McGregor, 2005; Blumstein,

Daniel, Schnell, Ardron, & Evans, 2002; Monclús, Rödel, Von Holst, & De Miguel, 2005).

In contrast, prey species that have not been evolutionarily exposed to predators may learn through ontogenetic experience to recognize and respond to predators' olfactory cues (Anson & Dickman, 2013; Berger, Swenson, & Persson, 2001) or to their visual cues (Atkins et al., 2016).

The ‘predator archetype’ hypothesis suggests that for many prey species, their capacity to recognize and respond to cues associated with predators may be generalized and not be limited to specific predators (Cox & Lima, 2006). As a result, prey may exhibit antipredator responses towards cues that share characteristics with those with which they have coevolved or cohabited (Cox & Lima, 2006). For example, the ‘common constituents’ hypothesis posits that odours from predators share common compounds that prey should respond to regardless of the predator that

produced it (Apfelbach et al., 2015; Nolte, Mason, Epple, Aronov, & Campbell, 1994). It has also been suggested that a prey's ability to discriminate between predator odours is influenced by its body size (Apfelbach et al., 2015; Woolhouse & Morgan, 1995).

Small prey are more likely to encounter predators at close quarters and thus may have little opportunity to assess the threat posed by different predators (Apfelbach et al.,

2015). Hence smaller, less mobile prey species are predicted to fear almost any carnivore odour (McEvoy, Sinn, & Wapstra, 2008; Nolte et al., 1994).

In situations where the risk of predation is low or non-existent, the benefits of expressing antipredator behaviours may be outweighed by the costs of missed opportunities to obtain food resources or mates. Consequently, relaxed selection by predators on both ontogenetic and evolutionary timescales may result in ‘prey naïveté’, whereby species may have diminished antipredator behaviour and/or fail to recognize and/or mount effective responses against novel predators (Banks, 1998;

Blumstein, 2006; Blumstein, Daniel, & Springett, 2004; Goldthwaite, Coss, & Owings,

1990).

Understanding of the factors that dictate prey species' abilities to recognize and respond to predators is an important theoretical issue (Cox & Lima, 2006; Ferrari,

Messier, & Chivers, 2008; Parsons, Apfelbach, Banks, Cameron, Dickman, Frank et al.,

2017), as well as an important applied topic. Evolutionary prey naïveté towards introduced predators has been hypothesized to be a major factor contributing to population declines of native prey species and failed attempts to reintroduce locally extinct species (Moseby et al., 2011; Salo, Korpimӓki, Banks, Nordström, & Dickman,

2007). However, our understanding of the evolutionary timeframes necessary for prey

species to maintain or acquire appropriate responses to introduced predators is poorly known. Many of the studies that have investigated the evolved abilities of wild prey to recognize cues associated with coevolved and novel predator species did not control for variation in ontogenetic exposure to predators (Anson & Dickman, 2013; Carthey &

Banks, 2012, 2016). Thus, it remains possible that responses of prey species to predator cues reported in many studies were, to some extent, shaped by generalization (Dickman & Doncaster, 1984) or reflect a result of both an individual's lifetime experience and the history of evolutionary exposure to predators (Blumstein,

2006; Hettena, Munoz, & Blumstein, 2014).

Knowledge of the extent to which prey species' responses to predators are the result of co-evolution or learning and the timeframes required for appropriate antipredator responses to be lost or develop has direct application to the development of programmes that attempt to overcome the problem of prey naïveté (Moseby,

Blumstein, & Letnic, 2016; West, Letnic, Blumstein, & Moseby, 2017). Indeed, if prey species can adequately recognize and appropriately respond to introduced predators, than it may no longer be necessary to classify them as introduced but instead naturalized (Carthey & Banks, 2012).

Prey are able to detect and respond to the presence of predators through the use of sight, sound and smell (Banks, Bytheway, Carthey, Hughes,&Price, 2014;

Parsons,Apfelbach, Banks, Cameron,Dickman, Frank et al., 2017). In coevolved predator prey systems, prey often use predator odours as cues to detect predators, gauge risk and respond accordingly (Anson&Dickman, 2013; Apfelbach et al., 2005;

Apfelbach et al., 2015; Parsons, Apfelbach, Banks, Cameron, Dickman, Frank et al.,

2017). Prey use refuges, such as burrows, to avoid predation (Martín & López, 1999;

Sih, Petranka, & Kats, 1988). Since predators can move through landscapes, the risk of predation outside burrows fluctuates through time. Consequently, prey must decide when it is safe tomove in and out of a refuge (Martín & López, 2015; Parsons,

Apfelbach, Banks, Cameron, Dickman, Frank et al., 2017; Sih, 1997).We predicted that if a prey species is able to detect a predator first, than it will optimize the avoidance of predators through appropriate risk assessment strategies (e.g. Lima & Dill, 1990).

However, this relies on a prey's ability to rapidly recognize and discriminate between predator cues.

Here we evaluated some of the theories on predator recognition discussed in

Chapter 1. Mainly the idea that a prey's ability to respond to predator odours is influenced by the duration of co-evolution, as opposed to a generalized response to shared characteristics of predator cues (Table 1.1). From a conservation perspective, understanding the mechanisms driving prey-predator recognition, such as innate versus learned recognition and the rate of behavioural adaptation by prey to a novel predator is of great importance. The success rate of reintroductions into areas where predators are excluded was 82% compared to only 8% when species were released

‘beyond the fence’ (Short et al., 1992). However, completely isolating populations from all placental predators may inhibit the ability for ‘naïve’ species to develop and/or be selected for, for appropriate anti-predator recognition and responses necessary for reintroductions into environments where predators are present (Moseby et al., 2016).

Our certainty regarding life time and evolutionary predator experiences of bilbies should provide a unique insight into the influence that predation pressure can have in

the development of anti-predator behaviours and the time frames over which this can occur.

We did this by quantifying the behavioural responses of an ontogenetically predator naïve population of wild living greater bilbies, Macrotis lagotis, to faecal samples from two introduced predators (dog, Canis familiaris, and cat, Felis catus), a herbivore (rabbit, Oryctolagus cuniculus) and a procedural control (no odour). Bilbies have shared varying periods of co-evolution with these predators and rabbits.

Dingoes/dogs and feral cats are both known to predate on bilbies, and have been implicated in previous reintroduction failures of bilbies beyond predator-free fenced reserves (Moseby et al., 2011; Southgate & Possingham, 1995). Dingoes and cats are found in the same order, Carnivora. Dingos utilise a conservative feeding strategy

(Corbett and Newsome, 1987), preferring large, infrequent meals, which is a reflection of the competitive feeding behaviour of their pack hunting ancestors, the wolf (Canis lupus) (Bradshaw, 2006). In contrast, cats are exclusively solitary hunters, and as such hunt prey that are typically smaller in body mass than themselves, resulting in cats having to eat several small meals per day (Bradshaw, 2006). Cats are found throughout Australia in a range of habitats such as tropical rainforests, alpine regions, deserts and rural and urban landscapes (Abbott, 2002, Denny & Dickman, 2010), whilst dingoes are restricted in their range to coastal Vic and NSW, Queensland and central

Australia, as a result of the dingo fence and culling (Letnic, Ritchie & Dickman, 2012).

Since many mammalian predators scent mark features in the landscape, such as the burrows of prey species, by depositing urinary and faecal odours (Corbett, 1995;

Gorman & Trowbridge, 1989), we deployed faeces at the entrance of bilbies' burrows.

The decision to emerge from a refuge, such as a burrow, requires prey to estimate predation risk outside the shelter versus the benefits of potential rewards (Martín &

López, 1999; Sih, 1997).

If the duration of co-evolution with predators influenced bilbies' ability to respond to predators, we would expect that bilbies should be more wary when emerging from burrows when dog rather than cat faeces are present. Bilbies have had more than 3000 years of co-evolution with dogs/dingoes (Savolainen, Leitner, Wilton,

Matisoo-Smith, & Lundeberg, 2004), but less than 200 years of co-evolution with cats

(Abbott, 2002). If bilbies generalized their response to placental predators, we expected that bilbies would respond similarly to dogs and cats, but not respond to rabbits or the control (no faeces). Rabbits are an introduced herbivore, harmless to bilbies, with which bilbies have had less than 160 years historical exposure (Zenger,

Richardson, & Vachot-Griffin, 2003). We restricted our test to introduced predators to which the source populations would have been exposed in the 20th century and did not include the scent of a marsupial predator, the western quoll, Dasyurus geoffroii, with which they would have had a longer period of evolutionary coexistence. The reasons for not including quoll scent were threefold. First, quolls and bilbies have not coexisted in the wild for over 100 years (Morris et al., 2003). At the time of European settlement western quolls were abundant and occupied 70% of the Australian continent, including areas of arid and semi-arid Australia (Morris et al., 2003).

Predation by introduced predators, such as the cat and red fox, has had similarly devastating effects on the quoll as it has the bilby. Due to predation, the western quoll population has significantly declined and the population is now restricted to the southern tip of Western Australia (Morris et al., 2003). The restriction of the quoll to

southern Western Australia means that bilbies and quolls have become geographically isolated for over 100 years. Second, it was not possible for us to obtain scent samples from captive quolls at the time the study was conducted. Third, although quolls were a historical predator, they are not a current predator to bilbies, unlike feral cats and dingoes. It is for this reason we have used the odour of these placental predators, as we wished to see whether these results would have implications beyond predator free fenced reserves.

Even though we did not have scents of a marsupial predator, we are confident that our test of the hypothesis, that the duration of co-evolution with a predator influences predator recognition, was not confounded by ontogenetic experience, as the population of bilbies within our study site have not been exposed to placental predators for more than 16 years.

2.3 MATERIALS AND METHODS

2.3.1 Study Area

We studied bilbies within the 60 km2 fenced exclosure at Arid Recovery Reserve, South

Australia (12 300 ha, 30°29’S, 136°53’E). Arid Recovery Reserve is in the arid zone, with an average rainfall of 149.4 mm (from 1997 and 2015; Bureau of Meteorology, 2015).

A 1.8 m high predator-proof fence surrounds the reserve. Dingoes, foxes, cats and rabbits are absent from the fenced exclosures where the study was undertaken.

Locally extinct mammals, including bilbies, were reintroduced to the Arid Recovery

Reserve in 2000 following the eradication of predators and introduced feral herbivores, such as rabbits (Moseby, Hill, & Read, 2009). All the mammals

reintroduced to Arid Recovery are wild, as they are not given supplementary food and are exposed to avian and reptilian predators.

2.3.2 Study species

Greater bilbies are an omnivorous, burrowing, nocturnal and largely solitary marsupial

(Moseby, Cameron, & Crisp, 2012). Male bilbies weigh 800e2500 g and females weigh

600 – 1200 g (Southgate, Christie, & Bellchambers, 2000). The distribution of bilbies has contracted markedly since European settlement of Australia in 1788 and they now occupy just 20% of their former range (Southgate, 1990). This decline has been attributed primarily to predation by introduced red foxes, Vulpes vulpes, and feral cats

(Moseby & O'Donnell, 2003; Southgate, 1990), as well as dingoes/wild dogs (Pavey,

2006). Naïveté towards introduced predators (such as feral cats and red foxes) has been implicated in the decline of many Australian mammals (Moseby et al., 2012).

Bilbies have been successfully reintroduced to some areas and islands within their former range where feral cats and foxes are absent, intensively controlled or eradicated (Moseby & O'Donnell, 2003).

The reintroduced population of bilbies at Arid Recovery Reserve were sourced from captive stock from Monarto Zoo (Moseby & O'Donnell, 2003), which descended from wild individuals captured from deserts in Western Australia and the Northern

Territory (Moseby et al., 2011). Bilbies can produce a litter of one to three young, four times a year, and have a captive longevity of 5 – 9 years (Southgate et al., 2000). Based on the reproductive rate of the bilby and historical source of the population of bilbies at Arid Recovery, we assumed that this population, in the wild, has gone through five predator-naïve generations over the past 16 years.

2.3.3 Sources and storage of treatment odours

We used faeces from three species: domestic dogs, domestic cats and wild rabbits along with a procedural control, which was no faeces present. We used domestic dog scats as previous studies have shown that they are chemically indistinguishable from those of dingoes (Carthey, 2013). To overcome the issue of decomposition of faecal odours after deposition, domestic dog and cat faecal samples were collected immediately from private pet owners and local veterinary hospitals, stored and sealed in airtight zip lock bags, and frozen at minus 20 °C (Carthey, 2013). Wild rabbit faecal samples were collected fresh from rabbit warrens. Disposable gloves were always worn when handling faeces to prevent cross contamination of odours. As faecal samples were collected from private pet owners and local veterinary hospitals, from multiple individual sources, the total number of donor individuals was unknown; however, it may be approximated that samples were sourced from between two and

10 separate individuals of each species. As rabbit faeces were collected from a wild population, the number of source individuals is unknown. To take potential donor effects into account, faeces allocation was randomized. Since the diets of domestic pets were consistent between individuals and were made up of a mix of raw meats and pet foods, we did not consider diet to be a potential confounding source in analysis

(Carthey, 2013).

2.3.4 Bilby burrow emergence behaviour

A total of 18 wild individual bilbies (10 females, eight males) were caught and fitted with a 9g core tail mount with whip antenna radio transmitter (Sirtrack, Havelock

North, New Zealand)between August and October 2015. Transmitters were attached

according to the protocol of the South Australian National Parks and Wildlife Service

(Moseby & O'Donnell, 2003). Individuals were radio tracked daily to their diurnal burrows for 2 – 8 weeks, with experiments commencing at least 2 nights after transmitter attachment.

We used a repeated measures design in which each individual was presented with each odour treatment once according to a predetermined balanced order. We controlled for order effects experimentally and assessed these effects statistically.

Odour treatments were presented on every third night of tracking, for a single night. There were two ‘baseline’ nights, where no odour was presented, to ensure that there was no residual odour from the previous treatment. Faeces were presented on the surface of the ground, within 20 cm of the burrow entrance. If there were multiple burrow entrances, faeces were placed at the burrow entrance that recorded the strongest VHF transmitter signal. One piece of cat and dog faeces of similar size and weight (approximately 25 – 30 g) and 20 pellets of rabbit faeces were presented outside the burrow accordingly.

At each burrow entrance on treatment and ‘baseline’ nights a metal post was positioned approximately 1 – 2 m from the burrow entrance, supporting either a

Bushnell Trophy Cam (Bushnell, Overland Park, KS, U.S.A.), Scoutguard SG550V or

Scoutguard Zeroglow (Scoutguard, Molendinar, Australia), infrared motion sensor video camera. Cameras were mounted 20 – 100 cm off the ground and were programmed to take 60 s of video, when triggered, to enable species identification and observe burrow emergence and behavioural responses to the odour treatments (Fig.

2.1), with a 0 s interval between possible triggers, from dusk until dawn (1700 – 0700 hours).

Figure 2.1 Experimental set up for bilby predator odour discrimination study. Infrared motion sensor video camera mounted on a metal post outside the burrow entrance of a radio tracked bilby.

2.3.5 Behavioural scoring

We employed an “expert-based” (EB) method to reduce the number of variables for analysis and create five main behavioural groups, based upon the initial observations of experimental videos, relying on ethological knowledge and video observations (Table 2.1) (Mazzamuto et al., 2018). With the EB approach the researcher defines groups of behaviours, with each group related to specific behavioural responses. . All behaviours were treated as mutually exclusive (Blumstein

& Daniel, 2007). We scored video recordings ≥ 60 s using the event recorder JWatcher

(Blumstein & Daniel, 2007). We focused on quantifying only the first 60 s video footage of each bilby at the burrow entrance. We did this because our study focused on quantifying bilbies' initial behavioural responses to the presence of predators' scats and we wanted to eliminate the potential for our observations to be influenced by habituation to the presence of scats. We were unable to test for the effect of odour treatments on latency to emerge as bilby burrows often have more than one entrance and/or exit (Thompson & Thompson, 2008). Cameras were placed at the most active burrow entrance, however bilbies could still exit through an alternative exit. For this reason we focused our research on the initial burrow emergence behaviour and excluded latency to emerge.

We calculated the proportion of time in sight allocated to each behaviour. We quantified the behaviour of both identified bilbies (i.e. those with a tail transmitter), as well as other individuals that shared the burrows with marked subjects. Behavioural scoring of the videos commenced at the start of each 60 s video, with comparisons only made between ‘treatment’ nights. The inclusion of ‘no odour’ treatments ensured

we were able to compare behavioural responses to the different odour treatments and as such we did not compare ‘baseline’ and ‘treatment’ nights.

For analysis we combined behaviours in which the bilby was fully emerged from the burrow to create a new category ‘fully emerged’ (Table 2.1). It was not possible to record data blind because our study involved focal animals in the field and it was possible to visually identify the odour treatments.

Table 2.1 Ethogram of greater bilby (Macrotis lagotis) burrow emergence behaviour.

Behaviour Description a

Partially emerged Individual at burrow entrance, with at most head and shoulders out. Head fixated, potentially looking or sniffing or looking and sniffing.

Fully emerged Individual standing quadrupedally or bi-pedally, fully emerged from burrow. Head fixated, potentially looking or sniffing, or looking and sniffing.

Animal moving slowly when exiting and fully emerged Walk from burrow.

Animal moving rapidly when exiting and fully emerged Run from burrow.

Out of sight in Individual seen on camera and retreated back out of sight burrow into burrow. a Description of postures associated with a particular behaviour and/or behaviours

2.3.6 Analysis of behavioural data

We fitted a series of linear mixed-effects models in SPSS-22 (IBM Corp., Armonk, NY,

U.S.A.) with diagonal error structure to test bilby burrow emergence behaviour in response to treatment. A diagonal error structure assumes no correlation between the random effects (Seltman, 2009). Emergence behaviour (i.e. partially emerged, fully emerged etc.) was made the dependent variable. We had two fixed effects in our models: treatment and presentation order. To account for non-independence between observations, we included individuals as a random effect. In preliminary analyses, we also tested for the effects of moon phase; as this was never significant, however, we did not include it as a predictor variable in our final model. In no case was presentation order significant; we retained it as a blocking factor in the analysis, however, to control for its effect statistically (Quinn & Keough, 2002). In instances where the treatment effect was significant (P < 0.05), we used Fisher's least significant difference (LSD) post hoc test to examine planned comparisons (Fisher, 1935, Quinn & Keough, 2002) to understand the differences in response to each odour treatment. . We chose not to use Bonferroni post hoc as there is an increased chance of Type II errors (rejecting an incorrect null hypothesis) due to the substantial reduction in statistical power

(Nakagawa, 2004). In the field of behavioral ecology and animal behavior it is difficult to obtain large sample sizes due to practical and ethical reasons. This results in low sample sizes which than unacceptably increases the probability of Type II errors if

Bonferroni post hoc analysis is applied (Nakagawa, 2004). It is for these reasons that we employed LSD post hoc analysis, rather than Bonferroni.

Mazzamuto et al. (2018) found that Principal Component Analysis (PCA) scores correlated significantly with EB behaviours. To validate our EB approach with a statistical method, we ran a PCA of the five behaviours measured within this study. The initial factorial solutions were rotated by the Varimax procedure (Sokal & Rohlf 1995).

2.3.7 Ethical Note

Work was conducted under animal ethics APEC Approval Number 1/2014

‘Tackling Prey Naivety in Australia's Threatened Mammals' and in accordance with the

South Australian Wildlife Ethics Committee.

Bilbies were captured with either cage traps (45 x 20 cm and 20 cm high), baited with a combination of peanut butter and rolled oats, or hand-held fishing nets as described by Moseby et al. (2012). As bilbies did not readily enter the cage traps, 17 of 18 bilbies were captured with nets. Bilbies that were netted were located during night-time searches conducted with spotlights from a vehicle. When sighted, they were approached and netted with a hand-held net. On capture, bilbies were transferred from the net to a dark nylon fleece bag for processing and transmitter attachment. Bilbies were securely restrained within the processing bag, rather than anaesthetized during the attachment of the radio transmitter. The transmitter weighed 1.25% of an 800 g female and 0.07% of a 1400 g male bilby.

For transmitter attachment, hair on the tail of the bilby was removed using scissors and disposable razors, and a transmitter attached using Leukoplast adhesive tape. To prevent the formation of tail ulcers, extra care was taken to ensure that the transmitter was not firmly pressed to the tail (Moseby & O'Donnell, 2003). Only trained personnel were responsible for transmitter attachment.

To ensure that animal movements were not hindered by the capture and processing procedure, daily radio tracking of individuals commenced immediately after transmitter attachment. For 15 bilbies, the tail transmitters fell off after approximately

2 – 3 months.

For three bilbies, the transmitters did not fall off and were manually removed.

These bilbies were captured by placing cage traps near their burrows within a temporary pen constructed of wire netting (Southgate, McRae,&Atherton,1995). The bilbies were restrained as described above and the transmitters removed by cutting the tape with scissors. Each of the bilbies was deemed healthy on release; however, further checks were not possible because we could not locate individual bilbies without transmitters.

2.4 RESULTS

The PCA for burrow emergence behavioural measurements of bilbies in response to odour treatments produced two components that together accounted for 65.98% of the variance. The first component of the PCA (component 1, Eigen value = 2.218; % variance explained = 44.36) was positively correlated with variables describing fully emerged (rotated PCA score = 0.862, Fig. 2.2) and walk – slow locomotion behaviours

(rotated PCA score = 0.609, Fig. 2.2). Partially emerged behaviour was found to be negatively correlated with component 1 (rotated PCA score = -0.851, Fig. 2.2). The means derived from the factor analysis for component 1 for dog odour (predator) indicates that it is negatively correlated with PCA component 1 Fig.2.2). The means derived from the factor analysis for component 1 for cat, rabbit and the control were found to be positively correlated with component 1 Fig 2.2).Component 2 (Eigen value

= 1.081; % variance explained = 21.62) was characterised primarily by a positive correlation of run – fast locomotion (rotated PCA score = 0.977, Fig. 2.2) behaviour.

Based on the PCA it appears that individuals are trading off partial emergence against full emergence behaviours. There also appears to be a good correspondence between the grouping of behaviours from our EB method and that obtained with the PCA.

1.200

1.000 Run - Fast locomotion 0.800

0.600

0.400 Control

0.200

0.000

Partially -0.200 emerged Out of sight in Fully emerged Component 2 Component Dog Rabbit burrow Cat Walk - Slow locomotion -0.400 -1.000 -0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600 0.800 1.000 Component 1

Figure 2.2 Relationship between component scores derived (diamonds) from a principal component analysis of bilby behaviour (PCA) and the mean factor scores

(circles ± SEM) representing the behaviours displayed by bilbies emerging from their burrows at each of the odour treatments.

There was no effect of treatment on the proportion of time that bilbies spent out of sight in the burrow (F3, 29.836 = 0.036, P = 0.991; Fig. 2.3a), walking (F3, 34.225 =

0.634, P = 0.598; Fig. 2.3b) and running (F3, 11.195 = 1.054, P = 0.407; Fig. 2.3c, Table 2.2).

There was a significant effect of treatment on the proportion of time that bilbies spent partially emerged from their burrows, with at most their head and shoulders exposed (F3, 34.389 = 5.974, P = 0.002; Fig. 2.3d, Table 2.2). Planned comparisons (Fig. 2.3d) revealed that bilbies spent more time partially emerged when dog faeces were present compared to cat faeces (Fisher's LSD, dog versus cat: P =

0.013), rabbit faeces (Fisher's LSD, dog versus rabbit: P = 0.015) and the control (no faeces; Fisher's LSD, dog versus control: P ≤ 0.001). There were no significant differences in time spent partially emerged when cat and rabbit faeces (Fisher's LSD, cat versus rabbit: P = 0.922), cat faeces and the control (Fisher's LSD, cat versus control: P = 0.135), and rabbit faeces and the control were present (Fisher's LSD, rabbit versus control: P = 0.213; Fig. 2.3d).

There was a significant effect of treatment on the combined proportion of time spent fully emerged (F3, 32.283 = 3.134, P = 0.039; Fig. 2.3e, Table 2.2). Bilbies spent less time fully emerged from the burrow when dog faeces were present compared to the control (no faeces; Fisher's LSD, dog versus control: P = 0.006; Fig. 2.3e). There was no significant difference between time spent fully emerged when dog faeces were present compared to cat faeces (Fisher's LSD, dog versus cat: P = 0.180) and rabbit faeces (Fisher's LSD, dog versus rabbit: P = 0.078). There were no differences in the time spent fully emerged when cat and rabbit faeces (Fisher's LSD, cat versus rabbit: P

= 0.676), cat faeces and the control (Fisher's LSD, cat versus control: P = 0.184), and rabbit faeces and the control were present (Fisher's LSD, rabbit versus control: P =

0.407; Fig. 2.3e).

Table 2.2 Results from linear mixed-effects models to test bilby (Macrotis lagotis)

burrow emergence behaviour in response to treatment (cat, dog, rabbit, and control –

no odour), including order effects.

Behaviour df Denominator df F p-value

Out of sight in burrow 3 29.836 0.036 0.991

Order (out of sight in burrow) 1 28.710 1.370 0.251

Walk – slow locomotion 3 34.225 0.634 0.598

Order (walk) 1 36.687 1.404 0.244

Run – fast locomotion 3 11.195 1.054 0.407

Order (run) 1 8.527 2.630 0.141

Partially emerged 3 34.389 5.974 0.002**

Order (partially emerged) 1 33.989 0.320 0.575

Fully emerged 3 32.283 3.134 0.039**

Order (fully emerged) 1 35.761 0.021 0.886 Values in italic** indicate significant differences

Similar letters (e.g. a or b) above bars identify pairwise differences that are not statistically distinguishable (P > 0.05).

2.5 DISCUSSION

Our results provide support for the ‘ghosts of predators past’ hypothesis (Peckarsky &

Penton, 1988) which posits that prey species' ability to respond to predator cues scales

with the duration of their co-evolution. This finding was evidenced by the greater

proportion of time that bilbies spent partially emerged from the burrow as opposed to

fully emerged, when dog faeces were present. In contrast, bilbies spent

Proportion of time Proportion (seconds) of time proportionately more time fully emerged from their burrows when cat (an introduced

predator) and rabbit faeces (an introduced herbivore) and the procedural control (no

odour) were presented. Despite complete ontogenetic naïveté and at least 16 years of

evolutionary isolation, bilbies at the Arid Recovery Reserve appear to have retained

specific antipredator responses towards the olfactory cues of dogs/dingoes, but have a

negligible response to cats. Bilbies have shared over 3000 years of evolutionary history

with dogs/dingoes, compared to cats with which they have had less than 200 years of

evolutionary exposure. Our results support the idea that in coevolved predator–prey

systems, prey may possess innate abilities to detect the risk associated with predator

cues and respond accordingly, but lack this form of recognition when predators are

novel (Banks et al., 2014; Zhang, Zhao, Zhang, Messenger, & Wang, 2015).

Our results showed that, while partially emerged, bilbies appeared to

discriminate between the odours of dogs and cats. They similarly showed a weak

response to the odours of cats, harmless rabbits and the unscented control, while

partially emerged. These results contradict the ‘predator archetype’ hypothesis, which

suggests prey may exhibit a generalized response towards predator cues that share

characteristics with their coevolved predators (Cox & Lima, 2006). Our results further

contradict the ‘common constituents’ hypothesis, which suggests that odours from placental predators share common sulphur- and nitrogen-rich compounds that prey should respond to regardless of the predator that produced it (Apfelbach et al., 2015;

Nolte et al., 1994). These findings further suggest that bilbies responded most to the predator with which they have shared the longest period of co-evolution, rather than displaying a generalized response to predator odours.

Bilbies spent the greatest proportion of time partially emerged and the least amount of time fully emerged from the burrow when dog faeces were present. This finding may be due to bilbies making a trade-off between costs and benefits of staying within or leaving their refuges. Predator evasion is often costly in terms of time and energy. Thus, theory predicts that prey individuals should not flee or seek shelter immediately when they detect a predator, but instead should adjust their response according to the level of threat perceived (Ydenberg & Dill, 1986). Many animals modify their refuge use and burrow emergence behaviour according to the estimated levels of predation risk (Martín & López, 1999; Sih et al., 1988; Sparrow, Parsons, &

Blumstein, 2016). However, animals require information to make such decisions

(Bouskila & Blumstein, 1992). As such, by allocating more time to assessing the potential risks associated with the presence of dog faeces, while in the safety of their burrow entrances, bilbies may have reduced the potential for lethal encounters with a dog/dingo outside their burrow.

Our certainty regarding lifetime predator experiences in this study gave us unique insight into the influence of selection pressure in the retention and development of antipredator behaviours. We know the evolutionary history of

predator exposure of the bilby population at Arid Recovery Reserve. We also know that these bilbies have had no ontogenetic exposure to mammalian predators. This is in contrast to most other studies of free-ranging wildlife in which history of predator exposure is unknown. A study of wild bush rats, Rattus fuscipes, a species suspected to coexist with free ranging dogs, showed they had no aversion to dog faecal odours; however, it was acknowledged that the risk posed to rodents by feral dogs in the study area was unknown (Banks, Nelika, Hughes, & Rose, 2002).

A caveat of this study is that it is unable to test for whether predator recognition confers survival benefits. It is possible that although bilbies appear to be responding to the odour of their long term, historical predator, dingos, they may not be able to survive an encounter with this predator. Further research is needed to determine whether predator recognition and predator response are liked. It is also possible that the expression of certain behaviours by bilbies in response to dog and not cat odour treatments may be a reflection of threat sensitive predator avoidance strategies influenced by an evolutionary “memory” of the hunting styles of these predators (Bradshaw, 2006), as dingos may simply be more risky than cats. However this is unlikely given cats are known to predate on small prey species, such as bilbies, more commonly than dingos.

A further caveat is that we are unsure of the quantity and quality of the donors from which odour samples were collected. Previous studies have found that the quality of donor can influence a prey’s response to predator cues. In a study of convict cichlids body condition influenced the intensity of antipredator responses to conspecific alarm cues (Roh et al., 2004). The intensity of the antipredator response was greater

following exposure to the alarm cues of high condition stimulus versus low condition donors, suggesting that the chemical alarm cues from high condition donors are qualitatively and quantitatively greater than those of low condition donors (Roh et al.,

2004).

There has been little research into when an introduced predator may be considered naturalized. Carthey and Banks (2012) proposed that introduced predators should be considered native predators when their prey species are no longer naïve towards them. That bilbies with no lifetime exposure to mammalian predators appear to possess an innate ability to discriminate and respond to dog/dingo scent by being more reluctant to leave their burrows thus supports the idea that dingoes should be regarded as naturalized (Carthey & Banks, 2012; Frank, Carthey, & Banks, 2016). In contrast to their response to dog faeces, bilbies spent more time fully emerged and less time partially emerged from their burrows in the presence of cat faeces, rabbit faeces and the unscented control. This finding implies that bilbies are naïve towards cats and that less than 200 years of evolutionary exposure to cats may not be long enough for bilbies to develop and retain appropriate predator discrimination abilities

(Frank et al., 2016). Like the study by Frank et al. (2016), our study raises the question of how long is long enough before a novel predator, such as a feral cat, may be considered naturalized? In theory, this question could be answered by evaluating the magnitude of native prey's responses to introduced predator cues at many different locations and using time since predator arrival as a predictor variable. Finally, our finding that bilbies have limited ability to discriminate cat scent is also of great applied interest as it better defines the problem that reintroduction programmes of predator- naïve populations face. That is, native Australian mammals in the critical weight range

(Burbidge & McKenzie, 1989) facing entirely novel predators may not be able to identify them as a threat.

2.6 DATA ACCESSIBILITY

Data available from the Dryad Digital Repository: http://doi.org/10.5061/dryad.20tq5

2.7 ACKNOWLEDGEMENTS

We thank local pet owners and the Roxby Downs Vet Clinic - Whyalla Vet for supply of odours, Arid Recovery staff and many volunteers for their assistance with the study.

2.8 FUNDING

This research was funded by the Australian Research Council (ARC-Linkage Grant

(#LP130100173) to M.L., K.M. and D.T.B.) and Holsworth Wildlife Research Endowment

(to L.S and M.L).

2.9 REFERENCES

Abbott, I. (2002). Origin and spread of the cat, Felis catus, on mainland Australia, with

a discussion of the magnitude of its early impact on native fauna. Wildlife

Research, 29(1), 51–74. https://doi.org/10.1071/WR01011.

Anson, J. R., & Dickman, C. R. (2013). Behavioral responses of native prey to disparate

predators: Naiveté and predator recognition. Oecologia, 171(2), 367–377.

https://doi.org/10.1007/s00442-012-2424-7.

Apfelbach, R., Blanchard, C. D., Blanchard, R. J., Hayes, R. A., & McGregor, I. S. (2005).

The effects of predator odors in mammalian prey species: A review of field and

laboratory studies. Neuroscience & Biobehavioral Reviews, 29(8), 1123–1144.

https://doi.org/10.1016/j.neubiorev.2005.05.005.

Apfelbach, R., Parsons, M. H., Soini, H. A., & Novotny, M. V. (2015). Are single odorous

components of a predator sufficient to elicit defensive behaviors in prey species?

Frontiers in Neuroscience, 9(263). https://doi.org/10.3389/fnins.2015.00263.

Atkins, R., Blumstein, D. T., Moseby, K. E., West, R., Hyatt, M., & Letnic, M. (2016).

Deep evolutionary experience explains mammalian responses to predators.

Behavioral Ecology and Sociobiology, 70(10), 1–9.

https://doi.org/10.1007/s00265-016-2181-4.

Banks, P. B. (1998). Responses of Australian Bush Rats, Rattus fuscipes, to the odor of

introduced Vulpes vulpes. Journal of Mammalogy, 79(4), 1260–1264.

https://doi.org/10.2307/1383017.

Banks, P. B., Bytheway, J. P., Carthey, A. J., Hughes, N. K., & Price, C. J. (2014). Olfaction

and predator-prey interactions amongst mammals in Australia. In A. Glen, & C. R.

Dickman (Eds.), Carnivores of Australia: past, present and future (pp. 389–412).

Collingwood, Australia: CSIRO Publishing.

Banks, P. B., & Dickman, C. R. (2007). Alien predation and the effects of multiple levels

of prey naiveté. Trends in Ecology & Evolution, 22(5), 229–230.

https://doi.org/10.1016/j.tree.2007.02.006.

Banks, P., Nelika, K., Hughes, A., & Rose, T. (2002). Do native Australian small mammals

avoid faeces of domestic dogs? Responses of Rattus fuscipes and Antechinus

stuartii. Australian Zoologist, 32(3), 406–409.

https://doi.org/10.7882/AZ.2002.018.

Berger, J. 1998. Future prey: some consequences of the loss and restoration of large

carnivores, in: Caro, T. (Ed), Behavioral ecology and conservation biology,

Oxford University Press, New York, pp. 80-100.

Berger, J., Swenson, J. E., & Persson, I. L. (2001). Recolonizing carnivores and naïve

prey: Conservation lessons from pleistocene extinctions. Science, 291(5506),

1036–1039. https://doi.org/10.1126/science.1056466.

Blumstein, D. T. (2002). Moving to suburbia: Ontogenetic and evolutionary

consequences of life on predator-free islands. Journal of Biogeography, 29(5–6),

685–692. https://doi.org/10.1046/j.1365-2699.2002.00717.x.

Blumstein, D. T. (2006). The multipredator hypothesis and the evolutionary persistence

of antipredator behavior. Ethology, 112(3), 209–217.

https://doi.org/10.1111/j.1439-0310.2006.01209.x.

Blumstein, D. T., & Daniel, J. C. (2007). Quantifying Behaviour the JWatcher Way.

Sunderland, MA: Sinauer.

Blumstein, D. T., Daniel, J. C., Schnell, M. R., Ardron, J. G., & Evans, C. S. (2002).

Antipredator behaviour of red-necked pademelons: A factor contributing to

species survival? Animal Conservation, 5(4), 325–331.

https://doi.org/10.1017/S1367943002004080.

Blumstein, D. T., Daniel, J. C., & Springett, B. P. (2004). A test of the multi-predator

hypothesis: Rapid loss of antipredator behavior after 130 years of isolation.

Ethology, 110(11), 919–934. https://doi.org/10.1111/j.1439-0310.2004.01033.x.

Bouskila, A., & Blumstein, D. T. (1992). Rules of thumb for predation hazard

assessment: Predictions from a dynamic model. American Naturalist, 139(1),

161–176. https://doi.org/10.1086/285318.

Bradshaw, J. W. S. 2006. The evolutionary basis for the feeding behavior of domestic

dogs (Canis familiaris) and cats (Felis catus). The Journal of Nutrition

136(7):1927S-1931S. doi: 10.1093/jn/136.7.1927S

Burbidge, A. A., & McKenzie, N. L. (1989). Patterns in the modern decline of Western

Australia's vertebrate fauna: Causes and conservation implications. Biological

Conservation, 50(1–4), 143–198.

Bureau of Meteorology, A. G. (2015). Mean rainfall (mm) for for Roxby Downs, South

Australia, years 1997 to 2015. Retrieved from

http://www.bom.gov.au/climate/data.

Carthey, A. J. (2013). Naiveté, novelty and native status: Mismatched ecological

interactions in the Australian environment. PhD Thesis. Sydney, Australia:

University of Sydney.

Carthey, A. J., & Banks, P. B. (2012). When does an alien become a native species? A

vulnerable native mammal recognizes and responds to its long-term alien

predator. PLoS One, 7(2), e31804.

https://doi.org/10.1371/journal.pone.0031804.

Carthey, A. J., & Banks, P. B. (2016). Naiveté is not forever: Responses of a vulnerable

native rodent to its long term alien predators. Oikos, 125(7).

https://doi.org/10.1111/oik.02723.

Corbett, L. K. (1995). Dingo in Australia and Asia. Sydney, Australia: University of New

South Wales Press.

Cox, J. G., & Lima, S. L. (2006). Naiveté and an aquatic–terrestrial dichotomy in the

effects of introduced predators. Trends in Ecology & Evolution, 21(12), 674–680.

https://doi.org/10.1016/j.tree.2006.07.011.

Denny, E.A. & Dickman, C.R. (2010) Review of cat ecology and management strategies

in Australia. Invasive Animal Cooperative Research Centre, Canberra.

Dickman, C. R., & Doncaster, C. P. (1984). Responses of small mammals to Red fox

(Vulpes vulpes) odour. Journal of Zoology, 204(4), 521–531.

https://doi.org/10.1111/j.1469-7998.1984.tb02384.x.

Ferrari, M. C., Messier, F., & Chivers, D. P. (2008). Can prey exhibit threat-sensitive

generalization of predator recognition? Extending the predator recognition

continuum hypothesis. Proceedings of the Royal Society of London B: Biological

Sciences, 275(1644), 1811–1816. https://doi.org/10.1098/rspb.2008.0305.

Fisher, R. A. 1935. The design of experiments. Oliver and Boyd, Edinburgh.

Frank, A. S. K., Carthey, A. J. R., & Banks, P. B. (2016). Does historical coexistence with

dingoes explain current avoidance of domestic dogs? Island bandicoots are naïve

to dogs, unlike their mainland counterparts. PLoS One, 11(9), e0161447.

https://doi.org/10.1371/journal.pone.0161447.

Goldthwaite, R. O., Coss, R. G., & Owings, D. H. (1990). Evolutionary dissipation of an

antisnake system: Differential behavior by California and arctic ground squirrels

in above- and below-ground contexts. Behaviour, 112(3/4), 246–269.

suhttps://doi.org/10.1163/156853990X00220.

Gorman, M. L., & Trowbridge, B. J. (1989). The role of odor in the social lives of

carnivores. In J. L. Gittleman (Ed.), Carnivore Behaviour, Ecology, and Evolution

(pp. 57–88). Springer US.

Hettena, A. M., Munoz, N., & Blumstein, D. T. (2014). Prey responses to predator's

sounds: A review and empirical study. Ethology, 120(5), 427–452.

https://doi.org/10.1111/eth.12219.

Letnic, M., Ritchie, E.G. & Dickman, C.R., (2012). Top predators as biodiversity

regulators: the dingo Canis lupus dingo as a case study. Biological Reviews, 87

(2), 390–413. http://onlinelibrary.wiley.com/doi/10.1111/j.1469-

185X.2011.00203.x/full.

Lima, S. L., & Dill, L. M. (1990). Behavioral decisions made under the risk of predation:

A review and prospectus. Canadian Journal of Zoology, 68(4), 619–640.

https://doi.org/10.1139/z90-092.

Martín, J., & L_opez, P. (1999). When to come out from a refuge: Risk-sensitive and

state-dependent decisions in an alpine lizard. Behavioral Ecology, 10(5), 487–

492. https://doi.org/10.1655/07-083R2.1.

Martín, J., & Löpez, P. (2015). Hiding time in refuge. In W. E. Cooper, Jr., & D. T.

Blumstein (Eds.), Escaping from predators: an integrative view of escape

decisions (pp. 227–262). Cambridge, U.K.: Cambridge University Press.

McEvoy, J., Sinn, D. L., & Wapstra, E. (2008). Know thy enemy: Behavioural response of

a native mammal (Rattus lutreolus velutinus) to predators of different

coexistence histories. Austral Ecology, 33(7), 922–931.

https://doi.org/10.1111/j.1442-9993.2008.01863.x.

Monclús, R., Rödel, H. G., Von Holst, D., & De Miguel, J. (2005). Behavioural and

physiological responses of naïve European rabbits to predator odour. Animal

Behaviour, 70(4), 753–761. https://doi.org/10.1016/j.anbehav.2004.12.019.

Morris, K., Johnson, B., Orell, P., Gaikhorst, G., Wayne, A., & Moro, D. (2003). Recovery

of the threatened chuditch (Dasyurus geoffroii): A case study. In M. Jones, M.

Archer, & C. Dickman (Eds.), Predators with pouches: the biology of carnivorous

marsupials (pp. 435–451). Clayton, Australia: CSIRO Publishing.

Moseby, K. E., Blumstein, D. T., & Letnic, M. (2016). Harnessing natural selection to

tackle the problem of prey naïveté. Evolutionary applications, 9(2), 334–343.

https://doi.org/10.1111/eva.12332.

Moseby, K. E., Cameron, A., & Crisp, H. A. (2012). Can predator avoidance training

improve reintroduction outcomes for the greater bilby in arid Australia? Animal

Behaviour, 83(4), 1011–1021. https://doi.org/10.1016/j.anbehav.2012.01.023.

Moseby, K. E., Hill, B. M., & Read, J. L. (2009). Arid Recovery – A comparison of reptile

and small mammal populations inside and outside a large rabbit, cat and fox-

proof exclosure in arid South Australia. Austral Ecology, 34(2), 156–169.

https://doi.org/10.1111/j.1442-9993.2008.01916.x.

Moseby, K. E., & O'Donnell, E. (2003). Reintroduction of the greater bilby, Macrotis

lagotis (Reid) (marsupialia : Thylacomyidae), to northern South Australia:

Survival, ecology and notes on reintroduction protocols. Wildlife Research, 30(1),

15–27. https://doi.org/10.1071/WR02012.

Moseby, K. E., Read, J. L., Paton, D. C., Copley, P., Hill, B. M., & Crisp, H. A. (2011).

Predation determines the outcome of 10 reintroduction attempts in arid South

Australia. Biological Conservation, 144(12), 2863–2872.

https://doi.org/10.1016/j.biocon.2011.08.003.

Nakagawa, S. (2004). A farewell to Bonferroni: the problem of low statistical power

and publication bias. Behavioral Ecology, 15(6), 1044-1045.

https://doi.org/10.1093/beheco/arh107

Nolte, D. L., Mason, J. R., Epple, G., Aronov, E., & Campbell, D. L. (1994). Why are

predator urines aversive to prey? Journal of Chemical Ecology, 20(7), 1505–1516.

Parsons, M. H., Apfelbach, R., Banks, P. B., Cameron, E. Z., Dickman, C. R., Frank, A. S.,

et al. (2018). Biologically meaningful scents: A framework for understanding

predator–prey research across disciplines. Biological Reviews, 93(1), 98–114.

https://doi.org/10.1111/brv.12334.

Pavey, C. (2006). National recovery plan for the greater bilby Macrotis lagotis. Northern

Territory Department of Natural Resources, Environment and the Arts. Retrieved

from http://www.environment.gov.au/biodiversity/threatened/recovery-

plans/national-recovery-plan-greater-bilby-macrotis-lagotis.

Peckarsky, B. L., & Penton, M. A. (1988). Why do Ephemerella nymphs scorpion

posture: A ‘ghost of predation past’? Oikos, 53(2), 185–193.

https://doi.org/10.2307/3566061.

Quinn, G. P., & Keough, M. J. (2002). Experimental design and data analysis for

biologists. Cambridge, U.K.: Cambridge University Press.

Roh, E., Mirza, R., & Brown, G. (2004). Quality or quantity? The role of donor condition

in the production of chemical alarm cues in juvenile convict cichlids. Behaviour,

141, 1235-1248. https://doi.org/10.1163/1568539042729667

Salo, P., Korpimӓki, E., Banks, P. B., Nordström, M., & Dickman, C. R. (2007). Alien

predators are more dangerous than native predators to prey populations.

Proceedings of the Royal Society B: Biological Sciences, 274(1615), 1237–1243.

https://doi.org/10.1098/rspb.2006.0444.

Savolainen, P., Leitner, T., Wilton, A. N., Matisoo-Smith, E., & Lundeberg, J. (2004). A

detailed picture of the origin of the Australian dingo, obtained from the study of

mitochondrial DNA. Proceedings of the National Academy of Sciences of the

United States of America, 101(33), 12387–12390.

https://doi.org/10.1073/pnas.0401814101.

Seltman, H. J. (2009). Mixed models. A flexible approach to correlated data.

Experimental design and analysis, 357-377.

Short, J., Bradshaw, S. D., Giles, J., Prince, R. I. T. & Wilson, G. R. (1992).

Reintroduction of macropods (Marsupialia: Macropodoidea) in Australia—A

review. Biological Conservation, 62(3), 189-204. https://doi: 10.1016/0006-

3207(92)91047-V

Sih, A. (1997). To hide or not to hide? Refuge use in a fluctuating environment. Trends

in Ecology & Evolution, 12(10), 375–376. https://doi.org/10.1016/S0169-

5347(97)87376-4.

Sih, A., Petranka, J.W., & Kats, L. B. (1988). The dynamics of prey refuge use: A model

and tests with sunfish and salamander larvae. American Naturalist, 132(4), 463–

483. https://doi.org/10.1086/284865.

Southgate, R. (1990). Distribution and abundance of the greater bilby Macrotis lagotis

Reid (Marsupialia: Peramelidae). In J. H. Seeback, P. R. Brown, R. I. Wallis, & C. M.

Kemper (Eds.), Bandicoots and bilbies (pp. 293–302). Sydney, Australia: Surrey

Beatty.

Southgate, R., Christie, P., & Bellchambers, K. (2000). Breeding biology of captive,

reintroduced and wild greater bilbies, Macrotis lagotis (Marsupialia:

Peramelidae). Wildlife Research, 27(6), 621–628.

https://doi.org/10.1071/WR99104.

Southgate, R., McRae, P., & Atherton, R. (1995). Trapping techniques and a pen design

for the greater bilby Macrotis lagotis. Australian Mammalogy, 18, 101–104.

Southgate, R., & Possingham, H. (1995). Modelling the reintroduction of the greater

bilby Macrotis lagotis using the metapopulation model analysis of the likelihood

of extinction (ALEX). Biological Conservation, 73(2), 151–160.

Sparrow, E. E., Parsons, M. H., & Blumstein, D. T. (2016). Novel use for a predator

scent: Preliminary data suggest that wombats avoid recolonising collapsed

burrows following application of dingo scent. Australian Journal of Zoology,

64(3), 192–197. https://doi.org/10.1071/ZO15068.

Thompson, Graham G., and S. A. Thompson. (2008). Greater bilby (Macrotis lagotis)

burrows, diggings and scats in the Pilbara. Journal of the Royal Society of Western

Australia 91, 21-25.

West, R., Letnic, M., Blumstein, D. T., & Moseby, K. E. (2017). Predator exposure

improves anti-predator responses in a threatened mammal. Journal of Applied

Ecology, 0(0), 1–10. https://doi.org/10.1111/1365-2664.12947.

Woolhouse, A. D., & Morgan, D. R. (1995). An evaluation of repellents to suppress

browsing by possums. Journal of Chemical Ecology, 21(10), 1571–1583.

https://doi.org/10.1007/BF02035153.

Ydenberg, R. C., & Dill, L. M. (1986). The economics of fleeing from predators.

Advances in the Study of Behavior, 16, 229–249. https://doi.org/10.1016/S0065-

3454(08)60192-8.

Zenger, K. R., Richardson, B. J., & Vachot-Griffin, A. M. (2003). A rapid population

expansion retains genetic diversity within European rabbits in Australia.

Molecular Ecology, 12(3), 789–794. https://doi.org/10.1046/j.1365-

294X.2003.01759.x.

Zhang, F., Zhao, J., Zhang, Y., Messenger, K., & Wang, Y. (2015). Antipredator

behavioural responses of native and exotic tadpoles to novel predator. Asian

Herpetological Research, 6(1), 51–58. doi:

https://doi.org/10.16373/j.cnki.ahr.140023.

2.10 SUPPLEMENTARY MATERIAL

Table S2.1 Ethogram of greater bilby (Macrotis lagotis) behaviour.

Behaviour Description a Dig at burrow Dig at burrow entrance Dig outside burrow Dig anywhere within camera frame other than at burrow entrance STLk at burrow Individual at burrow entrance, with at most head and shoulders out. Head fixated, potentially looking or sniffing or looking and sniffing. STLk at burrow fully emerged Individual standing quadrupedally or bi-pedally, fully emerged from burrow. Head fixated, potentially looking. ReLk Individual standing quadrupedally or bi-pedally, fully emerged from burrow. Head fixated, potentially looking behind individual. ReLk sniff Individual standing quadrupedally or bi-pedally, fully emerged from burrow. Head fixated, potentially sniffing behind individual. STLk sniff Stand look sniff at burrow with at most head and shoulders out STLk sniff Stand look sniff at burrow fully emerged Walk Animal moving slowly when exiting and fully emerged from burrow. Run Animal moving rapidly when exiting and fully emerged from burrow. OS iB Individual seen on camera and retreated back out of sight into burrow. OS off camera Out of sight off camera Other Other behaviour/interactions with other species and individuals a Definition of postures associated with particular behaviours

CHAPTER 3:

The influence of evolutionary history with predators on olfactory predator recognition in endangered marsupials

3. 1 ABSTRACT

Anti-predator behaviour is developed through either lifetime experiences and/or evolutionary history, and may either be retained for thousands of years after isolation from predators or lost rapidly following isolation. Here we test the idea that a prey’s ability to respond to predator odours is influenced by a shared co-evolutionary history, as opposed to a generalised response to shared characteristics of predator cues. We studied two species of wild, ontogenetically predator naïve bettongs (burrowing bettongs, Bettongia lesueur and brush-tailed bettongs, Bettongia penicillata) living within large predator-free exclosures, to determine whether they modified their behavioural responses to ‘whole body’ odour samples from two predators (the native dingo, Canis familiaris and the introduced feral cat, Felis catus), an introduced herbivore (European rabbit, Oryctolagus cuniculus) and a procedural control (no odour). Brush-tailed bettongs have shared at least 3000 years of co-evolutionary history with dingos and 150 years with feral cats. In contrast, burrowing bettongs have shared a minimal evolutionary history with these placental predators. Burrowing bettongs did not appear to differentiate their behavioural responses to odour treatments of two types of predators (cat and dingo), herbivore (rabbit) and the procedural control (no odour). It is possible that burrowing bettongs did not differentiate between odours due to variations of the study design. In contrast, brush- tailed bettongs adjusted their behaviour in response to cat and dingo odours, allocating the most time to wary approach behaviour when a cat odour was present and retreating more from dingo and rabbit odour compared to the control. The similarities in response of brush-tailed bettongs to a potential competitor (rabbits) and a predator (dingo) are unexpected, however may reflect the limited discriminative

ability of this species, rather than the fact that it is not a fearful response. Our results suggest that a prey species ability to recognise predatory cues may be a result of their evolutionary history with those specific predators and they are able to retain predator recognition abilities, despite isolation from all terrestrial, mammalian predators over their lifetime. However, further research is needed in order to tease apart the unexpected similar patterns of response to an herbivore and carnivore.

Keywords: Anti-predator behaviour, Brush-tailed bettong, Burrowing bettong,

Evolutionary history, Ontogenetic naïveté, Predator odour discrimination

3.2 INTRODUCTION

Anti-predator behaviour is developed through either lifetime experiences or evolutionary history with predators, or a combination of both (Blumstein, 2002). Yet following isolation from predators, some anti-predator behaviours may be retained for thousands of years (Curio, 1966; Coss, 1999; Blumstein et al., 2000; Orrock, 2010), while others may be lost rapidly (Blumstein & Daniel, 2005; Jolly et al., 2018). If experience with a predator is important for the retention of anti-predator responses and recognition skills, then the complete isolation of prey from all predators over both ontogenetic and evolutionary time scales, may lead to the rapid loss of these abilities and cause ‘prey naïveté’ (Blumstein, 2002; Cox & Lima, 2006). Prey may be considered naïve when they fail to display appropriate anti-predator behaviours and/or fail to recognise and/or mount effective responses against novel predators (Goldthwaite,

Coss, & Owings, 1990; Blumstein, 2006; Carthey & Blumstein, 2018).

The prey naïveté hypothesis suggests that prey will be unable to recognise predator with which they have no ontogenetic or evolutionary history (Cox & Lima,

2006). Thus, for prey to recognise a potential predator, they must be able to perceive cues from that predator, such as odour, and match it against knowledge developed through ontogenetic and/or evolutionary history (Carthey & Banks, 2014; Steindler,

Blumstein, West, Moseby, & Letnic, 2018). Since selection should favour individuals that respond correctly to the presence of potential predators (McCormick & Holmes,

2006), and since learning through experience comes at a cost, (i.e. potentially the prey’s life (Griffin, Evans, & Blumstein, 2001)), it may be beneficial for prey to retain knowledge of their historically important predators, despite a lifetime of isolation.

It is potentially dangerous for prey to directly observe or have physical contact with predators (Powell & Banks, 2004). Thus, prey commonly rely on visual (Blumstein,

Daniel, Griffin, & Evans, 2000), acoustic (Deecke, Slater, & Ford, 2002), and/or olfactory cues (Ferrari, Gonzalo, Messier, & Chivers, 2007) to reduce their chances of encountering a predator. Nocturnal mammals have developed strong chemical sensory abilities (Monclús, Rödel, Von Holst, & De Miguel, 2005) that are crucial to finding food and avoiding predators (Bytheway, Carthey, & Banks, 2013). The difficulty for many prey is negotiating environments that contain odours of varying compositions, strengths and ages (Bytheway, Carthey, & Banks, 2013), as well as accurately recognising and responding appropriately to predatory risks (Parsons et al., 2018). Prey that are able to identify a predatory threat will exploit odour cues produced by predators to avoid them, under the assumption that the more concentrated the predator odour is, the higher the risk of imminent predation (Powell and Banks, 2004).

To avoid predation by novel predators, some prey have generalised predator recognition skills based on similarities between novel and known predators. This has been shown in mammals (Griffin, Evans, & Blumstein, 2001), reptiles (Webb, Du, Pike,

& Shine, 2009), amphibians (Ferrari & Chivers, 2009), and fish (Ferrari, Gonzalo,

Messier, & Chivers, 2007). The ability to generalise recognition from an unknown predatory threat can be beneficial for some species, because it allows them to avoid and survive a potentially lethal encounter with a novel predator (Ferrari, Crane, &

Chivers, 2016). Conversely, avoiding areas due to generalised aversion of predator cues, or modifying time budgets after detection of a predator cue, may be costly in terms of lost opportunities (Blumstein, 2002). To reduce the risks associated with the misidentification of predatory threats, the optimal response of prey to predator cues

should be fine-tuned through ontogenetic and evolutionary history (McEvoy, Sinn, &

Wapstra, 2008).

Australian mammals provide a unique opportunity to study the relative importance of evolutionary history on predator odour recognition abilities. The introduction of predators such as red foxes (Vulpes vulpes) and feral cats (Felis catus) has been identified as the major contributing factor to the extinction of 22 mammalian species over the last two centuries in Australia (1788 – current) and coincides with

European settlement (Johnson, 2006, Woinarski, Burbidge, & Harrison, 2015). It is believed that prey naiveté of native mammals towards introduced predators has resulted in limited reintroduction and translocation success of a number of species

(Moseby et al., 2011). Some of these species include the burrowing bettongs

(Bettongia lesueur) and brush-tailed bettongs (Bettongia penicillata). Both of these species fall within the critical weight range of small to medium sized mammals (0.35 to

5.5 kg) that have declined and/or become extinct since European settlement (Burbidge

& McKenzie, 1989) and have been successfully introduced to predator-free exclosures in mainland Australia (Short & Turner, 2000).

Here we test the idea that a prey’s ability to respond to predator odours is influenced by a shared co-evolutionary history, as opposed to a generalised response to shared characteristics of predator cues. We did this by quantifying the behavioural responses of ontogenetically predator naïve populations of burrowing and brush-tailed bettongs living within predator-free exclosures to ‘whole body’ odour samples from two predators (dingos, Canis familiaris and cats), a herbivore (rabbit, Oryctolagus cuniculus) and a procedural control (no odour). The original intention of these

experiments was that they would be stand-alone experiments and would not be compared, hence the slight variations in the methodology applied between the two species of bettong. However, since the overall question was the same, we wished to compare and contrast their responses to the treatments within each location, to assess the influence of evolutionary history on bettong odour recognition abilities and behaviour.

Bettongs were studied at two geographically separated fenced reserves, with the populations in these reserves differing in their evolutionary exposure to predators and rabbits. Burrowing bettongs were formerly found throughout much of arid and semi- arid Australia (Short and Turner, 1993; Short and Turner, 2000). Burrowing bettongs were driven extinct on mainland Australia in the 20th century primarily by predation from introduced predators, the red fox and feral cats (Short and Turner, 1993; Short and Turner, 2000). However, populations of burrowing bettongs persisted on Bernier,

Dorre and Barrow Islands off the Western Australia coast (Short and Turner, 1993;

Short and Turner, 2000). Individuals from these island populations were initially reintroduced to a predator-proof exclosure at Heirisson Prong on the mainland. The island populations of burrowing bettongs from Bernier Island and reintroduced populations on mainland Australia have had limited contact with placental mammalian predators. Rising sea levels and wave erosion meant Bernier Island became separated from mainland Australia approximately 8,000 years ago and it is thought to have been uninhabited prior to European exploration (Hancock et al., 2000). Feral cats and domestic dogs were noted as being present on Bernier and Dorre Islands in the early

20th century, however, dogs only persisted on the islands for a short period

(approximately 20 years) and feral cat populations did not persist into the second half

of the 20th century (Shortridge, 1910; Ride et al., 1962). Feral cat and red fox incursions occurred within the fenced reserve at Heirisson Prong in the 1990’s (Short and Turner,

2000). As a result of the presence of cats on Bernier Island and Heirisson Prong, it is likely that the bettongs were subjected to a low level of predation pressure by feral cats within their recent evolutionary history (less than 100 years). Dingoes are not known to have occurred on Bernier and Dorre Islands, though it remains possible that they may have been taken to the islands by indigenous people during the last 3,000 –

4,000 years (Atkins et al., 2016). Thus, we assume that the burrowing bettongs at Arid

Recovery have had minimal evolutionary exposure to dingoes or domestic dogs.

Brush-tailed bettongs were found throughout much of arid and semi-arid

Australia up until the late 19th and early 20th centuries (Marlow et al., 2015). By 1980, the brush-tailed bettong population had declined by 99% (Short et al., 2005) and was restricted to three small relict populations in Western Australia (two within Dryandra

Woodland and one at Tutanning Nature Reserve) (Wheeler and Priddel, 2009; Marlow et al., 2015). The severe range contraction of this species appears to have correlated with the spread of the red fox (Short, 1998). However, recent studies have shown that relict brush-tailed bettong populations are also vulnerable to predation by feral cats

(Marlow et al., 2015).

Feral cats were noted as being present in Western Australia from the 1850s, likely establishing in the south-west first (1850s) and moving in towards the semi-arid areas in the 1870s and 1880s (Abbott, 2002). Therefore, populations of brush-tailed bettongs have been subjected to a high level of predation pressure by feral cats within their recent evolutionary history (> 150 years). Dingoes are thought to have been

present on mainland Australia for between 3,000 – 4,000 years (Letnic et al., 2014).

Thus, we assume that brush-tailed bettongs at Scotia Sanctuary have had more than

3,000 years of co-evolution with dingoes.

Given their lack of ontogenetic experience and minimal evolutionary history with placental predators and rabbits, we expected that burrowing bettongs living within

Arid Recovery reserve would lack or have a negligible response to predator odours, rabbits and the procedural control. However, if the duration of evolutionary history influences predator odour recognition in prey species, we would expect that ontogenetically naïve brush-tailed bettongs at Scotia Wildlife Sanctuary should exhibit behavioural responses to predator odours (dingos and cats). Brush-tailed bettongs have co-evolved with dingoes for at least 3000 years and with cats for at least 150 years. We would also expect brush-tailed bettongs to display negligible responses to rabbit odours (a herbivore with less than 110 years of co-evolution) and the procedural control.

3.3 MATERIAL AND METHODS

3.3.1 Study species and study sites

3.3.1.1 Burrowing bettongs (Bettongia lesueur) at Arid Recovery Reserve

Burrowing bettongs are a small (800 – 2000 g), nocturnal, omnivorous, burrowing macropod (Short and Turner, 1993).. Subsequently, bettongs from the island populations and Herisson Prong were reintroduced to the predator-proof exclosure at

Arid Recovery (30°29′S, 136°53′E) in 1999/2000 (Richards, 2008; Moseby et al., 2011).

Arid Recovery has a mean annual rainfall of 149.6 mm (from 1997 to 2016) (Bureau of

Meteorology, 2016). A 1.8 m high fence, with floppy top and foot netting surrounds the Arid Recovery reserve (Moseby and Read, 2006), with all mammalian predators, including dingos, cats and foxes absent and excluded from the section of the reserve where experiments were conducted (Fig. 3.1).

Burrowing bettongs can produce a litter of one young, three times per year, are sexually mature at 7 – 8 months and have a captive longevity of 3 – 11 years (Short and Turner, 1999). If reproduction occurred at least once in their lifetime after they reached sexual maturation and based on the historical source of the population of burrowing bettongs at Arid Recovery, a minimum of 8,000 predator naïve generations could have occurred over the past 8,000 years.

3.3.1.2 Brush-tailed bettongs (Bettongia penicillata) at Scotia Sanctuary

Brush-tailed bettongs are a small (750 – 1850 g), nocturnal, omnivorous macropod.

Bettongs sourced from wild populations at Dryanda Woodland were translocated to predator free fenced exclosures in 1995 (Yeatman and Groom, 2012). Subsequently, in

2008 bettongs from predator free populations were translocated into the 80 km2 predator free fenced areas at Scotia Wildlife Sanctuary (NSW), where all mammalian predators and introduced herbivores have been removed and excluded (Page, 2008).

The Australian Wildlife Conservancy’s (AWC) Scotia Sanctuary (Fig. 3.2, 64 000-ha, 33°

10′S, 144° 10′E) is located on the boundary of the arid and semiarid climatic zones, approximately 150 km south of Broken Hill, with an average rainfall of 214.4 mm

(Bureau of Meteorology, 2015).

Figure 3.2 Map of Scotia Wildlife Sanctuary where brush-tailed bettong (Bettongia penicillata) odour recognition studies were conducted. Brush-tailed bettongs were reintroduced into the 40 km2 predator free fenced reserve in 2008.

Brush-tailed bettongs are sexually mature at 4 – 6 months and can produce a litter of one young, three times a year (Delroy et al., 1986, Thompson et al., 2015), and have a captive longevity of 6 – 8 years (Short and Turner, 1999). If reproduction occurred at least once in their lifetime after they reached sexual maturation and based on the historical source of the population at Scotia Wildlife Sanctuary, a minimum of

20 predator naïve generations could have occurred over the past 20 years.

3.3.2 Experimental rationale

We used video footage from infrared motion sensor cameras placed at foraging stations to gauge the behavioural response of bettongs to ‘whole body’ odour treatments of two predators (dingo and cat), as well as a herbivore (rabbit) and a procedural control (no odour). ‘Whole body’ odour, rather than faecal or urine samples were used, because ‘whole body’ odour represents imminent danger through proximity or likely revisitation of a predator (Carthey, 2013).

3.3.3 Sources and storage of treatment odours

We used the ‘whole body’ odour of three different species; dingo (n = 4, 2 females, 2 males), domestic cat (n = 14, 5 females, 6 males, 3 unknown), and domestic rabbit (n =

24, 6 females, 5 males, 13 unknown) and a procedural control (clean towel). Pieces of odour-impregnated towel were cut from a larger towel (50 x 40 cm) that had been placed in the sleeping quarters of captive animals of each species for a minimum of 14 to 28 days. This duration was chosen to ensure that towels were imbued with the

‘whole body’ odour of the animals (Carthey, 2013). Odour samples were obtained from multiple individual sources from private pets, animals owned by Kindifarm (Sydney),

Symbio Wildlife Park (Sydney), Taronga Western Plains Zoo (Dubbo) and animals housed at the Cat Protection Society (Sydney). To take potential donor effects into account, odour allocation was randomized for the experimental procedures.

We used 100 % cotton sand-coloured towels (to prevent odour-imbued towels being visually discriminated from clean ones), washed without detergent and air-dried before being placed with an animal. After towels had absorbed the odour for 14 – 28 days, towels were sealed in an airtight zip lock bag and stored at -20°C to -80°C for a

100

maximum of three months before use. Control towels were also washed without detergent, dried and kept in an open area for 28 days and then frozen at -20°C.

Disposable gloves were worn at all times when handling the towels. Instruments and work surfaces were cleaned with ethanol and then rinsed with water and dried between cutting each sample, to prevent cross-contamination of odour.

3.3.4 Population level vigilance behaviour field methods

3.3.4.1 Burrowing bettongs (Bettongia lesueur)

To investigate burrowing bettong’s response to predator odours a total of 64 odour stations were set along roads, approximately 400 m apart. Experiments were conducted across 16 nights (8 nights in October 2015 and 8 nights in March 2016)

(Table 3.1).

Table 3.1 The number of foraging stations (n) by treatment type deployed during the

2015 and 2016 sampling periods.

Treatment

Year Cat Dingo Rabbit Control Total GRAND TOTAL

2015 16 16 16 16 64 128 2016 16 16 16 16 64

101

To test whether odour stations were independent of one another, we tested for spatial autocorrelation in the residuals of the fitted values for each behaviour, using

Moran’s index (i), calculated in the spatial analyst module of ArcGis v10.2 (Fortin &

Dale, 2005). Spatial autocorrelation occurs when the value of a variable at any one location in space can be predicted by the values of nearby locations. The existence of spatial autocorrelation indicates that sampling units are not independent of one another. Despite burrowing bettongs travelling an average of 500 – 600 m per night

(Short & Turner, 2000), analysis showed experimental sites were independent (Table

S3.1 in the Supplementary information).

Each station comprised a metal post, positioned approximately 2 m from the road, supporting either a Bushnell Trophy Cam (Bushnell, USA), Scoutguard SG550V or

Scoutguard Zeroglow (Scoutguard, Australia), infrared motion sensor video camera.

Cameras were mounted 50 – 150 mm off the ground and were programmed to take

60-s videos when triggered to enable species identification and observe responses to the odour treatments. There was a 10-s interval between possible triggers, from dusk until dawn (1700-0700 hours). A food lure, approximately 100 g dog pellets

(Homebrand Adult Dog Food Beef & Vegetable; Woolworths, Australia) was buried just beneath the surface and mixed into the inedible substrate from the surrounding environment, approximately 5 m from the road. We swept a 50 cm area directly surrounding the food lure to detect tracks and validate visitation. The site was checked the next day for signs of foraging. Sites were reset the following day, with approximately 100 g of dog pellets, in the late afternoon till dusk for that night.

Burrowing bettongs were allowed to acclimatise to feeding from the foraging sites for one night. On the second afternoon, odour treatments were applied for that night.

102

Treatments were randomly allocated to stations and a piece (10 x 10 cm) of predator, herbivore or control towel was stationed at the foraging site with a peg (Fig. 3.3a, b).

The towel was placed on the side opposite the camera.

a b

103

3.3.4.2 Brush-tailed bettongs (Bettongia penicillata)

To investigate brush-tailed bettongs responses to predator odours a total of 112 odour stations were set along roads, approximately 400 m apart (Table 3.2).

Table 3.2 The number of foraging stations (n) by treatment type deployed across night

2 and 3.

Treatment

Night Cat Dingo Rabbit Control Total GRAND TOTAL

2 28 28 28 28 112 224 3 28 28 28 28 112

The average home range of a brush-tailed bettong is 0.654 km2 (Yeatman and

Wayne, 2015) and analysis of spatial autocorrelation in the residuals of the fitted values for each behaviour showed experimental sites were independent (Table S3.1 in the Supplementary information). Experiments were conducted across 12 nights in June

2015. Each station had a camera placed using the procedure as described above for burrowing bettongs at Arid Recovery. At each station, a foraging tub (round plastic basin, ~ 20 L) was buried approximately 5 m from the road, with the rim of the basin set flush with the surface of the soil. Each tub was filled with 200 (±1) g of dog pellets

(Great Barko; Laucke Mills; Australia) mixed into approximately 20 L of inedible substrate from the surrounding environment. The area directly surrounding the tub was swept smooth to a radius of 50 cm, in order to detect tracks and validate

104

visitation. Tubs were reset with 200 (±1) g of dog pellets on the second and third night of treatment, in the late afternoon till dusk, for that night.

Brush-tailed bettongs were allowed to acclimate to feeding from the buried tubs for one night. Following this, odour treatments were applied for the next two nights, where odour stations received fresh odour on both treatment nights. The odour from treatment night one was completely removed and replaced with a fresh sample from the same type of animal on treatment night two to ensure this did not constitute over-marking or revisitation. Treatments were randomly allocated to stations, and a piece (10 x 10 cm) of predator, herbivore or control towel was adhered to the foraging tray with double-sided tape (Fig. 3.4a, b). The towel was placed opposite the camera.

105

a b

Figure 3.4 Experimental set up for brush-tailed bettong (Bettongia penicillata) population level predator odour recognition experiment at Scotia Wildlife Sanctuary

(a) 10 x 10 cm odour impregnated towel (cat, dingo, rabbit or control – no odour), positioned at foraging site, 5 m off the road, and (b) an infrared motion sensor video camera mounted on a metal stake 2 m from the foraging site.

3.3.5 Behavioural scoring

(Table 3.3) (Mazzamuto et al., 2018). With the EB approach the researcher defines groups of behaviours, with each group related to specific behavioural responses. We then scored video recordings ≤ 60-s using the event recorder JWatcher (Blumstein &

Daniel, 2007) from which we calculated the proportion of time in sight allocated to each behaviour. Since prey may habituate to the presence of an odour cue when not

106

accompanied by a predator (Parsons & Blumstein, 2010), we quantified the behaviour of the first bettong seen on camera to approach the treatment site. Behavioural scoring commenced when a bettong was within the field of view of the camera. Videos with more than one individual foraging were excluded to reduce the influence of conspecific presence on the behavioural response to the treatments. We also excluded any video where it could clearly be seen that the site had already been foraged, to ensure we were capturing the response of the first visit to the site. Sites were identified as ‘foraged’ if the tubs or foraging sites were no longer flush to the surface of the surrounding area. We were unable to group individuals into age or sex categories given the limitations of infrared camera technology.

For analysis we combined behaviours in order to form five main categories.

“Wary approach” was comprised of any sniffing behaviour, including bi-pedal and quadrupedal sniffing and slow movement towards the feeding station. “Bold approach” comprised of rapid movement towards the feeding station, as well as relaxed grooming behaviours. “Foraging” included foraging with the head raised and lowered. “Investigate” comprised of any sniffing, touching or chewing behaviour of the odour treatments. “Retreat” included rapid and slow movement away from the foraging site. Videos were scored blind with respect to treatment.

107

Table 3.3 Ethogram of burrowing bettong (Bettongia lesueur) and brush-tailed bettong

(Bettongia penicillata) behaviour.

Behaviour Category Behaviour Description a

Wary approach Bi-pedal sniff Animal looks and/or sniffs air whilst standing upright on hind limbs

Prone sniff Animal looks and/or sniffs the air and/or ground whilst standing on all four limbs

Slow approach Animal moves slowly towards foraging site

Bold approach Fast approach Animal moves quickly and directly towards foraging site

Grooming Animals grooms itself using grooming claw on hind limb and/or mouth

Foraging Vigilant foraging Animal chews with its head up and observing surroundings

Relaxed foraging Animal forages for food and chews with its head down without observing surroundings

Investigate Investigate towel Olfactory investigation of towel

Chew towel Animal attempts to chew and/or remove towel away from foraging site

Retreat Recoil Animal recoils away from the foraging site and/or visual treatment

Retreat Animal retreats away from foraging site

a Definition of postures associated with particular behaviours

108

3.3.6 Analysis of behavioural data

We fitted generalized estimating equation (GEE) models with independent error structure and binomial distribution in SPSS-25 (IBM Corp. Armonk, NY, U.S.A) to investigate if bettong visitation to feeding stations on treatment nights were dependent on the type of treatment placed at the feeding station. The response variable was visit (1) or no visit (0).

Because our studies were conducted on different species, across different sites

(Arid Recovery reserve and Scotia Wildlife Sanctuary) and differed in methodology

(number of treatment nights and quantity of food lure), we were unable to directly compare our data sets statistically. However, we have contrasted their responses to the treatments within each location, to assess the influence of evolutionary history on bettong odour recognition abilities and behaviour.

Mazzamuto et al. (2018) found that Principal Component Analysis (PCA) scores correlated significantly with EB behaviours. To validate our EB approach with a statistical method, we ran a PCA of the five behaviours measured within this study for burrowing and brush-tailed bettongs independently. The initial factorial solutions were rotated by the Varimax procedure (Sokal & Rohlf 1995) (Fig. 3.5 and 3.6).

To test whether odour treatment caused burrowing and brush-tailed bettongs to allocate different proportions of time to composite behaviours we fitted a series of generalized estimating equation (GEE) models with an independent correlation matrix and linear distribution (Horton and Lipsitz, 1999, Quinn and Keough, 2002). The factor in our model was odour treatment (cat, dingo, rabbit and control). Sampling period was made a covariate. For burrowing bettongs sampling period was year (2015 and

109

2016), whilst for brush-tailed bettongs it was night (2 and 3). Foraging station was made a repeated measure in the models. We used a GEE model, rather than a

Generalized Linear Mixed Model (GLMM), as GEEs allow analysis of a repeated measures design (Quinn and Keough 2002). A GEE model can handle correlated non- normally distributed data, which is very common in the behavioural sciences (Pekár and Brabec, 2017). They are most suitable when the main factor or interest is between subjects and within subjects components represent repeated observations through time (Omar et al., 1999, Quinn and Keough 2002). GEEs are also able to handle missing data effectively, as long as the observations missing are completely random, unlike classical ANOVA type models for repeated measure designs that are unable to handle missing observations very effectively (Quinn and Keough 2002). One of the strengths of GEEs is that, although the correct specification of the correlation structure makes estimation more efficient, parameter estimates are usually consistent even if the wrong correlation structure is used, meaning the estimates of the model parameters are not very sensitive to the choice of correlation structure (Quinn and Keough, 2002,

Omar et al. 1999). A limitation to a GEE is that it is very efficient when the design of a study can be modelled using any of the predefined structures, however if the design is more complicated, such as more random effects either crossed or nested (multiple level design), than use of a GLMM is more efficient (Pekár and Brabec, 2017). A further limitation to the use of GEEs it that they are most applicable when the pattern of observation through time for experimental units is not the main research question

(Omar et al. 1999). GEEs are less useful if the within-subjects component is a factor of specific interest (Quinn and Keough 2002). The data fulfilled all the assumptions required to perform a GEE analysis.

110

In instances where the effect of odour was significant (P < 0.05), we used Fisher’s

Least Significant Difference (LSD) post-hoc analysis to examine planned comparisons

(Fisher, 1935, Quinn and Keough, 2002) to understand the differences in response to each odour treatment. We investigated the planned comparisons: dingo vs. cat, dingo vs. rabbit, dingo vs. control, cat vs. rabbit, cat vs. control and rabbit vs. control. We chose not to use Bonferroni post hoc as there is an increased chance of Type II errors

(rejecting an incorrect null hypothesis) due to the substantial reduction in statistical power (Nakagawa, 2004). In the field of behavioral ecology and animal behavior it is difficult to obtain large sample sizes due to practical and ethical reasons. This results in low sample sizes which than unacceptably increases the probability of Type II errors if

Bonferroni post hoc analysis is applied (Nakagawa, 2004). It is for these reasons that we employed LSD post hoc analysis, rather than Bonferroni.

111

3.4 RESULTS

3.4.1 Visits to Station

There was no effect of odour treatment on visitation to feeding stations by burrowing bettongs (Wald χ2 = 0.281, P = 0.964, Table 3.4) or for brush-tailed bettongs (Wald χ2 =

1.551, P = 0.172, Table 3.4). Thus, bettongs were equally likely to visit the stations regardless of the specific odour treatments present.

Treatment Treatment Treatment Treatment

Cat Dingo Rabbit Control

Burrowing bettongs 22 (8) 22 (17) 23 (14) 21 (8)

Brush-tailed bettongs 54 (21) 51 (27) 53 (31) 51 (22)

3.4.2 Behavioural response to odour treatment

Burrowing bettongs: The PCA for behavioural measurements of burrowing bettongs in response to odour treatments produced two components that together accounted for

57.41% of the variance. The first component of the PCA (component 1, Eigen value =

1.690; % variance explained = 33.79) was positively correlated with variables describing wary approach behaviour (rotated PCA score = 0.932, Fig. 3.5). The means

112

derived from the factor analysis for component 1 for dingo odour were positively correlated with PCA component 1 (Fig. 3.5). Foraging behaviour was found to be negatively correlated with component 1 (rotated PCA score = -0.870, Fig. 3.5). The means derived from the factor analysis for component 1 for cat, rabbit and procedural control – no odour were also found to be negatively correlated with component 1 (Fig.

3.5). Component 2 (Eigen value = 1.181; % variance explained = 23.62) was characterised primarily by a positive correlation of investigate (rotated PCA score =

0.993, Fig. 3.5) behaviour and a negative correlation of fast approach (rotated PCA score = -0.052, Fig. 3.5) and retreat (rotated PCA score = -0.046) behaviours. Although burrowing bettongs appear to be trading off wary approach and forage behaviours, this does not appear to be in response to the odour treatments. As such there appears to be a good correspondence between the grouping of behaviours from our EB method and that obtained with the PCA.

We found no significant effect of odour treatment on the proportion of time burrowing bettongs allocated to wary approach, forage, retreat, fast approach and investigate behaviours (Table 3.5, Fig 3.6a – d). There was no spatial autocorrelation in the residuals of the fitted values for any of the analysed behaviours for burrowing bettongs (Table S3.1 in the Supplementary information). These results indicate that the feeding stations were independent for the purpose of our analysis.

113

Burrowing bettong 1.200

1.000 Investigate

0.800

0.600

0.400 Control

0.200 Cat

0.000 Retreat Rabbit

Component 2 Component Fast Approach -0.200 Dingo -0.400 Wary Approach Forage -0.600 -1.000 -0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600 0.800 1.000 1.200 Component 1

114

Figure 3.6 The mean (± SEMs) proportion of time in sight (PIS) that burrowing bettongs

(Bettongia lesueur) allocated to the behaviours in response to odour treatment (cat n =

8, dingo n = 17, rabbit n = 14 and a procedural control n = 8) in the first minute of visitation to the foraging station. No letters indicate there are no significant differences.

115

Brush-tailed bettongs: The PCA for behavioural measurements of brush-tailed bettongs in response to odour treatments produced two components that together accounted for 65.85% of the variance. The first component of the PCA (component 1,

Eigen value = 2.167; % variance explained = 43.34) was positively correlated with variables describing wary approach (rotated PCA score = 0.819) and retreat behaviours

(rotated PCA score = 0.627). The means derived from the factor analysis for component 1 for predator odours (cat and dingo) were also positively correlated with

PCA component 1 (Fig. 3.7). Foraging behaviour was found to be negatively correlated with component 1 (rotated PCA score = -0.969). The means derived from the factor analysis for component 1 for the non-predator treatment odours (rabbit and procedural control – no odour) were also found to be negatively correlated with component 1 (Fig. 3.7). Component 2 (Eigen value = 1.126; % variance explained =

22.51) was characterised primarily by a positive correlation of fast approach (rotated

PCA score = 0.912) behaviours and a negative correlation of investigate behaviours

(rotated PCA score = -0.528). As such there appears to be a good correspondence between the grouping of behaviours from our EB method and that obtained with the

PCA.

116

Brush-tailed bettong 1

0.8 Fast Approach

0.6

0.4

0.2 Control Dingo 0 Retreat Rabbit Cat -0.2 Forage Wary Approach

-0.42 Component

-0.6 Investigate -0.8

-1 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Component 1

We found there was a significant effect of odour treatment on the proportion of time that brush-tailed bettongs allocated to wary approach (Wald χ2 = 7.914, P = 0.048,

Table 3.5, Fig. 3.8a) and retreat (Wald χ2 = 9.138, P = 0.028, Table 3.5, Fig. 3.8c). We found no significant effect of odour treatments on the proportion of time that brush- tailed bettongs allocated to foraging, fast approach and investigate behaviours (Table

3.5, Fig. 3.8b, d and e). There was no spatial autocorrelation in the residuals of the fitted values for any of the analysed behaviours (Table S3.1 in the Supplementary information). These results indicate that the feeding stations were independent for the purpose of our analysis.

117

Planned post-hoc comparisons revealed that brush-tailed bettongs spent significantly more time warily approaching cat odour compared to dingo, rabbit and the control (Fishers LSD, cat vs. dingo, P = 0.048, cat vs. rabbit, P = 0.013, and cat vs. control, P = 0.014, Fig. 3.8a). Brush-tailed bettongs did not modify time spent warily approaching dingo or rabbit compared to the control (Fishers LSD, dingo vs. control, P

= 0.488, and rabbit vs. control, P = 0.946, Fig. 3.8a). There was no significant difference between the proportion of time spent in wary approach between dingo and rabbit

(Fishers LSD, dingo vs. rabbit, P = 0.517, Fig. 3.8a).

Brush-tailed bettongs spent significantly more time retreating from dingo and rabbit odour compared to the control (Fishers LSD, dingo vs. control, P = 0.026, rabbit vs. control, P = 0.015, Fig. 3.8d). There was no difference in the proportion of time spent retreating between the other odour treatments (Fishers LSD, dingo vs. cat, P =

0.109, dingo vs. rabbit, P = 0.375, and cat vs. rabbit, P = 0.216, Fig. 3.8d). There was no difference in time spent retreating from cat odour compared to the control (Fishers

LSD, cat vs. control, P = 0.233, Fig. 3.8d).

118

Species of Bettong Behaviour df Wald χ2 p-value

Burrowing bettong Wary Approach 3 0.797 0.850

Forage 3 3.970 0.265

Retreat 3 3.907 0.272

Fast Approach 3 4.541 0.209

Investigate 3 2.446 0.485

Brush-tailed bettong Wary Approach 3 7.914 0.048**

Forage 3 3.478 0.324

Retreat 3 9.138 0.028**

Fast Approach 3 1.256 0.740

Investigate 3 1.791 0.617

Values in italic** indicate significant differences

119

Figure 3.8 The mean (± SEMs) proportion of time in sight (PIS) that brush-tailed bettongs (Bettongia penicillata) allocated to the behaviours in response to odour treatment (cat n = 21, dingo n = 27, rabbit n = 31 and a procedural control n = 22) in the first minute of visitation to the foraging station. Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05).

No letters indicate there are no significant differences.

120

3.5 DISCUSSION

The results from our study suggest that a prey species’ ability to recognise predatory cues may be a result of their evolutionary history with those specific predators (Griffin,

Blumstein, & Evans, 2000). Burrowing bettongs did not differentiate their behavioural responses to odour treatments of two types of predators (cat and dingo), herbivore

(rabbit) and the procedural control (no odour). These findings suggest that burrowing bettongs do not recognise odour cues of predators and are unable to differentiate between carnivore and herbivore ‘whole body’ odours and thus are ‘naïve’.

In stark contrast, brush-tailed bettongs adjusted their behaviour in response to cat and dingo odours. Brush-tailed bettongs allocated the most time to wary approach behaviour when a cat odour was present. Brush-tailed bettongs also spent significantly more time retreating from the dingo and rabbit odours compared to the control. That brush-tailed bettongs were significantly more wary of cat odour and allocated the most time retreating from dingo odour, suggests that odour recognition is associated with co-evolutionary history. This finding is consistent with previous studies that have found that prey may be able to retain predator recognition abilities, despite isolation from all predators over their lifetime (Curio, 1966; Coss, 1999; Steindler, Blumstein,

West, Moseby, & Letnic, 2018).

Brush-tailed bettongs also appeared to respond to rabbit odour, expressed through retreat behaviour. One hypothesis to explain this response is that rabbits have been competitors with brush-tailed bettongs since their introduction to Australia

(Zenger, Richardson, & Vachot-Griffin, 2003), and this competition has resulted in the evolution of an aversive response to rabbits. The similarities in response of brush-

121

tailed bettongs to a potential competitor (rabbits) and a predator (dingo) could also reflect the limited discriminative ability of this species, rather than the fact that it is not a fearful response. The use of herbivore odour as a control within this study may have confounded our results. Most other studies of free-ranging wildlife do not account for a herbivore control and rather test for behavioural responses to a number of predators, with a procedural control of no odour (Carthey and Banks, 2016, Mella et al., 2010, Russell and Banks, 2007). Future studies assessing bettong predator odour recognition may have to exclude the use of a herbivore. It is also possible that the response elicited by bettongs in response to rabbits may be a result of the use of

‘whole’ body odour, rather than faeces. ‘Whole body’ odour represents imminent danger through proximity or likely revisitation of a predator (Carthey, 2013), whilst faeces indicate that a predator has been active in the area (Powell and Banks, 2004;

Monclús et al., 2005), however may not currently be present. In order to discern the influence of the type of odour used and whether this elicits different behavioural responses within the studied species, further field tests would need to be conducted that account for these differences.

Although both the burrowing and brush-tailed bettong populations used within this study exist within predator free fenced reserves and have done so for 16 and 20 years respectively, only brush-tailed bettongs showed any evidence of differentiation in their behavioural responses towards predator, herbivore and the control odours.

Burrowing bettongs have been isolated from placental predators for approximately

8,000 years (Hancock, Brown, & Stephens, 2000). Although populations of burrowing bettongs on islands off the Western Australia coast may have been exposed short term to feral cats and domestic dogs (Shortridge, 1910; Ride & Fraser, 1962), the results

122

from this study suggest that this experience was not sufficient to develop the ability to discriminate between predatory and non-predatory odours, at least in the absence of super-imposed developmental experience. A lack of evolutionary and ontogenetic experience with predators means that burrowing bettongs appear to be unable to differentiate between the odours of two predators (dingo and cat), herbivore (rabbit) and control (no odour).

In contrast, brush-tailed bettongs which have had more than 3,000 years of evolutionary history with dingoes (Letnic, Fillios, & Crowther, 2014) and over 150 years with cats (Abbott, 2002); appear to respond to the odour of both predators. Brush- tailed bettongs invested significantly more time retreating from dingo odour and significantly more time warily approaching cat odour compared to the control. These results suggest that between 150 – 3,000 years of evolutionary history may be sufficient for brush-tailed bettongs to develop hard-wired predator odour discrimination abilities towards cats and dingos. These results similarly support those found by Anson and Dickman (2013), which suggest that 150 years of co-evolutionary history may be sufficient for some native ‘naïve’ species to develop appropriate predator recognition of novel introduced predators. However caution should be taken when interpreting these results as brush-tailed bettongs did also respond to odour of a herbivore (rabbit).

Prey may exhibit predator-specific responses, behaving differently in the presence of odour from different predators (Jedrzejewski, Rychlik, & Jedrzejwska,

1993). Brush-tailed bettongs spent significantly more time warily approaching cat odour, than they did dingo, rabbit and the control. Conversely, brush-tailed bettongs

123

also spent the greatest proportion of time retreating from dingo odour compared to the control. The threat sensitive hypothesis suggests that prey alter their predator avoidance response in a manner that reflects the magnitude of predatory threat

(Helfman, 1989). The expression of certain behaviours by brush-tailed bettongs in response to odour treatments may also be a reflection of threat sensitive predator avoidance strategies influenced by an evolutionary “memory” of the hunting styles of these predators (Bradshaw, 2006).

It has been suggested that the generalisation of predator recognition plays a key role in naïve prey responding to novel, introduced predators (Ferrari, Crane, & Chivers,

2016), with some prey displaying a strong aversive response towards the odours of both familiar and unfamiliar predators (Ferrero et al., 2011). ‘Whole body’ odour is a complex odour composed of secretions from the entire animal, including scent glands, sebaceous and skin secretions, waste secretions and potentially salivary secretions

(Ferrero et al., 2011; Carthey & Banks, 2014; Apfelbach, Parsons, Soini, & Novotny,

2015). Some prey may utilise the similarities in the scent compounds of predators, resulting from a carnivorous diet (Kats & Dill, 1998), to generalise their response to all predator odours. Previous studies have also suggested that the size of a prey animal will influence their predator discrimination abilities, with small (< 2 kg) and medium (2

– 10 kg) sized prey animals more sensitive and less discriminatory of unfamiliar predators (Apfelbach, Parsons, Soini, & Novotny, 2015). As such, small to medium sized prey, such as bettongs, are presumed to fear almost any sulphur rich or nitrogen containing odour, regardless of the predator that produced it (Spencer, Crowther, &

Dickman, 2014; Apfelbach, Parsons, Soini, & Novotny, 2015). Despite this, the ability of prey species to detect and avoid predators may be more dependent on the

124

evolutionary and ontogenetic history of the prey species (McEvoy, Sinn, & Wapstra,

2008) rather than shared chemical cues, which appears to be the case for burrowing and brush-tailed bettongs.

It is possible that burrowing bettongs did not differentiate between odours, due to the ‘aging’ of the odour signals. In arid areas of Australia, odour signals may ‘age’ faster. Aging of odours leads to decreased concentrations or loss of some of the components and/or changing ratios between the compounds (Apfelbach, Parsons,

Soini, & Novotny, 2015). Since the arid zone is a hot and dry environment, compounds evaporate more quickly due to increased vapour pressures at elevated environmental temperatures and low humidity (Apfelbach, Parsons, Soini, & Novotny, 2015). If odour cannot be readily detected in a system, prey will be unlikely to use it as a cue for predation risk (Spencer, Crowther, & Dickman, 2014). If ‘aging’ of chemical signals influences burrowing bettong’s response to the treatments, we would have expected a similar non-response by brush-tailed bettongs because the brush-tailed bettongs we studied inhabited a semi-arid environment, that experiences similar humidity levels and temperature extremes to that at the more xeric Arid Recovery site (Bureau of

Meteorology, 2015; Bureau of Meteorology, 2016). In order to test whether ‘aged’ signals played an influence on the results, future studies may have to account for how long the odour treatments were out in the environment before visitation by prey animals.

A further caveat to our study is that we had a relatively small number of videos available to assess burrowing bettong predator odour recognition behaviour (cat n = 8, dingo n = 17, rabbit n = 14 and a procedural control n = 8) compared to brush-tailed

125

bettongs (cat n = 21, dingo n = 27, rabbit n = 31 and a procedural control n = 22). This discrepancy in the number of videos available is related to the number of sites that we could sample at each location, which reflects the relative size of the study areas (the

Arid Recovery paddock we studied bettongs in – 22 km2 vs. Scotia Wildlife Sanctuary –

40 km2). It is important to note too that because our studies were conducted on different species at different sites, we did not statistically compare their responses, but instead examined their responses to the treatments within each location. Our results suggest that the two species of bettong, the burrowing bettong and brush-tailed bettong did respond differently to the odour treatments, with their responses consistent with the idea that prey responses to predator cues scale with their duration of co-evolution (Anson & Dickman, 2013; Steindler, Blumstein, West, Moseby, &

Letnic, 2018). Due to differences in sample design and size, caution is warrented when interpretting these results. In order to truly test for differences in the influence of evolutionary experience on burrowing and brush-tailed bettong predator recognition abilities, sample design and size should be replicated across the study sites.

Although the results from this study suggest that brush-tailed bettongs have retained hard–wired predator recognition abilities, developed through their shared evolutionary history with dingos and cats, it does not mean that this species will have a superior advantage when faced with these predators outside predator proof fences.

There have been several attempts to reintroduce both burrowing bettongs

(Christensen & Burrows, 1995; Bannister, Lynch, & Moseby, 2016) and brush-tailed bettongs (Wheeler & Priddel, 2009) into areas where they previously existed. Despite efforts to reduce predator densities (cat, dingo and fox), prior to release, all releases ended in eventual failure. Cat predation was identified as the leading cause for

126

burrowing and brush-tailed bettong mortalities in these cases (Christensen & Burrows,

1995; Wheeler & Priddel, 2009). It has been suggested that the proper performance of antipredator behaviour may require both heritable predisposition (Riechert & Hedrick,

1990), as well as experience (Magurran, 1990). It is feasible that bettongs failed to survive outside predator proof fences as they lacked the appropriate anti-predator behaviours that can only be acquired through lifetime experience and natural selection for appropriate antipredator traits (Moseby, Blumstein, & Letnic, 2016). Future reintroduction attempts of bettongs may have to take into consideration the influence of evolutionary history on predator recognition skills, as well as the importance of learned antipredator responses that can only be acquired through ontogenetic exposure to predators (West, Letnic, Blumstein, & Moseby 2018).

Our results suggest that burrowing bettongs, which have no ontogenetic and no evolutionary exposure to predators, are unable to distinguish between predatory, herbivore and control odours (no odour). In contrast, brush-tailed bettongs with no ontogenetic exposure to predators, but between 150 – 3,000 years of evolutionary history with predators may have retained what appear to be hard-wired predator discrimination abilities. From a reintroduction perspective this lends support to the notion that anti-predator recognition in ontogenetically predator naïve prey can be maintained despite isolation from all mammalian predators. The next phase to ensuring successful reintroductions of these species outside predator proof fences is to harness their innate antipredator responses and couple them with real life scenarios, where there are consequences for failed predator recognition (Moseby, Blumstein, &

Letnic, 2016). The results from this study also support the idea proposed by Anson and

Dickman (2013), that Australian native mammals will not be eternally naïve and that

127

less than 200 years of evolutionary history can be sufficient to develop and retain antipredator behaviours.

128

3.6 ACKNOWLEDGEMENTS

We thank private pet owners, Kindifarm, Symbio Wildlife Park, Taronga Western Plains

Zoo and animals housed at the Cat Protection Society Sydney for supply of odours and the Arid Recovery staff for their assistance with the study.

3.7 REFERENCES

Abbott, I. 2002. Origin and spread of the cat, Felis catus, on mainland Australia, with a

discussion of the magnitude of its early impact on native fauna. Wildlife

Research 29(1):51-74. doi: 10.1071/WR01011

Anson, J. R., and C. R. Dickman. 2013. Behavioral responses of native prey to disparate

predators: naiveté and predator recognition. Oecologia 171(2):367-377. doi:

10.1007/s00442-012-2424-7

Apfelbach, R., M. H. Parsons, H. A. Soini, and M. V. Novotny. 2015. Are single odorous

components of a predator sufficient to elicit defensive behaviors in prey

species? Frontiers in Neuroscience 9(263):1-14. doi: 10.3389/fnins.2015.00263

Atkins, R., D. T. Blumstein, K. E. Moseby, R. West, M. Hyatt, and M. Letnic. 2016. Deep

evolutionary experience explains mammalian responses to predators.

Behavioral Ecology and Sociobiology 70(10):1-9. doi: 10.1007/s00265-016-

2181-4

Bannister, H. L., C. E. Lynch, and K. E. Moseby. 2016. Predator swamping and

supplementary feeding do not improve reintroduction success for a threatened

Australian mammal, Bettongia lesueur. Australian Mammalogy 38(2):177-187.

doi: 10.1071/AM15020

129

Blumstein, D. T. 2002. Moving to suburbia: ontogenetic and evolutionary

consequences of life on predator-free islands. Journal of Biogeography 29(5-

6):685-692. doi: 10.1046/j.1365-2699.2002.00717.x

Blumstein, D. T. 2006. The multipredator hypothesis and the evolutionary persistence

of antipredator behavior. Ethology 112(3):209-217. doi: 10.1111/j.1439-

0310.2006.01209.x

Blumstein, D. T., and J. C. Daniel. 2005. The loss of anti-predator behaviour following

isolation on islands. Proceedings of the Royal Society B: Biological Sciences

272(1573):1663-1668. doi: 10.1098/rspb.2005.3147

Blumstein, D. T., and J. C. Daniel. 2007. Quantifying Behaviour the JWatcher Way.

Sinauer Associates, Inc., Sunderland, Massachusetts, U.S.A.

Blumstein, D. T., J. C. Daniel, A. S. Griffin, and C. S. Evans. 2000. Insular tammar

wallabies (Macropus eugenii) respond to visual but not acoustic cues from

predators. Behavioral Ecology 11(5):528-535. doi: 10.1093/beheco/11.5.528

Bradshaw, J. W. S. 2006. The evolutionary basis for the feeding behavior of domestic

dogs (Canis familiaris) and cats (Felis catus). The Journal of Nutrition

136(7):1927S-1931S. doi: 10.1093/jn/136.7.1927S

Burbidge, A. A., and N. McKenzie. 1989. Patterns in the modern decline of Western

Australia's vertebrate fauna: causes and conservation implications. Biological

Conservation 50(1):143-198. doi: 10.1016/0006-3207(89)90009-8

Bureau of Meteorology, A. G. 2015. Mean rainfall (mm) for Wentworth (Scotia

Sanctuary), New South Wales, years 1997 to 2015.

http://www.bom.gov.au/climate/data.

130

Bureau of Meteorology, A. G. 2016. Mean rainfall (mm) for Roxby Downs, South

Australia, years 1997 to 2016 http://www.bom.gov.au/climate/data.

Bytheway, J., A. R. Carthey, and P. Banks. 2013. Risk vs. reward: how predators and

prey respond to aging olfactory cues. Behavioral Ecology and Sociobiology

67(5):715-725. doi: 10.1007/s00265-013-1494-9

Carthey, A. J. 2013. Naiveté, novelty and native status: Mismatched ecological

interactions in the Australian environment. Ph.D. Dissertation, University of

Sydney, Australia.

Carthey, A. J., and P. B. Banks. 2014. Naïveté in novel ecological interactions: lessons

from theory and experimental evidence. Biological Reviews 89(4):932-949. doi:

10.1111/brv.12087

Carthey, A.J., Banks, P.B., 2016. Naiveté is not forever: responses of a vulnerable native

rodent to its long term alien predators. Oikos 125(7):918-926. doi:

10.1111/oik.02723

Carthey, A. J., and D. T. Blumstein. 2018. Predicting predator recognition in a changing

world. Trends in Ecology & Evolution 33(2):106-115. doi:

10.1016/j.tree.2017.10.009

Christensen, P., and N. Burrows. 1995. Project desert dreaming: experimental

reintroduction of mammals to the Gibson Desert, Western Australia. In: M.

Serena, editors, Reintroduction Biology of Australian and New Zealand Fauna.

Surrey Beatty & Sons, Chipping Norton. p. 199-207.

Corbett, L. K. 1995. Dingo in Australia and Asia. University of New South Wales Press,

Sydney.

131

Coss, R. G. 1999. Effects of relaxed natural selection on the evolution of behavior. In: S.

A. Foster and J. A. Endler, editors, Geographic variation in behavior:

Perspectives on evolutionary mechanisms. Oxford University Press, Inc. , New

York. p. 180-208.

Cox, J. G., and S. L. Lima. 2006. Naiveté and an aquatic–terrestrial dichotomy in the

effects of introduced predators. Trends in Ecology & Evolution 21(12):674-680.

doi: 10.1016/j.tree.2006.07.011

Curio, E. 1966. How finches react to predators. Animals 9:142-143.

DEC. 2007. Records held in Department of Environment and Conservation’s

Threatened Fauna Database and rare fauna files. In: P. Department of

Environment and Conservation (ed.).

Deecke, V. B., P. J. Slater, and J. K. Ford. 2002. Selective habituation shapes acoustic

predator recognition in harbour seals. Nature 420(6912):171-173. doi:

10.1038/nature01030

Delroy, L., Earl, J., Radbone, I., Robinson, A., Hewett, M., 1986. The Breeding and

Reestablishment of the Brush-Tailed Bettong, Bettongia penicillata, in South-

Australia. Wildlife Research 13(3): 387-396. doi: 10.1071/WR9860387

Ferrari, M. C., and D. P. Chivers. 2009. Temporal variability, threat sensitivity and

conflicting information about the nature of risk: understanding the dynamics of

tadpole antipredator behaviour. Animal Behaviour 78(1):11-16. doi:

10.1016/j.anbehav.2009.03.016

Ferrari, M. C., A. L. Crane, and D. P. Chivers. 2016. Certainty and the cognitive ecology

of generalization of predator recognition. Animal Behaviour 111:207-211. doi:

10.1016/j.anbehav.2015.10.026

132

Ferrari, M. C., A. Gonzalo, F. Messier, and D. P. Chivers. 2007. Generalization of learned

predator recognition: an experimental test and framework for future studies.

Proceedings of the Royal Society of London B: Biological Sciences

274(1620):1853-1859. doi: 10.1098/rspb.2007.0297

Ferrero, D. M., J. K. Lemon, D. Fluegge, S. L. Pashkovski, W. J. Korzan, S. R. Datta, M.

Spehr, M. Fendt, and S. D. Liberles. 2011. Detection and avoidance of a

carnivore odor by prey. Proceedings of the National Academy of Sciences

108(27):11235-11240. doi: 10.1073/pnas.1103317108

Fisher, R. A. 1935. The design of experiments. Oliver and Boyd, Edinburgh.

Fortin, M. J., and M. R. Dale. 2005. Spatial analysis: a guide for ecologists. Cambridge

University Press, Cambridge.

Goldthwaite, R. O., R. G. Coss, and D. H. Owings. 1990. Evolutionary dissipation of an

antisnake system: differential behavior by California and arctic ground squirrels

in above- and below-ground contexts. Behaviour 112(3/4):246-269. doi:

10.1163/156853990X00220

Griffin, A. S., D. T. Blumstein, and C. S. Evans. 2000. Training captive‐bred or

translocated animals to avoid predators. Conservation Biology 14(5):1317-

1326. doi: 10.1046/j.1523-1739.2000.99326.x

Griffin, A. S., C. S. Evans, and D. T. Blumstein. 2001. Learning specificity in acquired

predator recognition. Animal Behaviour 62(3):577-589. doi:

10.1006/anbe.2001.1781

Hancock, S., P. Brown, and B. Stephens. 2000. Shark Bay Terrestrial Reserves

Management Plan 2000–2009. Western Australia, Department of Conservation

133

and Land Management for the National Parks and Nature Conservation

Authority

Helfman, G. 1989. Threat-sensitive predator avoidance in damselfish-trumpetfish

interactions. Behav Ecol Sociobiol 24(1):47-58. doi: 10.1007/BF00300117

Horton, N.J. and S. R. Lipsitz. 1999. Review of software to fit generalized estimating

equation regression models. The American Statistician 53(2):160–169. doi:

10.2307/2685737

Jedrzejewski, W., L. Rychlik, and B. Jedrzejwska. 1993. Responses of bank voles to

odours of seven species of predators: experimental data and their relevance to

natural predator-vole relationships. Oikos 68(2):251-257. doi: 10.2307/3544837

Johnson, C. 2006. Australia's Mammal Extinctions: A 50,000 year history. Cambridge

University Press, Port Melbourne. p. 1-278.

Jolly, C. J., J. K. Webb, and B. L. Phillips. 2018. The perils of paradise: an endangered

species conserved on an island loses antipredator behaviours within 13

generations. Biology Letters 14(6)doi: 10.1098/rsbl.2018.0222

Kats, L. B., and L. M. Dill. 1998. The scent of death: Chemosensory assessment of

predation risk by prey animals. Écoscience 5(3):361-394. doi:

10.1080/11956860.1998.11682468

Letnic, M., M. Fillios, and M. S. Crowther. 2014. The arrival and impacts of the dingo.

In: A. Glen and C. R. Dickman, editors, Carnivores of Australia: Past, Present and

Future. CSIRO Publishing p. 55 - 69.

Magurran, A. E. 1990. The inheritance and development of minnow anti-predator

behaviour. Animal Behaviour 39(5):834-842. doi: 10.1016/S0003-

3472(05)80947-9

134

Marlow, N. J., N. D. Thomas, A. A. Williams, B. Macmahon, J. Lawson, Y. Hitchen, J.

Angus, and O. Berry. 2015. Cats (Felis catus) are more abundant and are the

dominant predator of woylies (Bettongia penicillata) after sustained fox (Vulpes

vulpes) control. Australian Journal of Zoology 63(1):18-27. doi:

10.1071/ZO14024

McCormick, M., and T. Holmes. 2006. Prey experience of predation influences

mortality rates at settlement in a coral reef fish, Pomacentrus amboinensis.

Journal of Fish Biology 68(3):969-974. doi: 10.1111/j.0022-1112.2006.00982.x

McEvoy, J., D. L. Sinn, and E. Wapstra. 2008. Know thy enemy: Behavioural response of

a native mammal (Rattus lutreolus velutinus) to predators of different

coexistence histories. Austral Ecology 33(7):922-931. doi: 10.1111/j.1442-

9993.2008.01863.x

Mella, V. S. A., C. E. Cooper, and S. J. J. F. Davies. 2010. Predator odour does not

influence trappability of southern brown bandicoots (Isoodon obesulus) and

common brushtail possums (Trichosurus vulpecula). Australian Journal of

Zoology 58(5):267-272. doi: 10.1071/ZO10049

Monclús, R., H. G. Rödel, D. Von Holst, and J. De Miguel. 2005. Behavioural and

physiological responses of naïve European rabbits to predator odour. Animal

Behaviour 70(4):753-761. doi: 10.1016/j.anbehav.2004.12.019

Moseby, K., and J. Read. 2006. The efficacy of feral cat, fox and rabbit exclusion fence

designs for threatened species protection. Biological Conservation 127(4):429-

437. doi: 10.1016/j.biocon.2005.09.002

135

Moseby, K. E., D. T. Blumstein, and M. Letnic. 2016. Harnessing natural selection to

tackle the problem of prey naïveté. Evolutionary Applications 9(2):334-343. doi:

10.1111/eva.12332

Moseby, K. E., J. L. Read, D. C. Paton, P. Copley, B. M. Hill, and H. A. Crisp. 2011.

Predation determines the outcome of 10 reintroduction attempts in arid South

Australia. Biological Conservation 144(12):2863-2872. doi:

10.1016/j.biocon.2011.08.003

Nakagawa, S. (2004). A farewell to Bonferroni: the problem of low statistical power

and publication bias. Behavioral Ecology, 15(6), 1044-1045.

https://doi.org/10.1093/beheco/arh107

Omar, R. Z., Wright, E. M., Turner, R. M., Thompson, S. G., 1999. Analysing repeated

measurements data: a practical comparison of methods. Statistics in Medicine

18(13): 1587-1603. doi: 10.1002/(SICI)1097-0258(19990715)18:13<1587::AID-

SIM141>3.0.CO;2-Z

Orrock, J. L. 2010. When the ghost of predation has passed: do rodents from islands

with and without fox predators exhibit aversion to fox cues? Ethology

116(4):338-345. doi: 10.1111/j.1439-0310.2010.01740.x

Page, M. J. 2008. Karakamia Wildlife Sanctuary Annual Report 2008

https://library.dbca.wa.gov.au/static/Journals/080197/080197_2009.pdf.

Parsons, M. H., R. Apfelbach, P. B. Banks, E. Z. Cameron, C. R. Dickman, A. S. K. Frank,

M. E. Jones, I. S. McGregor, S. McLean, D. Müller‐Schwarze, E. E. Sparrow, and

D. T. Blumstein. 2018. Biologically meaningful scents: a framework for

understanding predator–prey research across disciplines. Biological Reviews

93(1):98-114. doi: 10.1111/brv.12334

136

Parsons, M. H., and D. T. Blumstein. 2010. Familiarity Breeds Contempt: Kangaroos

Persistently Avoid Areas with Experimentally Deployed Dingo Scents. PLoS ONE

5(5):e10403. doi: 10.1371/journal.pone.0010403

Pekár, S. and M. Brabec. 2017. Generalized estimating equations: A pragmatic and

flexible approach to the marginal GLM modelling of correlated data in the

behavioural sciences. Ethology 124(2):86-93. doi.org/10.1111/eth.12713

Powell, F., and P. B. Banks. 2004. Do house mice modify their foraging behaviour in

response to predator odours and habitat? Animal Behaviour 67(4):753-759.

doi: 10.1016/j.anbehav.2003.08.016

Quinn, G.P., Keough, M.J., 2002. Experimental design and data analysis for biologists.

Cambridge University Press.

Richards, J., Morris, K. and A. Burbidge. 2008. Bettongia lesueur. The IUCN Red List of

Threatened Species 2008. doi:

10.2305/IUCN.UK.2008.RLTS.T2784A9480530.en.

Ride, W. D. L. and A. J. Fraser. 1962. The Results of an expedition to Bernier and Dorre

Islands, Shark Bay, Western Australia, in July 1959. Perth, Western

Australia: Chief Warden of Fauna p. 64-69 & 84.

Riechert, S. E., and A. V. Hedrick. 1990. Levels of predation and genetically based anti-

predator behaviour in the spider, Agelenopsis aperta. Animal Behaviour

40(4):679-687. doi: 10.1016/S0003-3472(05)80697-9

Russell, B.G., and P. B. Banks. 2007. Do Australian small mammals respond to native

and introduced predator odours? Austral Ecology 32:277-286.

doi:10.1111/j.1442-9993.2007.01685.x

137

Saxon-Mills, E. C., K. Moseby, D. T. Blumstein, and M. Letnic. 2018. Prey naïveté and

the anti-predator responses of a vulnerable marsupial prey to known and novel

predators. Behavioral Ecology and Sociobiology 72(9):151. doi:

10.1007/s00265-018-2568-5

Short, J. 1998. The extinction of rat-kangaroos (Marsupialia:Potoroidae) in New South

Wales, Australia. Biological Conservation 86(3):365-377. doi: 10.1016/S0006-

3207(98)00026-3

Short, J., L. Atkins, and B. Turner. 2005. Diagnosis of mammal decline in Western

Australia, with particular emphasis on the possible role of feral cats and poison

peas. Report to National Geographic Society. In: Perth: CSIRO Sustainable

Ecosystems p. 2-43.

Short, J., and B. Turner. 1993. The distribution and abundance of the burrowing

bettong (Marsupialia : Macropoidea). Wildlife Research 20(4):525-533. doi:

10.1071/WR9930525

Short, J., Turner, B., 1999. Ecology of burrowing bettongs, Bettongia lesueur

(Marsupialia: Potoroidae), on Dorre and Bernier Islands, Western Australia.

Wildlife Research 26(5), 651-669. doi: 10.1071/WR98039

Short, J., and B. Turner. 2000. Reintroduction of the burrowing bettong Bettongia

lesueur (Marsupialia: Potoroidae) to mainland Australia. Biological

Conservation 96(2):185-196. doi: 10.1016/S0006-3207(00)00067-7

Shortridge, G. 1910. An Account of the Geographical Distribution of the Marsupials and

Monotremes of South‐West Australia, having special reference to the

specimens collected during the Balston Expedition of 1904–1907. In:

Proceedings of the Zoological Society of London. p 803-848.

138

Spencer, E. E., M. S. Crowther, and C. R. Dickman. 2014. Risky business: do native

rodents use habitat and odor cues to manage predation risk in Australian

deserts? PLoS One 9(2):e90566. doi: 10.1371/journal.pone.0090566

Steindler, L. A., D. T. Blumstein, R. West, K. E. Moseby, and M. Letnic. 2018.

Discrimination of introduced predators by ontogenetically naïve prey scales

with duration of shared evolutionary history. Animal Behaviour 137:133-139.

doi: 10.1016/j.anbehav.2018.01.013

Thompson, C.K., Wayne, A.F., Godfrey, S.S., Thompson, R.C.A., 2015. Survival, age

estimation and sexual maturity of pouch young of the brush-tailed bettong

(Bettongia penicillata) in captivity. Australian Mammalogy 37(1), 29-38. doi:

10.1071/AM14025

Webb, J. K., W. G. Du, D. A. Pike, and R. Shine. 2009. Chemical cues from both

dangerous and nondangerous snakes elicit antipredator behaviours from a

nocturnal lizard. Animal Behaviour 77(6):1471-1478. doi:

10.1016/j.anbehav.2009.02.020

West, R., M. Letnic, D. T. Blumstein, and K. E. Moseby. 2018. Predator exposure

improves anti‐predator responses in a threatened mammal. Journal of Applied

Ecology 55(1): 147-156. doi: 10.1111/1365-2664.12947

Wheeler, R., and D. Priddel. 2009. The impact of introduced predators on two

threatened prey species: A case study from western New South Wales.

Ecological Management & Restoration 10:S117-S123. doi: 10.1111/j.1442-

8903.2009.00457.x

Woinarski, J. C., A. A. Burbidge, and P. L. Harrison. 2015. Ongoing unraveling of a

continental fauna: decline and extinction of Australian mammals since

139

European settlement. Proceedings of the National Academy of Sciences USA

112(15):4531-4540. doi: 10.1073/pnas.1417301112

Yeatman, G., and C. Groom. 2012. National Recovery Plan for the woylie Bettongia

penicillata. Wildlife Management Program (51)

Yeatman, G. J., and A. F. Wayne. 2015. Seasonal home range and habitat use of a

critically endangered marsupial (Bettongia penicillata ogilbyi) inside and

outside a predator-proof sanctuary. Australian Mammalogy 37(2):157-163. doi:

https://doi.org/10.1071/AM14022

Zenger, K. R., B. J. Richardson, and A. M. Vachot-Griffin. 2003. A rapid population

expansion retains genetic diversity within European rabbits in Australia.

Molecular Ecology 12(3):789-794. doi: 10.1046/j.1365-294X.2003.01759.x

140

3.8 SUPPLEMENTARY MATERIAL

Species of bettong Behaviour Moran’s (i) z p-value

Burrowing bettong Wary Approach 0.477 1.403 0.161

Forage 0.543 1.622 0.105

Retreat 0.371 1.213 0.225

Fast Approach 0.287 1.116 0.264

Investigate 0.376 1.259 0.208

Brush-tailed bettong Wary Approach 0.075 0.489 0.625

Forage 0.268 1.587 0.113

Retreat 0.050 0.247 0.805

Fast Approach 0.151 0.853 0.393

Investigate 0.158 0.886 0.376

141

Table S3.2 Brush-tailed bettongs: Results of Fisher’s Least Significant Difference (LSD) post-hoc comparisons testing for differences between odour treatments: cat, dingo, rabbit, and control – no odour, in the mean proportion of time spent on each behaviour by brush-tailed bettongs.

Behaviour groups Mean difference SEM p-value

Wary approach

Dingo vs. cat 0.1585 0.08007 0.048**

Dingo vs. rabbit 0.0479 0.07393 0.517

Dingo vs. control 0.0533 0.07684 0.488

Cat vs. rabbit 0.2064 0.08324 0.013**

Cat vs. control 0.2118 0.08583 0.014**

Rabbit vs. control 0.0054 0.08013 0.946

Retreat

Dingo vs. cat 0.0455 0.02845 0.109

Dingo vs. rabbit 0.0257 0.02899 0.375

a Dingo vs. control 0.0620 0.02786 0.026**

Cat vs. rabbit 0.0198 0.01600 0.216

Cat vs. control 0.0165 0.01385 0.233

Rabbit vs. control 0.0363 0.01493 0.015**

Values in italic** indicate significant differences

142

Table S3.3 Repeated measures results from generalized estimating equations model

testing for differences between odour treatments (cat, dingo, rabbit and control – no

odour) on the mean proportion of time spent on each behaviour by burrowing

bettongs (Bettongia lesueur) and brush-tailed bettong (Bettongia penicillata).

Species Behaviour N Min Max Mean Std. Dev

Burrowing bettongs

(Bettongia lesueur)

Dependent Variable Wary Approach 47 ≤0.005 0.999 0.421 0.318

Covariate Year 47 2015 2016 2015.57 0.500

ID 47 2 64 31.38 19.390

Dependent Variable Forage 47 ≤0.005 0.999 0.239 0.322

Covariate Year 47 2015 2016 2015.57 0.500

ID 47 2 64 31.38 19.390

Dependent Variable Retreat 47 ≤0.005 0.999 0.054 0.173

Covariate Year 47 2015 2016 2015.57 0.500

ID 47 2 64 31.38 19.390

Dependent Variable Fast Approach 47 ≤0.005 0.870 0.066 0.177

Covariate Year 47 2015 2016 2015.57 0.500

ID 47 2 64 31.38 19.390

Dependent Variable Investigate 47 ≤0.005 0.880 0.164 0.248

143

Covariate Year 47 2015 2016 2015.57 0.500

ID 47 2 64 31.38 19.390

Brush-tailed bettong

(Bettongia penicillata)

Dependent Variable Wary Approach 101 ≤0.005 0.999 0.434 0.289

Covariate ID 101 1 85 42.64 26.278

Dependent Variable Forage 101 ≤0.005 0.999 0.380 0.361

Covariate ID 101 1 85 42.64 26.278

Dependent Variable Retreat 101 ≤0.005 0.610 0.066 0.093

Covariate ID 101 1 85 42.64 26.278

Dependent Variable Fast Approach 101 ≤0.005 0.850 0.049 0.139

Covariate ID 101 1 85 42.64 26.278

Dependent Variable Investigate 101 ≤0.005 0.440 0.039 0.094

Covariate ID 101 1 85 42.64 26.278

144

Table S3.4 List of odour treatment (cat, dingo, rabbit and control – no odour) deployed across sites (ID) per year (2015 and 2016) for burrowing bettongs (Bettongia lesueur).

Year ID Treatment Year ID Treatment Year ID Treatment 2015 1 Cat 2016 14 Rabbit 2015 28 Rabbit 2016 1 Dingo 2015 15 Dingo 2016 28 Dingo 2015 2 Control 2016 15 Dingo 2015 29 Dingo 2016 2 Rabbit 2015 16 Rabbit 2016 29 Cat 2015 3 Dingo 2016 16 Control 2015 30 Control 2016 3 Cat 2015 17 Rabbit 2016 30 Dingo 2015 4 Cat 2016 17 Cat 2015 31 Cat 2016 4 Rabbit 2015 18 Control 2016 31 Dingo 2015 5 Control 2016 18 Control 2015 32 Dingo 2016 5 Control 2015 19 Dingo 2016 32 Rabbit 2015 6 Rabbit 2016 19 Rabbit 2015 33 Control 2016 6 Control 2015 20 Dingo 2016 33 Control 2015 7 Dingo 2016 20 Control 2015 34 Dingo 2016 7 Rabbit 2015 21 Control 2016 34 Rabbit 2015 8 Rabbit 2016 21 Dingo 2015 35 Control 2016 8 Cat 2015 22 Cat 2016 35 Rabbit 2015 9 Cat 2016 22 Control 2015 36 Cat 2016 9 Control 2015 23 Control 2016 36 Cat 2015 10 Rabbit 2016 23 Control 2015 37 Rabbit 2016 10 Dingo 2015 24 Rabbit 2016 37 Cat 2015 11 Dingo 2016 24 Rabbit 2015 38 Control 2016 11 Dingo 2015 25 Cat 2016 38 Dingo 2015 12 Cat 2016 25 Rabbit 2015 39 Dingo 2016 12 Cat 2015 26 Cat 2016 39 Rabbit 2015 13 Control 2016 26 Cat 2015 40 Rabbit 2016 13 Cat 2015 27 Rabbit 2016 40 Rabbit 2015 14 Control 2016 27 Cat 2015 41 Dingo

145

Year ID Treatment Year ID Treatment Year ID Treatment 2016 41 Control 2015 52 Rabbit 2016 62 Cat 2015 42 Rabbit 2016 52 Dingo 2015 63 Control 2016 42 Control 2015 53 Cat 2016 63 Cat 2015 43 Cat 2016 53 Dingo 2015 64 Dingo 2016 43 Dingo 2015 54 Dingo 2016 64 Cat 2015 44 Dingo 2016 54 Control 2016 44 Control 2015 55 Rabbit 2015 45 Cat 2016 55 Rabbit 2016 45 Dingo 2015 56 Control 2015 46 Control 2016 56 Control 2016 46 Dingo 2015 57 Control 2015 47 Rabbit 2016 57 Rabbit 2016 47 Cat 2015 58 Control 2015 48 Cat 2016 58 Dingo 2016 48 Cat 2015 59 Dingo 2015 49 Cat 2016 59 Rabbit 2016 49 Rabbit 2015 60 Rabbit 2015 50 Rabbit 2016 60 Cat 2016 50 Control 2015 61 Cat 2015 51 Cat 2016 61 Dingo 2016 51 Control 2015 62 Dingo

146

Table S3.5 List of odour treatment (cat, dingo, rabbit and control – no odour) deployed across sites (ID) per night (2 and 3) for brush-tailed bettong (Bettongia penicillata).

ID Treatment Night ID Treatment Night ID Treatment Night

1 Dingo 2 14 Rabbit 3 28 Control 2 1 Dingo 3 15 Cat 2 28 Control 3 2 Control 2 15 Cat 3 29 Dingo 2 2 Control 3 16 Rabbit 2 29 Dingo 3 3 Cat 2 16 Rabbit 3 30 Dingo 2 3 Cat 3 17 Rabbit 2 30 Dingo 3 4 Rabbit 2 17 Rabbit 3 31 Cat 2 4 Rabbit 3 18 Rabbit 2 31 Cat 3 5 Cat 2 18 Rabbit 3 32 Cat 2 5 Cat 3 19 Dingo 2 32 Cat 3 6 Rabbit 2 19 Dingo 3 33 Rabbit 2 6 Rabbit 3 20 Dingo 2 33 Rabbit 3 7 Control 2 20 Dingo 3 34 Cat 2 7 Control 3 21 Control 2 34 Cat 3 8 Dingo 2 21 Control 3 35 Dingo 2 8 Dingo 3 22 Cat 2 35 Dingo 3 9 Rabbit 2 22 Cat 3 36 Cat 2 9 Rabbit 3 23 Control 2 36 Cat 3 10 Control 2 23 Control 3 37 Dingo 2 10 Control 3 24 Control 2 37 Dingo 3 11 Rabbit 2 24 Control 3 38 Control 2 11 Rabbit 3 25 Rabbit 2 38 Control 3 12 Dingo 2 25 Rabbit 3 39 Cat 2 12 Dingo 3 26 Dingo 2 39 Cat 3 13 Control 2 26 Dingo 3 40 Control 2 13 Control 3 27 Cat 2 40 Control 3 14 Rabbit 2 27 Cat 3 41 Dingo 2

147

ID Treatment Night ID Treatment Night ID Treatment Night 41 Dingo 3 57 Control 2 72 Rabbit 2 42 Control 2 57 Control 3 72 Rabbit 3 42 Control 3 58 Rabbit 2 73 Dingo 2 43 Dingo 2 58 Rabbit 3 73 Dingo 3 43 Dingo 3 59 Cat 2 74 Rabbit 2 44 Control 2 59 Cat 3 74 Rabbit 3 44 Control 3 60 Dingo 2 75 Control 2 45 Rabbit 2 60 Dingo 3 75 Control 3 45 Rabbit 3 61 Rabbit 2 76 Control 2 46 Dingo 2 61 Rabbit 3 76 Control 3 46 Dingo 3 62 Rabbit 2 77 Rabbit 2 47 Dingo 2 62 Rabbit 3 77 Rabbit 3 47 Dingo 3 63 Cat 2 78 Control 2 48 Control 2 63 Cat 3 78 Control 3 48 Control 3 64 Dingo 2 79 Cat 2 49 Rabbit 2 64 Dingo 3 79 Cat 3 49 Rabbit 3 65 Cat 2 80 Cat 2 50 Cat 2 65 Cat 3 80 Cat 3 51 Rabbit 2 66 Control 2 81 Cat 2 51 Rabbit 3 66 Control 3 81 Cat 3 52 Control 2 67 Cat 2 82 Dingo 2 52 Control 3 67 Cat 3 82 Dingo 3 53 Cat 2 68 Cat 2 83 Rabbit 2 53 Cat 3 68 Cat 3 83 Rabbit 3 54 Control 2 69 Cat 2 84 Cat 2 54 Control 3 69 Cat 3 84 Cat 3 55 Rabbit 2 70 Dingo 2 85 Dingo 2 55 Rabbit 3 70 Dingo 3 85 Dingo 3 56 Dingo 2 71 Dingo 2 86 Rabbit 2 56 Dingo 3 71 Dingo 3 86 Rabbit 3

148

ID Treatment Night ID Treatment Night ID Treatment Night 87 Rabbit 2 96 Control 2 105 Dingo 2 87 Rabbit 3 96 Control 3 105 Dingo 3 88 Cat 2 97 Rabbit 2 106 Rabbit 2 88 Cat 3 97 Rabbit 3 106 Rabbit 3 89 Cat 2 98 Control 2 107 Cat 2 89 Cat 3 98 Control 3 107 Cat 3 90 Control 2 99 Cat 2 108 Rabbit 2 90 Control 3 99 Cat 3 108 Rabbit 3 91 Control 2 100 Control 2 109 Rabbit 2 91 Control 3 100 Control 3 109 Rabbit 3 92 Dingo 2 101 Dingo 2 110 Cat 2 92 Dingo 3 101 Dingo 3 110 Cat 3 93 Control 2 102 Control 2 111 Dingo 2 93 Control 3 102 Control 3 111 Dingo 3 94 Dingo 2 103 Control 2 112 Dingo 2 94 Dingo 3 103 Control 3 112 Dingo 3 95 Rabbit 2 104 Cat 2 95 Rabbit 3 104 Cat 3

149

Table S3.6 Ethogram of burrowing bettong (Bettongia lesueur) and brush-tailed bettong (Bettongia penicillata) behaviour.

Behaviour Description a

Bi-pedal sniff Animal looks and/or sniffs air whilst standing upright on hind limbs

Prone sniff Animal looks and/or sniffs the air and/or ground whilst standing on all four limbs

Slow approach Animal moves slowly towards foraging site

Fast approach Animal moves quickly and directly towards foraging site

Grooming Animals grooms itself using grooming claw on hind limb and/or mouth

Vigilant foraging Animal chews with its head up and observing surroundings

Relaxed foraging Animal forages for food and chews with its head down without observing surroundings

Investigate towel Olfactory investigation of towel

Chew towel Animal attempts to chew and/or remove towel away from foraging site

Recoil Animal recoils away from the foraging site and/or visual treatment

Retreat Animal retreats away from foraging site

Fighting Animal fights with con-specifics

Scent marking Animal has its rear dropped to the ground and tail lifted and appears to drag cloaca along the ground scent marking

Out of sight Animal moves out of sight from within camera frame

Other Animal displays other behaviour not listed within ethogram

a Definition of postures associated with particular behaviour

150

CHAPTER 4:

Exposure to a novel predator induces visual predator

recognition by naïve prey

151

4.1 ABSTRACT

The ‘life-dinner principle’ posits that there is greater selection pressure on the species that have more to lose in an interaction. Thus, based on the asymmetry within predator-prey interactions there is an advantage for prey to learn quickly, especially in response to novel, introduced predators. Within ecosystems in which novel predators are newly introduced, their impacts are often most severe during the initial phase of invasion, due to a lack of prey-predator recognition and inappropriate anti-predator response behaviours. Here we test the ‘learned recognition’ hypothesis that posits that naïve prey species’ ability to recognize and respond to introduced predators can be induced through experience. We quantified the behavioural response of initially predator naïve burrowing bettongs (Bettongia lesueur) that had been living in the presence (for 8 — 15 months) and absence of an introduced predator (feral cats —

Felis catus) to models of cats, a herbivore (rabbit — Oryctolagus cuniculus), novel object (plastic bucket) and no object (control). We expected that if bettongs recognized cats as a threat they would be more wary in the presence of cat models than either rabbit models, buckets or the control. Bettongs living without predators approached all models cautiously in comparison to the control, suggesting that bettongs responded to the presence of an object and/or model, but did not discriminate between them. Bettongs living with cats spent more time cautiously approaching the cat model compared to the rabbit, bucket and control. Our results are consistent with the learned recognition hypothesis which suggests that a predator- naïve prey species ability to recognize novel predators is inducible through experience.

Our finding suggests that anti-predator recognition of reintroduced species could be

152

improved prior to release by exposing them to predators under carefully controlled conditions.

Keywords: Anti-predator behaviour, Burrowing bettong, Learned recognition hypothesis, Ontogenetic experience, Visual discrimination

153

4.2 INTRODUCTION

If a prey animal fails to escape a predation event they lose their life, whilst if they survive, the predator simply loses their meal. The ‘life-dinner principle’ posits that there is greater selection pressure on the species that have more to lose in an interaction (Dawkins and Krebs, 1979). It is assumed that this asymmetry within the predator-prey system results in an unequal selection pressure in favour of the prey.

Prey must evolve more rapidly than predators, thereby leaving predators less able to exploit them (Brodie and Brodie, 1999).

There is a large source of literature on the invasion of novel predators and their effects on predator-naïve prey. Within ecosystems in which predators are newly introduced, their impacts are often most severe during the initial phase of invasion, compared to the chronic phase, where native prey may learn to respond to novel enemies (Bytheway et al., 2016). Some prey are able to innately recognize their predators (Veen et al., 2000; Apfelbach et al., 2005; Epp and Gabor, 2008; Steindler et al., 2018), whilst others have the ability to learn how to respond to an introduced predator (Griffin et al., 2001; Epp and Gabor, 2008; Anson and Dickman, 2013). The role of learned behaviour and naïveté are often unclear and confounded (Martin,

2014) due to lack of knowledge of evolutionary and ontogenetic history of prey species. Based on the natural asymmetry within the predator prey system there is an advantage for prey to learn quickly (Dawkins and Krebs, 1979), particularly in response to novel, introduced predators.

The degree to which antipredator behaviour towards a novel introduced predator may be learnt or induced and the time frame over which this may occur is of

154

considerable theoretical interest. The ‘prey naïveté’ hypothesis suggests that animals isolated from predators over both ontogenetic and evolutionary time scales may lose antipredator behaviour that is costly and no longer relevant (van Damme and Castilla,

1996; Blumstein and Daniel, 2005). For example, a number of species isolated on islands appear to have responded to isolation from predators by reducing antipredator vigilance behaviour (Beauchamp, 2004; Blumstein and Daniel, 2005), as well as other behaviours, such as flight initiation distance (Cooper et al., 2014). However, we do not expect that novel predators will remain eternally novel.

The ‘learned recognition’ hypothesis suggests that through lifetime experience with predators, naïve prey may enhance their ability to recognize and respond to predators (Turner et al., 2006). This ability to develop learnt antipredator recognition skills towards previously evolutionary and ontogenetically unfamiliar predators has been shown in fish (Ferrari et al., 2005; Holmes and McCormick, 2010; Ferrari, 2014), birds (Maloney and McLean, 1995) and mammals (Mineka and Cook, 1988; Griffin et al., 2000; Webb et al., 2008). Although animals have been shown to learn for decades, this research aims to assess theories of on predator recognition, of which the ‘learned recognition’ hypothesis is one. We also wished to understand the time frames over which predator recognition behaviour may be learnt or induced.

An understanding of the time frames over which predator recognition behaviour may be learnt or induced is also of great practical conservation interest. The more rapidly a naïve prey can learn to recognize predators, the better their chances of survival (McCormick and Holmes, 2006). The introduction of novel predators has led to the rapid loss of ‘naïve’, endemic species within a number of areas (Engbring and

155

Fritts, 1988; Johnson, 2006). Since the introduction of predators such as the red fox

(Vulpes vulpes) and feral cat (Felis catus) (Woinarski et al., 2015), Australia has experienced the highest contemporary mammal extinction rate in the world, with 22 species driven extinct over the last two centuries (Johnson, 2006, Hayward et al.,

2014). Prey naiveté towards introduced predators has resulted in limited reintroduction and translocation success (Moseby et al., 2011; Moseby et al., 2012).

Reintroduction of naïve species that have already failed to survive in the presence of novel predators will ultimately result in failure, unless the reintroduced animals acquire survival skills not present in the original populations (McLean et al., 1996;

Moseby et al., 2016).

Anti-predator behaviour may be lost over ontogenetic (Carrete and Tella, 2015) or evolutionary time (Blumstein and Daniel, 2005), but the question still than remains: how long does it take species to display anti-predator behaviour that have had minimal evolutionary and no previous ontogenetic exposure to predators? Previous studies have shown that species are able to adapt readily to change (Berger et al., 2001;

Phillips and Shine, 2006; Webb et al., 2008). The ability to learn to respond to predators permits flexible responses to variable threats (Brown and Chivers, 2005;

Berger et al., 2010). Individuals should be at a selective advantage if they are capable of reliably assessing local predation risk and adjusting the intensity of their antipredator behaviour to match their current risk (Chivers et al., 2001; Brown and

Chivers, 2005; Ferrari et al., 2005; Ferrari, 2014). The challenge, however is to determine what level of environmental change or predation pressure will encourage learning and, ultimately, adaptation rather than extinction (Phillips and Shine, 2006).

156

Much of the literature investigating the evolved abilities of wild prey to recognize cues associated with novel predator species cannot control for variation in ontogenetic exposure to predators (Carthey and Banks, 2012; Anson and Dickman, 2013; Carthey and Banks, 2014; 2016). Here we test the idea that a prey species’ ability to visually recognize novel predators can be induced as a result of experience. We did this by quantifying the behavioural response of burrowing bettongs (Bettongia lesueur) living within fenced exclosures at Arid Recovery Reserve, South Australia, in the presence

(for 8 – 15 months) and absence of feral cats, to taxidermy models of cats (an introduced novel predator) and European rabbits (Oryctolagus cuniculus - herbivore), plastic buckets (a novel object), and no object (a procedural control). We expected that if bettongs recognized cats as a threat they would be more wary in the presence of the cat models than rabbit models, the novel control and the procedural control.

4.3 MATERIALS AND METHODS

4.3.1 Study species

Burrowing bettongs (Fig. 4.1) are a small (800 – 2000 g), nocturnal, omnivorous, burrowing macropod (Short and Turner, 1993), which live in large social groups of 20 –

40 individuals (Sander et al., 1997). It is the only species of macropod to construct and live in warrens and burrows (Short and Turner, 1993).

157

Figure 4.1 Burrowing bettong (Bettongia lesueur).

Burrowing bettongs were driven extinct on mainland Australia in the 20th century due primarily to predation by introduced red foxes and feral cats (Short and Turner,

1993; Short and Turner, 2000), as well as competition with rabbits and pastoral activities (Short & Turner, 2000). Populations of burrowing bettongs persisted on

Bernier, Dorre and Barrow Islands off the coast of Western Australia (Short and Turner,

1993; Short and Turner, 1999; 2000). Individuals from these island populations were initially reintroduced to an exclosure free of mammalian predators on the mainland, at

Heirisson Prong. Subsequently, bettongs from the island populations and Herisson

Prong were reintroduced to a number of fenced reserves free of mammalian predators on mainland Australia, including Arid Recovery in 1999/2000 (Richards, 2008; Moseby et al., 2011).

The island populations of burrowing bettongs from Bernier Island and reintroduced populations on mainland Australia have had limited contact with

158

placental mammalian predators. Rising sea levels and wave erosion meant Bernier

Island became separated from mainland Australia approximately 8,000 years ago and it is thought to have been uninhabited prior to European exploration (Hancock et al.,

2000). Feral cats were noted as being present on Bernier and Dorre Islands in the early

20th century; however, these cat populations did not persist into the second half of the

20th century (Shortridge, 1910; Ride et al., 1962). Feral cat and red fox incursions within the fenced reserve at Heirisson Prong resulted in the loss of some individuals during reintroductions in the 1990’s (Short and Turner, 2000). Thus, it is likely that the bettongs from Bernier Island and Heirisson Prong were subjected to a low level of predation pressure by feral cats within their recent evolutionary history (less than 100 years).

4.3.2 Study site

The study was conducted within two fenced areas within the Arid Recovery Reserve,

South Australia (12,300-ha, 30°29′S, 136°53′E). All mammalian predators, including cats and foxes are absent and excluded from the 22 km2 “no predator” section of the reserve (Fig. 4.2). A total of 352 burrowing bettongs were reintroduced into the 26 km2

“cat enclosure” (Fig. 4.2) between October and December 2014. At the start of the bettong release in 2014 there was one cat of unknown sex present. Five additional feral cats were added between 6 and 8 months after the initial bettong release in

2014, however three were only detected during the study period. Rabbits were absent from the “no predator” exclosure and at low densities in the “cat enclosure” (West et al., 2017).

159

“cat enclosure” one cat of unknown sex was present in 2014, with five additional cats added between 6 and 8 months after the initial release of bettongs in 2014

160

4.3.3 Experimental rationale

Because burrowing bettongs are nocturnal it is difficult to perform focal observation experiments. For this reason, we utilized video-footage from infrared motion sensor cameras placed at foraging trays to gauge the behavioural responses of bettongs living within fenced exclosures in the presence and absence of feral cats, to taxidermy models of cats, rabbits, a procedural control which consisted of a novel object (bucket) and control with no object. We conducted experiments in the “no predator” and “cat enclosure” simultaneously.

If a lack of experience with predators influenced the ability of bettongs living in the “no predator” exclosure to recognize models, we would expect that bettongs should show a similar response to models of cats (a potential predatory threat), as they would to rabbits (a harmless herbivore) and our novel object (a bucket of similar height as our cat model). We would expect that bettongs should show a greater, generalized response to all the models than to the procedural control (no object).

If predator visual recognition abilities are inducible through experience, we expected that bettongs coexisting with cats, in the “cat enclosure”, should have greater responses to the predator model than rabbits, with which they have had some ontogenetic exposure to within the enclosure and to our novel object (plastic bucket) and the control (no object).

161

4.3.4 Population level vigilance behaviour field methods

A total of 96 feeding stations, with 48 in the “no predator” exclosure and 48 in the “cat enclosure”, were established along roads with sites set approximately 400 m apart. Experiments were conducted over 16 nights (8 nights in October 2015 and 8 nights in March 2016). Burrowing bettongs were allowed to acclimate to feeding at the feeding stations for one night. After one night of acclimatization, treatments (cat, rabbit, bucket or control) were randomly allocated to stations for the second night.

Treatments were applied only once per site per year (Table 4.1). We used life size taxidermy models of cats (Fig. 4.3a) and rabbits (Fig. 4.3b), and a 9.3 L plastic bucket

(Fig. 4.3c) to quantify the response to a novel object and a blank control (Fig. 4.3d) that allowed us to measure spontaneous behavioural change in the absence of a stimulus presentation.

162

Table 4.1 The number of foraging stations (n) by treatment type deployed during the

2015 and 2016 sampling periods.

Treatment

Exclosure Year Bucket Cat Control Rabbit TOTAL GRAND TOTAL

2015 12 12 12 12 48 No Predator 96 2016 12 12 12 12 48

2015 12 12 12 12 48 Cat Exclosure 96 2016 12 12 12 12 48

Each station comprised a metal post, positioned approximately 2 m from the road, supporting either a Bushnell Trophy Cam (Bushnell; America), Scoutguard

SG550V or Scoutguard Zeroglow (Scoutguard; Australia), infrared motion sensor video camera. Cameras were mounted 50 – 150 mm off the ground and were programmed to take 60-s videos when triggered, to enable species identification and observe behavioural responses to the model types. There was a 10-s interval between possible triggers, from dusk until dawn (1700-0700 hours). A food lure, approximately 100-g dog pellets (Home brand Adult Dog Food Beef & Vegetable; Woolworths; Australia), was buried beneath the surface and mixed into the inedible substrate from the surrounding environment, approximately 5 m from the road. We swept a 10 cm circle directly surrounding the food lure to detect tracks and validate visitation. The site was checked the following day for signs of foraging. Sites were reset with approximately

100-g of dog pellets in the late afternoon for the second night.

163

164

To investigate whether sample sites were independent of one another, we tested for spatial autocorrelation in the residuals of the fitted values for each behaviour, in each exclosure, using Moran’s index (i), calculated in the spatial analyst module of ArcGis v10.3 (Table S4.1 in the Supplementary information). Spatial autocorrelation occurs when the value of a variable at any one location in space can be predicted by the values of nearby locations. The existence of spatial autocorrelation indicates that sampling units are not independent of one another.

4.3.5 Behavioural scoring

(Table 4.2) (Mazzamuto et al., 2018). With the EB approach the researcher defines groups of behaviours, with each group related to specific behavioural responses. We then scored video recordings ≤ 60-s using the event recorder JWatcher (Blumstein and

Daniel, 2007) from which we calculated the proportion of time in sight allocated to each behaviour. We quantified the behaviour of the first burrowing bettong to approach and/or forage at the site. Videos with more than one individual foraging were excluded to reduce the influence of conspecific presence on the behavioural response to the treatments. We were unable to group individuals into age or sex demographics due to the limitations of infrared camera technology.

For analysis we combined behaviours in order to form five main categories.

“Wary approach” was comprised of any sniffing behaviour, including bi-pedal and quadrupedal sniffing and slow movement towards the feeding station. “Fast approach”

165

was comprised of rapid movement towards the feeding station. “Foraging” included foraging with the head raised and lowered. “Investigate” comprised of any sniffing, touching or chewing behaviour of the visual treatments. “Retreat” included rapid and slow movement away from the foraging site. Data was recorded blind in regards to exclosure; however it was not possible to record data blind in terms of treatment because our study involved quantifying the behaviour of animals interacting with a visual stimulus.

Table 4.2 Ethogram of burrowing bettong (Bettongia lesueur) behaviour.

Behaviour Category Behaviour Description a

Wary approach Bi-pedal sniff Animal looks and/or sniffs air whilst standing upright on hind limbs

Prone sniff Animal looks and/or sniffs the air and/or ground whilst standing on all four limbs

Slow approach Animal moves slowly towards feeding station

Fast approach Fast approach Animal moves quickly and directly towards feeding station

Foraging Vigilant foraging Animal chews with its head up and observing surroundings

Relaxed foraging Animal forages for food and chews with its head down without observing surroundings

166

Investigate model Investigate model Animal investigates model type through smell and/or touch

Chew model Animal attempts to eat the model type

Retreat Recoil Animal recoils away from the feeding station and/or model type

Retreat Animal retreats away from feeding station and/or model type

Out of sight Out of sight Animal seen on camera and retreated back out of sight away from camera a Definition of postures associated with particular behaviours

4.3.6 Analysis of behavioural data

Conducting experiments that manipulate the populations of reintroduced threatened species and introduced predators are rarely possible due to the large spatial extent required for manipulations, and legal, political, and ethical considerations. The introduction of a known number of feral cats into a managed, fenced population of reintroduced burrowing bettongs provided a rare opportunity to test naïve prey species’ ability to recognize and respond to introduced predators as a result of experience. Because feral cats were present only in one section of the reserve and absent throughout the remainder of the reserve, our treatments were by necessity spatially segregated such that, according to a strictly statistical approach, our experimental design does not allow us to run a model comparing “no predator” vs.

“cat enclosure” (Hurlbert, 1984). However, we have contrasted their responses to the

167

treatments within each location, in order to assess the influence of lifetime experience on bettong predator visual recognition abilities and behaviour.

We used a generalized estimating equation (GEE) model with independent error structure and binomial distribution in in SPSS-22 (IBM Corp. Armonk, NY, U.S.A) to investigate if bettong visitation to feeding stations on night 2 was dependent on the type of model placed at the feeding station. The response variable was visit (1) or no visit (0).

Mazzamuto et al. (2018) found that Principal Component Analysis (PCA) scores correlated significantly with EB behaviours. To validate our EB approach with a statistical method, we ran a PCA of the five behaviours measured within this study for brush-tailed bettongs living in the no predator and cat exclosure independently. The initial factorial solutions were rotated by the Varimax procedure (Sokal & Rohlf 1995)

(Fig. 4.4 and 4.6).

To test whether model type caused burrowing bettongs to allocate different proportions of time to composite behaviours, in the absence and presence of predators, we fitted a series of generalized estimating equation (GEE) models with an independent error structure and linear distribution. The factor in our model was model type (cat, rabbit, bucket and control). As we could not distinguish between individual burrowing bettongs and account for the possibility of non-independence between observations, we included foraging station location as a repeated measure in our models. Sampling period (October 2015 and March 2016) was made a covariate in the models. We used a GEE model, rather than a Generalized Linear Mixed Model (GLMM), as GEEs allow analysis of a repeated measures design (Quinn and Keough 2002). A GEE

168

model can handle correlated non-normally distributed data, which is very common in the behavioural sciences (Pekár and Brabec, 2017). They are most suitable when the main factor or interest is between subjects and within subjects components represent repeated observations through time (Omar et al., 1999, Quinn and Keough 2002). GEEs are also able to handle missing data effectively, as long as the observations missing are completely random, unlike classical ANOVA type models for repeated measure designs that are unable to handle missing observations very effectively (Quinn and Keough

2002). One of the strengths of GEEs is that, although the correct specification of the correlation structure makes estimation more efficient, parameter estimates are usually consistent even if the wrong correlation structure is used, meaning the estimates of the model parameters are not very sensitive to the choice of correlation structure

(Quinn and Keough, 2002, Omar et al. 1999). A limitation to a GEE is that it is very efficient when the design of a study can be modelled using any of the predefined structures, however if the design is more complicated, such as more random effects either crossed or nested (multiple level design), than use of a GLMM is more efficient

(Pekár and Brabec, 2017). A further limitation to the use of GEEs it that they are most applicable when the pattern of observation through time for experimental units is not the main research question (Omar et al. 1999). GEEs are less useful if the within- subjects component is a factor of specific interest (Quinn and Keough 2002). The data fulfilled all the assumptions required to perform a GEE analysis.For the analysis of

‘investigate visual’ behaviours, in which an individual investigated models through smell and/or touch, the control was removed from the analysis, because no model was present. In instances where the effect of model type was significant (P < 0.05), we used

Bonferroni post-hoc analysis to examine planned comparisons for differences in

169

response to each model type (cat vs. rabbit, cat vs. bucket, cat vs. control, rabbit vs. bucket, rabbit vs. control and bucket vs. control), as we wished to understand the pattern of responses.

4.3.7 Ethical Note

All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. Work was conducted under animal ethics APEC Approval

Number 1/2014 Tackling Prey Naivety in Australia’s Threatened Mammals, in accordance with South Australian Wildlife Ethics Committee, and ACEC Approval

Number 15/19A, in accordance with The Australian Code of Practice for the Care and

Use of Animals for Scientific Purposes (1997).

170

4.4 RESULTS

4.4.1 Visits to Station

There was no effect of model type (Wald χ2 = 0.304, P = 0.859) in the “no predator” and “cat enclosure” (Wald χ2 = 0.272, P = 0.873) on visitation to feeding stations (Table

4.3). Thus bettongs were equally likely to visit the stations regardless of the specific model present.

Table 4.3 The number of foraging stations (n) visited by bettongs by model type. The number in parentheses indicates the total number of feeding stations deployed with each model type from which videos were analysed for bettong behavioural responses.

Night 2 Night 2 Night 2 Night 2

Bucket Cat Control Rabbit

No Predator 23 (16) 23 (13) 19 (14) 22 (18)

Cat Enclosure 17 (11) 13 (12) 13 (13) 14 (8)

4.4.2 Behavioural Response to Model Type

4.4.2.1 No predator treatment

The PCA for behavioural measurements of burrowing bettongs in the “no predator” exclosure in response to odour treatments produced two components that together accounted for 59.26% of the variance. The first component of the PCA (component 1,

Eigen value = 1.701; % variance explained = 34.01) was positively correlated with variables describing wary approach (rotated PCA score = 0.889, Fig.4.3) and retreat

171

behaviours (rotated PCA score = 0.091, Fig.4.3). Foraging behaviours (rotated PCA score = -0.905, Fig.4.3) were found to be negatively correlated with component 1. The means derived from the factor analysis for component 1 for the bucket and no visual were also negatively correlated with PCA component 1 (Fig.4.3). The means derived from the factor analysis for component 1 for cat and rabbit were positively correlated with component 1 Fig.4.3). Component 2 (Eigen value = 1.262; % variance explained =

25.241) was characterised primarily by a positive correlation of investigate (rotated

PCA score = 0.788, Fig.4.3) and fast approach (rotated PCA score = 0.647, Fig.4.3) behaviours. Based on the PCA it appears that individuals are trading off wary approach against forage behaviours. Based on the PCA it would appear that burrowing bettongs in the “no predator” area are not differentiating their behaviour in response to the model types. There appears to be a good correspondence between the grouping of behaviours from our EB method and that obtained with the PCA.

172

No Predator 1.000 0.800 0.600 Investigate Fast Approach

0.400

0.200 Cat 0.000 Rabbit

-0.200 Bucket Retreat Component 2 Component -0.400 Control Wary Approach -0.600 -0.800 Forage -1.000 -1.000 -0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600 0.800 1.000 Component 1

In the “no predator” exclosure there was a significant effect of model type on the proportion of time that burrowing bettongs allocated to wary approach (Wald χ2 =

9.636, P = 0.022, Table 4.4, Fig. 4.5a) and retreat (Wald χ2 = 8.380, P = 0.039, Table 4.4,

Fig. 4.5c) behaviours. We found no significant effect of model types in the “no predator” exclosure on the proportion of time burrowing bettongs allocated to foraging, fast approach and investigate visual behaviours (Table 4.4, Fig. 4.5b, d and e).

There was no spatial autocorrelation in the residuals of the fitted values for any of the analysed behaviours in the “no predator” exclosure (Table S4.1 in the Supplementary

173

information). These results indicate that the feeding stations were independent for the purpose of our analysis.

Within the no predator area, planned post-hoc comparisons revealed that burrowing bettongs spent significantly more time in wary approach behaviour when approaching the rabbit compared to the control (Bonferroni, rabbit vs. control, P =

0.043; Fig. 4.5a). Bettongs did not modify time spent warily approaching the cat or bucket compared to the control (Bonferroni, cat vs. control, P = 0.074, bucket vs. control, P = 0.909; Fig 4.5a). There was no significant difference between the proportion of time spent in wary approach behaviours between the three model types

(Bonferroni, bucket vs. cat, P = 0.999, bucket vs. rabbit, P = 0.614 and cat vs. rabbit, P =

0.999; Fig. 4.5a).

There were no significant differences in the proportion of time spent by bettongs retreating between the models (Bonferroni, cat vs. rabbit, P = 0.999, cat vs. bucket, P = 0.999, and rabbit vs. bucket, P = 0.252) and procedural control (blank)

(Bonferroni, cat vs. control, P = 0.999, rabbit vs. control, P = 0.458, and bucket vs. control, P = 0.999). A detailed list of planned comparisons conducted for behaviours recorded in the no cat treatment may be found in Table S4.2 in the supplementary information.

174

Figure 4.5 The mean (± SEMs) proportion of time in sight (PIS) that burrowing bettongs

(Bettongia lesueur) allocated to the behaviours in response to model types, in the “no predator” study area (bucket n = 16, cat n = 13, control n = 14 and rabbit n = 18).

Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences.

175

4.4.2.2 Cat enclosure treatment

The PCA for behavioural measurements of burrowing bettongs in the cat exclosure in response to odour treatments produced two components that together accounted for

63.06% of the variance. The first component of the PCA (component 1, Eigen value =

1.782; % variance explained = 35.65) was positively correlated with variables describing wary approach (rotated PCA score = 0.889, Fig.4.6) and retreat behaviours

(rotated PCA score = 0.633, Fig.4.6).Fast approach (rotated PCA score = -0.121, Fig.4.6) and forage (rotated PCA score = -0.695, Fig.4.6) behaviours were found to be negatively correlated with component 1. The means derived from the factor analysis for component 1 for the rabbit, bucket and no visual were also negatively correlated with PCA component 1 (Fig.4.6). The means derived from the factor analysis for component 1 for predator visual treatment (cat) positively correlated with component

1 Fig.4.6). Component 2 (Eigen value = 1.371; % variance explained = 27.418) was characterised primarily by a positive correlation of investigate behaviour (rotated PCA score = 0.931, Fig.4.6). Based on the PCA it appears that individuals are trading off wary approach and retreat behaviours against fast approach and forage behaviours.

Based on the PCA it would appear that burrowing bettongs living in the presence of cats are investing more time in wary approach behaviours when a cat is present, compared to bucket, rabbit and control. There appears to be a good correspondence between the grouping of behaviours from our EB method and that obtained with the

PCA.

176

Cat Enclosure 1.200

1.000

0.800 Investigate

0.600

0.400

Bucket Cat 0.200 Fast Approach 0.000 Wary Approach

Component2 -0.200 Rabbit

-0.400 Retreat -0.600 Control Forage -0.800 -1.000 -0.500 0.000 0.500 1.000 1.500 Component 1

In the “cat enclosure” there was a significant effect of model type on the proportion of time that burrowing bettongs allocated to wary approach (Wald χ2 =

36.052, P ≤ 0.005, Table 4.4, Fig. 4.7a) and forage (Wald χ2 = 15.864, P ≤ 0.005, Table

4.4, Fig. 4.7b) behaviours. We found no significant effect of model type in the “cat enclosure” on the proportion of time burrowing bettongs allocated to retreat, fast approach and investigate visual behaviours (Table 4.4, Fig. 4.7c – e). There was no spatial autocorrelation in the residuals of the fitted values for any of the analysed behaviours in the “cat enclosure” (Table S4.1 in the Supplementary information).

177

These results indicate that the feeding stations were independent for the purpose of our analysis.

Post-hoc tests revealed that within the “cat enclosure”, burrowing bettongs spent significantly more time in wary approach behaviour when investigating the cat model than they did the rabbit (Bonferroni, cat vs. rabbit, P ≤ 0.005; Fig. 4.7a), the bucket (Bonferroni, cat vs. bucket, P ≤ 0.005; Fig. 4.7a) and the control (no model)

(Bonferroni, cat vs. control, P = 0.02; Fig. 4.7a). There were no differences in the proportion of time allocated to wary approach between the rabbit, novel object and control (Bonferroni, rabbit vs. bucket, P = 0.999, rabbit vs. control, P = 0.999 and bucket vs. control, P = 0.999; Fig. 4.7a).

Where cats were present, planned comparisons revealed that bettongs spent significantly less time foraging at the cat models than the control (Bonferroni, cat vs. control, P ≤ 0.005; Fig. 4.7b). There was no significant difference in time allocated to foraging between the models (Bonferroni, cat vs. rabbit, P = 0.208, cat vs. bucket, P =

0.647, and rabbit vs. bucket, P = 0.999; Fig. 4.7b), nor between the rabbit and novel object compared to the control (Bonferroni, rabbit vs. control, P = 0.999, and bucket vs. control, P = 0.727; Fig. 4.7b). A detailed list of planned comparisons conducted for behaviours recorded in the “cat enclosure” may be found in Table S4.3 in the supplementary information.

178

Exclosure Type Behaviour df Wald χ2 p-value

No Predator Wary Approach 3 9.636 0.022**

Forage 3 7.661 0.054

Retreat 3 8.380 0.039**

Fast Approach 3 1.215 0.749

Investigate 2 1.604 0.448

Cat Enclosure Wary Approach 3 36.052 ≤0.005**

Forage 3 15.864 ≤0.005**

Retreat 3 4.023 0.259

Fast Approach 3 2.321 0.508

Investigate 2 0.999 0.607

Values in italic** indicate significant differences

179

Figure 4.7 The mean (± SEMs) proportion of time in sight (PIS) that burrowing bettongs

(Bettongia lesueur) allocated to the behaviours in response to model types, in the “cat enclosure” study area (bucket n = 11, cat n = 12, control n = 13 and rabbit n = 8).

Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences.

180

4.5 DISCUSSION

The results from our experiments are consistent with the learned recognition hypothesis and our a priori predictions that a prey species’ ability to visually recognize predators is inducible through experience. Bettongs living in the predator free exclosure approached all models cautiously. This finding suggests that bettongs responded to the presence of a model, but did not discriminate between a predatory threat (cat), a non-threatening herbivore (rabbit) and a novel object (bucket).

Burrowing bettongs in the predator free exclosure also appeared to respond to the rabbit model compared to the control (no model), expressed through wary approach behaviour. Bettongs did not differentiate their wary approach behaviour between the models (cat, rabbit and bucket). One hypothesis to explain this response is that rabbits may be competitors with burrowing bettongs and this competition has resulted in the evolution of an aversive response to rabbits. The similarities in response of burrowing bettongs to a potential competitor (rabbits), a predator (cat) and procedural control

(bucket) could also reflect the limited discriminative ability of this species, based on their lack of ontogenetic experience with predators, rather than the fact that it is not a fearful response.

In stark contrast, bettongs living with cats adjusted their behaviour in the presence of cat models. Bettongs in the cat enclosure allocated the most time to wary approach behaviour when a cat model was present, compared to the rabbit, bucket and procedural control. That bettongs in the cat enclosure differentiated their response to the visual models and were only wary towards the cat models suggests that visual recognition is associated with a specific predatory threat and is consistent with previous studies that have investigated learned behavioural response towards

181

predators (Herzog and Hopf, 1984; Mineka and Cook, 1988). In addition, our results support the ideas proposed by Berger, Swenson, & Persson (2001) and Griffin, Evans &

Blumstein (2001) who suggested that if prey are exposed to a novel threat and survive the encounter, then their response towards predators may persist.

Although analysis found that there was a significant effect of model type in the proportion of time burrowing bettongs in the predator free exclosure allocated to retreat behaviour, post-hoc analysis suggested that there was no significant differences in the proportion of time spent by bettongs retreating in response to model type (bucket, cat, rabbit and control – no visual), This non-result may be a caveat of the type of post-hoc analysis used. Bonferroni post-hoc analysis adjusts significance levels to control for Type I error rates in multiple testing situations (Quinn and Keough 2002). Although Bonferroni post-hoc analysis provides great control over

Type I error, it is very conservative when there are lots of comparisons; causing comparisons to have decreasing power as the number of comparisons increase (Quinn and Keough 2002). Previous studies have found that animals living with predators commonly trade off foraging with anti-predator vigilance (Bednekoff and Lima, 1998;

Griffin et al., 2000; Beauchamp, 2015). Bettongs in the no predator exclosure did not adjust their foraging behaviour in response to the different model treatments. Our results suggest bettongs living with no predators were “predator naïve” because they were unable to differentiate between the model types, such as a predator, herbivore

(rabbit), novel object (bucket) and the procedural control (no model). We presume that this naïveté was due to their lack of ontogenetic and minimal evolutionary history with cats (Atkins, Blumstein, Moseby, West, Hyatt, & Letnic, 2016).

182

In contrast to the bettongs living in the predator free environment, bettongs that were coexisting with cats spent the least amount of time foraging when cat models were present. Previous studies have similarly found that experience, rather than evolutionary history, strongly influences prey response to predators (Berger et al.,

2001; Martin, 2014). For example, studies of learning in fish, have found that fish learned to respond more intensely to predator cues associated with high risk (Ferrari et al., 2005). This implies that while a species may not respond appropriately upon first encountering a predator (Mirza et al., 2006), experience and rapid learning may play a key role in the development of antipredator behaviours.

Visual predator recognition sometimes depends on cues such as shape or the presence of frontally located eyes (Curio, 1993; Coss and Goldthwaite, 1995).

Carnivores have binocular vision, whilst herbivores have eyes on the sides of their heads (Blumstein et al., 2000). Previous studies have also suggested that prey may use the apparent size of models to assess risk (Evans et al., 1993; Blumstein et al., 2002). If stimulus size were important we would expect similar responses to both the cat and bucket models because they are of similar height (Fig 3a, c). Bettongs living in the “no predator” exclosure did not appear to alter their behaviour according to model size or eye location, with no significant differences in behaviour towards cat, bucket and rabbit models. It is possible that olfactory cues associated with the taxidermy mounts explained some variation in the burrowing bettong response. However, taxidermy mounts of rabbits and cats were prepared in the same tanning solution and were moved across both study sites at random. As such if there was a significant effect of olfactory cues associated with the mounts, we would have expected to see this

183

response across both the “no predator” and “cat enclosure” study areas and this was not the case.

It is possible that the use of taxidermy models may underestimate the predatory response bettongs may display if exposed to a live predator. This has been noted in a study of peahens, which emitted louder antipredator calls when exposed to a live predator, compared to a model (Yorzinski & Platt, 2012). It was assumed that this response reflected the greater threat a live predator posed compared to a model in the captive experiments.

Studies have also found that the eyesight of nocturnal species is more acute and variable than previously recognised (Bearder et al., 2006). Although nocturnal mammals are primarily orientated by olfactory and audible senses (Mascalzoni and

Regolin, 2011), it has been found that nocturnal mammals are well adapted to low light environments (Heesy and Hall, 2010, Mascalzoni and Regolin, 2011), allowing them to visually discriminate objects (Wynne and McLean, 1999), as would appear to be the case in this study.

Prior to this study bettongs at Arid Recovery had no ontogenetic experience and minimal evolutionary exposure to placental predators. Although the founding population of bettongs at Arid Recovery may have had brief evolutionary exposure to cats on Bernier Island (Atkins, Blumstein, Moseby, West, Hyatt, & Letnic, 2016) and

Herrison Prong (Short and Turner, 2000), the results from the “no predator” exclosure suggest that this experience did not influence their visual recognition abilities. The ability of bettongs living within the “cat enclosure” to visually discriminate a predator from an herbivore and novel object suggests that experience is essential to develop

184

and induce learned predator recognition. These findings are consistent with previous studies that have reported learnt predator recognition (Brown et al., 2006; Ferrari et al., 2006; Ferrari et al., 2010) and the strong role that learning (Martin, 2014) and level of predation risk (Chivers and Smith, 1994; Bøving and Post, 1997; Griffin et al., 2001) play in structuring antipredator responses.

The ability to visually discriminate between a predator and herbivore and novel object might be characteristic of the early stages of the evolution of predator recognition. According to the Baldwin theory phenotypic plasticity allows organisms to adapt to its environment during its lifetime (Turney et al., 1996). In the first instance, phenotypic plasticity allows an individual to adapt, such as the ability to learn. Given sufficient time, evolution may find a rigid mechanism to replace the plastic mechanism, such that a learned behaviour may become instinct (Turney et al., 1996).

It is possible that over evolutionary time burrowing bettongs may be able to develop appropriate instinctive responses to cats.

A caveat of this study is that we were unable to test for the mechanisms leading to our results. As burrowing bettongs are a social species, it is possible that individuals may have exploited the expertise and individually acquired predator avoidance behaviours of others, through social learning (Kavaliers et al., 2001; Griffin and Evans, 2003). However due to the limitations of this study, we were unable to test for this. As we were studying a wild population of bettongs, we were unable to determine whether a bettong had encountered a predation event with a cat. As such we can only make assumptions that bettongs living within the “cat enclosure” have encountered cats, either directly or indirectly, within their lifetime, in comparison to bettongs living within the “no predator” enclosure, who have no ontogenetic

185

experience with cats. We further make the assumptions that through individual experience or through social learning, it may be presumed that learned recognition occurs because there is an evolutionary advantage for prey to learn quickly in response to novel predators.

From a conservation perspective, understanding the rate of behavioural adaptation by prey to a novel predator is of great importance. Our certainty regarding life time predator experiences of burrowing bettongs (whether individually and/or socially) in this study, provides a unique insight into the influence that predation pressure can have in the development of anti-predator behaviours and the time frames over which this can occur. We know that bettongs have had minimal evolutionary exposure to novel predators, such as cats. We know that the bettongs at

Arid Recovery have had no ontogenetic exposure to placental mammalian predators within the “no predator” exclosure, and we also know how long bettongs have been exposed to cat predation within the “cat enclosure”. This is in stark contrast to other studies in which history of predator exposure is unknown (Banks et al., 2002; Anson and Dickman, 2013).

Our results suggest that bettongs with no ontogenetic and minimal evolutionary exposure to feral cats can rapidly (within 8 – 15 months of predator exposure) acquire predator recognition abilities after exposure to a novel predator.

From a reintroduction perspective this suggests that anti-predator recognition in evolutionarily predator naïve prey can be induced through experience. Such learned antipredator responses could be utilized in pre-release prey training, through exposure of prey individuals to predators under carefully controlled conditions (Moseby et al.,

2016). However, we acknowledge that demonstrating the utility of predator exposure

186

as a pre-release strategy requires actually demonstrating that in-situ predator exposure confers a fitness advantage.

187

4.6 ACKNOWLEDGEMENTS

We thank the Arid Recovery staff for their assistance with the study. Thanks to Mike

Letnic, Sharon Ryall, Claire Sives and Jamie Dunlop for their assistance with taxidermy.

4.7 REFERENCES

Anson, J. R., Dickman, C. R., 2013. Behavioral responses of native prey to disparate

predators: naiveté and predator recognition. Oecologia 171(2):367-377. doi:

10.1007/s00442-012-2424-7

Apfelbach, R., Blanchard, C.D., Blanchard, R.J., Hayes, R.A., McGregor, I.S., 2005. The

effects of predator odors in mammalian prey species: a review of field and

laboratory studies. Neuroscience & Biobehavioral Reviews 29(8):1123-1144.

doi: 10.1016/j.neubiorev.2005.05.005

Atkins, R., D. T. Blumstein, K. E. Moseby, R. West, M. Hyatt, Letnic, M., 2016. Deep

evolutionary experience explains mammalian responses to predators.

Behavioral Ecology and Sociobiology 70(10):1-9. doi: 10.1007/s00265-016-

2181-4

Banks, P., Nelika, K., Hughes, A., Rose, T., 2002. Do native Australian small mammals

avoid faeces of domestic dogs? Responses of Rattus fuscipes and Antechinus

stuartii. Australian Zoologist 32(3):406-409. doi: 10.7882/AZ.2002.018

Bearder, S. K., Nekaris, K. A. I., Curtis, D. J., 2006. A re-evaluation of the role of vision in

the activity and communication of nocturnal primates. Folia Primatologica 77(1-

2): 50-71. doi: 10.1159/000089695

188

Beauchamp, G., 2004. Reduced flocking by birds on islands with relaxed predation.

Proceedings of the Royal Society of London, Series B: Biological Sciences

271(1543):1039-42. doi: 10.1098/rspb.2004.2703

Beauchamp, G., 2015. Animal Vigilance: Monitoring Predators and Competitors,

Academic Press, San Diego, USA, pp. 23-47.

Bednekoff, P.A., Lima, S.L., 1998. Randomness, chaos and confusion in the study of

antipredator vigilance. Trends in Ecology & Evolution 13(7):284-287. doi:

10.1016/S0169-5347(98)01327-5

Berger, J., Swenson, J.E., Persson, I. L., 2001. Recolonizing Carnivores and Naïve Prey:

Conservation Lessons from Pleistocene Extinctions. Science 291(5506):1036-

1039. doi: 10.1126/science.1056466

Berger, J., Terborgh, J., Estes, J., 2010. Fear-mediated food webs, in: Terborgh, J.E.,

James A., editors, Trophic Cascades: Predators, Prey, and the Changing

Dynamics of Nature, Island Press, Washington, DC, pp. 241-253.

Blumstein, D. T., Daniel. J. C., 2005. The loss of anti-predator behaviour following

isolation on islands. Proceedings of the Royal Society B: Biological Sciences

272(1573):1663-1668. doi: 10.1098/rspb.2005.3147

Blumstein, D.T., Daniel, J.C., 2007. Quantifying Bahaviour the JWatcher Way. Sinauer

Associates, Inc., Sunderland, Massachusetts, U.S.A.

Blumstein, D.T., Daniel, J.C., Griffin, A.S., Evans, C.S., 2000. Insular tammar wallabies

(Macropus eugenii) respond to visual but not acoustic cues from predators.

Behavioral Ecology 11(5):528-535. doi: 10.1093/beheco/11.5.528

189

Blumstein, D.T., Daniel, J.C., Schnell, M.R., Ardron, J.G., Evans, C.S., 2002. Antipredator

behaviour of red-necked pademelons: a factor contributing to species survival?

Animal Conservation 5(4):325-331. doi: 10.1017/S1367943002004080

Bøving, P.S., Post, E., 1997. Vigilance and foraging behaviour of female caribou in

relation to predation risk. Rangifer 17(2):55-63. doi: 10.7557/2.17.2.1302

Brodie, E.I., Brodie, E.J., 1999. Predator-prey arms races. Bioscience 49(7):557-568. doi:

10.2307/1313476

Brown, G.E., Chivers, D.P., 2005. Learning as an adaptive response to predation, in:

Barbosa, P., Castellanos, I., editors. Ecology of predator/prey interactions.,

Oxford University Press, Oxford, p. 34-54.

Brown, G.E., Rive, A.C., Ferrari, M.C.O., Chivers, D.P., 2006. The dynamic nature of

antipredator behavior: prey fish integrate threat-sensitive antipredator

responses within background levels of predation risk. Behavioral Ecology and

Sociobiology 61(1):9-16. doi: 10.1007/s00265-006-0232-y

Bureau of Meteorology, A.G., 2016. Mean rainfall (mm) for Roxby Downs, South

Australia, years 1997 to 2016 http://www.bom.gov.au/climate/data.

Bytheway, J.P., Price, C.J., Banks, P.B., 2016. Deadly intentions: naïve introduced foxes

show rapid attraction to odour cues of an unfamiliar native prey. Scientific

Reports 6:30078. doi: 10.1038/srep30078

Carrete, M., Tella, J.L., 2015. Rapid loss of antipredatory behaviour in captive-bred

birds is linked to current avian invasions. Scientific reports 5:18274. doi:

10.1038/srep18274

190

Carthey, A.J., Banks, P.B., 2012. When Does an Alien Become a Native Species? A

Vulnerable Native Mammal Recognizes and Responds to Its Long-Term Alien

Predator. PLoS One 7(2):e31804. doi: 10.1371/journal.pone.0031804

Carthey, A. J., Banks, P. B., 2014. Naïveté in novel ecological interactions: lessons from

theory and experimental evidence. Biological Reviews 89(4):932-949. doi:

10.1111/brv.12087

Carthey, A.J., Banks, P.B., 2016. Naiveté is not forever: responses of a vulnerable native

rodent to its long term alien predators. Oikos 125(7):918-926. doi:

10.1111/oik.02723

Chivers, D. P., R. S. Mirza, P. J. Bryer, Kiesecker, J. M., 2001. Threat-sensitive predator

avoidance by slimy sculpins: understanding the importance of visual versus

chemical information. Canadian Journal of Zoology 79(5):867-873. doi:

10.1139/z01-049

Chivers, D.P., Smith, R.J.F., 1994. Fathead minnows, Pimephales promelas, acquire

predator recognition when alarm substance is associated with the sight of

unfamiliar fish. Animal Behaviour 48(3): 597-605. doi: 10.1006/anbe.1994.1279

Cooper, W.E., Pyron, R.A., Garland, T., 2014. Island tameness: living on islands reduces

flight initiation distance. Proceedings of the Royal Society B: Biological Sciences

281(1777):20133019. doi: 10.1098/rspb.2013.3019

Coss, R.G., Goldthwaite, R.O., 1995. The persistence of old designs for perception. In:

N. S. Thompson, N.S., Editor. Perspectives in ethology, Plenum Press, New York,

USA, pp. 83-148.

191

Curio, E., 1993. Proximate and developmental aspects of antipredator behavior,

Advances in the Study of Behavior, Academic Press, San Diego, USA, pp. 135-

238.

Dawkins, R., Krebs, J.R., 1979. Arms races between and within species. Proceedings of

the Royal Society of London B: Biological Sciences 205(1161):489-511. doi:

10.1098/rspb.1979.0081

Engbring, J., Fritts, T.H., 1988. Demise of an insular avifauna: the brown tree snake on

Guam. Transactions of the Western Section of the Wildlife Society 24:31-37.

Epp, K. J., Gabor, C. R., 2008. Innate and learned predator recognition mediated by

chemical signals in Eurycea nana. Ethology 114(6):607-615. doi:

10.1111/j.1439-0310.2008.01494.x

Evans, C.S., Macedonia, J.M., Marler, P., 1993. Effects of apparent size and speed on

the response of chickens, Gallus gallus, to computer-generated simulations of

aerial predators. Animal Behaviour 46(1):1-11. doi: 10.1006/anbe.1993.1156

Ferrari, M.C., Messier, F., Chivers, D.P., 2006. The nose knows: minnows determine

predator proximity and density through detection of predator odours. Animal

Behaviour 72(4):927-932. doi: 10.1016/j.anbehav.2006.03.001

Ferrari, M. C., J. J. Trowell, G. E. Brown, Chivers, D. P., 2005. The role of learning in the

development of threat-sensitive predator avoidance by fathead minnows.

Animal Behaviour 70(4):777-784. doi: 10.1016/j.anbehav.2005.01.009

Ferrari, M.C., Wisenden, B.D., Chivers, D.P., 2010. Chemical ecology of predator-prey

interactions in aquatic ecosystems: a review and prospectus. Canadian Journal

of Zoology 88(7):698-724. doi: 10.1139/Z10-029

192

Ferrari, M. C. O. 2014. Short-term environmental variation in predation risk leads to

differential performance in predation-related cognitive function. Animal

Behaviour 95:9-14. doi: 10.1016/j.anbehav.2014.06.001

Finlayson, G. R., Moseby, K. E., 2004. Managing confined populations: the influence of

density on the home range and habitat use of reintroduced burrowing bettongs

(Bettongia lesueur). Wildlife Research 31(4):457-463. doi: 10.1071/WR03035

Griffin, A. S., D. T. Blumstein, Evans, C. S., 2000. Training captive‐bred or translocated

animals to avoid predators. Conservation biology 14(5):1317-1326. doi:

10.1046/j.1523-1739.2000.99326.x

Griffin, A.S., Evans, C.S., 2003. Social learning of antipredator behaviour in a marsupial.

Animal Behaviour 66(3):485-492. doi: 10.1006/anbe.2003.2207

Griffin, A.S., Evans, C.S., Blumstein, D.T., 2001. Learning specificity in acquired predator

recognition. Animal Behaviour 62(3):577-589. doi: 10.1006/anbe.2001.1781

Hancock, S., Brown, P., Stephens, B., 2000. Shark Bay Terrestrial Reserves Management

Plan 2000–2009. Western Australia, Department of Conservation and Land

Management for the National Parks and Nature Conservation Authority.

Hayward, M. W., Moseby, K., Read, J. L., 2014. The role of predator exclosures in the

conservation of Australian fauna. In A. Glen and C. Dickman, editors, Carnivores

of Australia: past, present and future. CSIRO Publishing Sydney. p. 353-371.

Heesy, C. P., Hall, M. I., 2010. The nocturnal bottleneck and the evolution of

mammalian vision. Brain, behavior and evolution 75(3): 195-203. doi:

10.1159/000314278

193

Herzog, M., Hopf, S., 1984. Behavioral responses to species‐specific warning calls in

infant squirrel monkeys reared in social isolation. American Journal of

Primatology 7(2):99-106. doi: 10.1002/ajp.1350070204

Holmes, T.H., McCormick, M.I., 2010. Smell, learn and live: the role of chemical alarm

cues in predator learning during early life history in a marine fish. Behavioural

Processes 83(3):299-305. doi: 10.1016/j.beproc.2010.01.013

Hurlbert, S.H., 1984. Pseudoreplication and the design of ecological field experiments.

Ecological monographs 54:187-211. doi: 10.2307/1942661

Johnson, C. 2006. Australia's Mammal Extinctions: A 50,000 year history. Cambridge

University Press, Port Melbourne. p. 1-278.

Kavaliers, M., Choleris, E., Colwell, D., 2001. Learning from others to cope with biting

flies: social learning of fear-induced conditioned analgesia and active

avoidance. Behavioral neuroscience 115(3):661-674. doi: 10.1037//0735-

7044.115.3.661

Maloney, R. F., McLean, I. G., 1995. Historical and experimental learned predator

recognition in free-living New-Zealand robins. Animal Behaviour 50(5):1193-

1201. doi: 10.1016/0003-3472(95)80036-0

Martin, C. W. 2014. Naive prey exhibit reduced antipredator behavior and survivorship.

PeerJ 2:e665. doi: 10.7717/peerj.665

Mascalzoni, E., Regolin, L., 2011. Animal visual perception. Wiley Interdisciplinary

Reviews: Cognitive Science, 2(1): 106-116. doi: 10.1002/wcs.97

McCormick, M., Holmes, T., 2006. Prey experience of predation influences mortality

rates at settlement in a coral reef fish, Pomacentrus amboinensis. Journal of

Fish Biology 68(3):969-974. doi: 10.1111/j.0022-1112.2006.00982.x

194

McLean, I.G., Lundie-Jenkins, G., Jarman, P.J., 1996. Teaching an endangered mammal

to recognise predators. Biological Conservation 75(1):51-62. doi: 10.1016/0006-

3207(95)00038-0

Mineka, S., Cook, M., 1988. Social learning and the acquisition of snake fear in

monkeys. In: T. R. Zentall and B. G. Galef, Jr., editors, Social learning:

Psychological and biological perspectives. Erlbaum, New Jersey. p. 51-73.

Mirza, R.S., Ferrari, M.C.O., Kiesecker, J.M., Chivers, D.P., 2006. Responses of American

Toad Tadpoles to Predation Cues: Behavioural Response Thresholds, Threat-

Sensitivity and Acquired Predation Recognition. Behaviour 143(7):877-889. doi:

10.1163/156853906778017926

Moseby, K., Read, J., 2006. The efficacy of feral cat, fox and rabbit exclusion fence

designs for threatened species protection. Biological Conservation 127(4):429-

437. doi: 10.1016/j.biocon.2005.09.002

Moseby, K. E., Blumstein, D. T., Letnic, M., 2016. Harnessing natural selection to tackle

the problem of prey naïveté. Evolutionary applications 9(2):334-343. doi:

10.1111/eva.12332

Moseby, K.E., Cameron, A., Crisp, H.A., 2012. Can predator avoidance training improve

reintroduction outcomes for the greater bilby in arid Australia? Animal

Behaviour 83(4):1011-1021. doi: 10.1016/j.anbehav.2012.01.023

Moseby, K. E., J. L. Read, D. C. Paton, P. Copley, B. M. Hill, Crisp, H. A., 2011. Predation

determines the outcome of 10 reintroduction attempts in arid South Australia.

Biological Conservation 144(12):2863-2872. doi: 10.1016/j.biocon.2011.08.003

Oksanen, L., 2001. Logic of experiments in ecology: is pseudoreplication a

pseudoissue? Oikos 94(1):27-38. doi: 10.1034/j.1600-0706.2001.11311.x

195

Omar, R. Z., Wright, E. M., Turner, R. M., Thompson, S. G., 1999. Analysing repeated

measurements data: a practical comparison of methods. Statistics in Medicine

18(13): 1587-1603. doi: 10.1002/(SICI)1097-0258(19990715)18:13<1587::AID-

SIM141>3.0.CO;2-Z

Pekár, S. and M. Brabec. 2017. Generalized estimating equations: A pragmatic and

flexible approach to the marginal GLM modelling of correlated data in the

behavioural sciences. Ethology 124(2):86-93. doi.org/10.1111/eth.12713

Phillips, B.L., Shine, R., 2006. An invasive species induces rapid adaptive change in a

native predator: cane toads and black snakes in Australia. Proceedings of the

Royal Society B: Biological Sciences 273(1593):1545-1550. doi:

10.1098/rspb.2006.3479

Quinn, G. P., Keough, M. J., 2002. Experimental design and data analysis for biologists.

Cambridge, U.K.: Cambridge University Press.

Richards, J., Morris, K., Burbidge, A., 2008. Bettongia lesueur. The IUCN Red List of

Threatened Species 2008.

Ride, W.D.L.M., G. F., Douglas, A.M., Royce, R.D., Tyndale-Biscoe, C.H., 1962. The

Results of an expedition to Bernier and Dorre Islands, Shark Bay, Western

Australia, in July 1959, Fisheries Department Western Australia, Perth, W.A, pp.

64-69 & 84.

Sander, U., Short, J., Turner, B., 1997. Social Organisation and Warren Use of the

Burrowing Bettong, Bettongia lesueur (Macropodoidea : Potoroidae). Wildlife

Research 24(2):143-157. doi: 10.1071/WR96021

196

Short, J., Turner, B., 1993. The distribution and abundance of the burrowing bettong

(Marsupialia : Macropoidea). Wildlife Research 20(4):525-533. doi:

10.1071/WR9930525

Short, J., Turner, B., 1999. Ecology of burrowing bettongs, Bettongia lesueur

(Marsupialia: Potoroidae), on Dorre and Bernier Islands, Western Australia.

Wildlife Research 26(5):651-669. doi: 10.1071/WR98039

Short, J., Turner, B., 2000. Reintroduction of the burrowing bettong Bettongia lesueur

(Marsupialia: Potoroidae) to mainland Australia. Biological Conservation

96(2):185-196. doi: 10.1016/S0006-3207(00)00067-7

Shortridge, G., 1910. An Account of the Geographical Distribution of the Marsupials

and Monotremes of South‐West Australia, having special reference to the

specimens collected during the Balston Expedition of 1904–1907, Proceedings

of the Zoological Society of London, Blackwell Publishing Ltd., pp. 803-848.

Steindler, L. A., Blumstein, D. T., West, R., Moseby, K. E., Letnic, M., 2018.

Discrimination of introduced predators by ontogenetically naïve prey scales

with duration of shared evolutionary history. Animal Behaviour 137:133-139.

doi: 10.1016/j.anbehav.2018.01.013

Turner, A. M., Turner, S. E., Lappi, H. M., 2006. Learning, memory and predator

avoidance by freshwater snails: effects of experience on predator recognition

and defensive strategy. Animal Behaviour 72(6):1443-1450. doi:

10.1016/j.anbehav.2006.05.010

Turney, P, D. Whitley, Anderson, R. W., 1996. Evolution, learning and instinct: 100

years of the Baldwin effect. Evolutionary Computation 4(3):iv-viii

197

van Damme, R., Castilla, A.M., 1996. Chemosensory predator recognition in the

lizardPodarcis hispanica: Effects of predation pressure relaxation. Journal of

Chemical Ecology 22(1):13-22. doi: 10.1007/BF02040196

Veen, T., Richardson, D.S., Blaakmeer, K., Komdeur, J., 2000. Experimental evidence for

innate predator recognition in the Seychelles warbler. Proceedings of the Royal

Society of London B: Biological Sciences 267(1459):2253-2258. doi:

10.1098/rspb.2000.1276

Webb, J.K., Brown, G.P., Child, T., Greenlees, M.J., Phillips, B.L., Shine, R., 2008. A

native dasyurid predator (common planigale, Planigale maculata) rapidly learns

to avoid a toxic invader. Austral Ecology 33(7):821-829. doi: 10.1111/j.1442-

9993.2008.01847.x

West, R., Letnic, M. , Blumstein, D. T., Moseby, K. E., 2018. Predator exposure improves

anti‐predator responses in a threatened mammal. Journal of Applied Ecology

55(1):147-156. doi: 10.1111/1365-2664.12947

Woinarski, J.C., Burbidge, A.A., Harrison, P.L., 2015. Ongoing unraveling of a

continental fauna: decline and extinction of Australian mammals since

European settlement. Proceedings of the National Academy of Sciences

112(15):4531-4540. doi: 10.1073/pnas.1417301112

Wynne, C. D. L., McLean, I. G., 1999. The comparative psychology of marsupials.

Australian Journal of Psychology, 51(2): 111-116.

Yorzinski, Jessica L., and Michael L. Platt. 2012. The difference between night and day:

antipredator behavior in birds. Journal of ethology 30(2):211-218.doi:

10.1007/s10164-011-0318-5

198

4.8 SUPPLEMENTARY MATERIAL

Table S4.1 Spatial autocorrelation of sample sites in the residuals of the fitted values for analysed behaviours within the “no predator” exclosure and “cat enclosure”.

Exclosure Type Behaviour Moran’s (i) Z p-value

No Predator Wary Approach 0.206 0.990 0.321

Forage 0.075 0.409 0.682

Retreat 0.015 0.154 0.877

Fast Approach 0.160 1.112 0.266

Investigate 0.186 0.803 0.422

Cat Enclosure Wary Approach 0.254 0.949 0.342

Forage 0.046 0.079 0.937

Retreat 0.018 0.016 0.986

Fast Approach 0.101 0.461 0.645

Investigate 0.300 1.000 0.316

199

Behaviour groups Mean difference SEM p-value

Wary approach

Cat vs. rabbit 0.083 0.132 0.999

Cat vs. bucket 0.127 0.113 0.999

Cat vs. control 0.270 0.108 0.074

Rabbit vs. bucket 0.210 0.128 0.614

Rabbit vs. control 0.353 0.131 0.043**

Bucket vs. control 0.143 0.100 0.909

Retreat

Cat vs. rabbit 0.008 0.008 0.999

Cat vs. bucket 0.012 0.013 0.999

Cat vs. control 0.024 0.020 0.999

Rabbit vs. bucket 0.019 0.009 0.252

Rabbit vs. control 0.032 0.018 0.458

Bucket vs. control 0.013 0.020 0.999

Values in italic* indicate significant differences

200

Behaviour groups Mean difference SEM p-value

Wary approach

Cat vs. rabbit 0.337 0.074 ≤0.005**

Cat vs. bucket 0.366 0.066 ≤0.005**

Cat vs. control 0.297 0.101 0.020**

Rabbit vs. bucket 0.030 0.074 0.999

Rabbit vs. control 0.040 0.095 0.999

Bucket vs. control 0.069 0.099 0.999

Foraging

Cat vs. rabbit 0.273 0.129 0.208

Cat vs. bucket 0.194 0.121 0.647

Cat vs. control 0.405 0.113 ≤0.005**

Rabbit vs. bucket 0.079 0.151 0.999

Rabbit vs. control 0.132 0.162 0.999

Bucket vs. control 0.211 0.136 0.727

Values in italic** indicate significant differences

201

Table S4.4 Repeated measures results from generalized estimating equations model

testing for differences between model types (cat, rabbit, bucket and control – no

visual) on the mean proportion of time spent on each behaviour by burrowing

bettongs (Bettongia lesueur) in the “no predator” and “cat enclosure”.

Exclosure Type Behaviour N Min Max Mean Std. Dev

No Predator

Dependent Variable Wary Approach 61 ≤0.005 0.999 0.458 0.358

Covariate Year 61 2015 2016 2015.52 0.504

ID 61 1 60 29.48 18.846

Dependent Variable Forage 61 ≤0.005 0.999 0.302 0.351

Covariate Year 61 2015 2016 2015.52 0.504

ID 61 1 60 29.48 18.846

Dependent Variable Retreat 61 ≤0.005 0.224 0.027 0.042

Covariate Year 61 2015 2016 2015.52 0.504

ID 61 1 60 29.48 18.846

Dependent Variable Fast Approach 61 ≤0.005 0.898 0.036 0.121

Covariate Year 61 2015 2016 2015.52 0.504

ID 61 1 60 29.48 18.846

Dependent Variable Investigate 47 ≤0.005 0.891 0.121 0.202

Covariate Year 47 2015 2016 2015.49 0.505

ID 47 1 58 29.57 18.342

202

Cat Enclosure

Dependent Variable Wary Approach 44 ≤0.005 0.977 0.361 0.269

Covariate Year 44 2015 2016 2015.48 0.505

ID 44 61 120 86.64 15.84

Dependent Variable Forage 44 ≤0.005 0.999 0.387 0.353

Covariate Year 44 2015 2016 2015.48 0.505

ID 44 61 120 86.64 15.842

Dependent Variable Retreat 44 ≤0.005 0.214 0.033 0.054

Covariate Year 44 2015 2016 2015.48 0.505

ID 44 61 120 86.64 15.842

Dependent Variable Fast Approach 44 ≤0.005 0.843 0.085 0.213

Covariate Year 44 2015 2016 2015.48 0.505

ID 44 61 120 86.64 15.842

Dependent Variable Investigate 31 ≤0.005 0.904 0.135 0.239

Covariate Year 31 2015 2016 2015.48 0.508

ID 31 61 120 86.10 16.809

203

Table S4.5 List of model types (bucket, cat, rabbit and control – no visual) deployed in the “no predator” enclosure across sites (ID) per year (2015 and 2016).

Year Expansion ID Treatment ID Treatment

2015 No Predator 1 Control 12 Cat

2016 No Predator 1 Bucket 12 Control

2015 No Predator 2 Bucket 13 Rabbit

2016 No Predator 2 Cat 13 Control

2015 No Predator 3 Cat 14 Bucket

2016 No Predator 3 Cat 14 Rabbit

2015 No Predator 4 Rabbit 15 Rabbit

2016 No Predator 4 Control 15 Cat

2015 No Predator 5 Rabbit 16 Control

2016 No Predator 5 Cat 16 Control

2015 No Predator 6 Rabbit 17 Bucket

2016 No Predator 6 Rabbit 17 Control

2015 No Predator 7 Cat 18 Bucket

2016 No Predator 7 Rabbit 18 Bucket

2015 No Predator 8 Bucket 19 Control

2016 No Predator 8 Control 19 Bucket

2015 No Predator 9 Control 20 Cat

2016 No Predator 9 Rabbit 20 Bucket

2015 No Predator 10 Control 21 Rabbit

2016 No Predator 10 Bucket 21 Cat

2015 No Predator 11 Bucket 22 Cat

2016 No Predator 11 Bucket 22 Cat

204

Year Expansion ID Treatment ID Treatment

2015 No Predator 23 Cat 35 Rabbit

2016 No Predator 23 Rabbit 35 Cat

2015 No Predator 24 Control 36 Rabbit

2016 No Predator 24 Rabbit 36 Control

2015 No Predator 25 Cat 37 Bucket

2016 No Predator 25 Rabbit 37 Cat

2015 No Predator 26 Cat 38 Control

2016 No Predator 26 Control 38 Cat

2015 No Predator 27 Bucket 39 Control

2016 No Predator 27 Bucket 39 Bucket

2015 No Predator 28 Control 40 Rabbit

2016 No Predator 28 Rabbit 40 Control

2015 No Predator 29 Control 41 Bucket

2016 No Predator 29 Control 41 Cat

2015 No Predator 30 Rabbit 42 Cat

2016 No Predator 30 Rabbit 42 Rabbit

2015 No Predator 31 Bucket 43 Cat

2016 No Predator 31 Cat 43 Rabbit

2015 No Predator 32 Bucket 44 Cat

2016 No Predator 32 Bucket 44 Control

2015 No Predator 33 Cat 45 Bucket

2016 No Predator 33 Bucket 45 Bucket

2015 No Predator 34 Control 46 Rabbit

2016 No Predator 34 Cat 46 Bucket

205

Year Expansion ID Treatment ID Treatment

2015 No Predator 47 Control 54 Bucket

2016 No Predator 47 Control 54 Cat

2015 No Predator 48 Rabbit 55 Bucket

2016 No Predator 48 Rabbit 55 Control

2015 No Predator 49 Control 56 Bucket

2016 No Predator 49 Bucket 56 Rabbit

2015 No Predator 50 Control 57 Control

2016 No Predator 50 Bucket 57 Cat

2015 No Predator 51 Cat 58 Rabbit

2016 No Predator 51 Control 58 Rabbit

2015 No Predator 52 Rabbit 59 Cat

2016 No Predator 52 Cat 59 Rabbit

2015 No Predator 53 Cat 60 Rabbit

2016 No Predator 53 Bucket 60 Control

206

Table S4.6 List of model types (bucket, cat, rabbit and control – no visual) deployed in the “Cat enclosure” across sites (ID) per year (2015 and 2016).

Year Expansion ID Treatment ID Treatment

2015 Cat Enclosure 61 Cat 72 Cat

2016 Cat Enclosure 61 Bucket 72 Bucket

2015 Cat Enclosure 62 Bucket 73 Control

2016 Cat Enclosure 62 Bucket 73 Bucket

2015 Cat Enclosure 63 Control 74 Rabbit

2016 Cat Enclosure 63 Cat 74 Control

2015 Cat Enclosure 64 Cat 75 Cat

2016 Cat Enclosure 64 Control 75 Bucket

2015 Cat Enclosure 65 Control 76 Control

2016 Cat Enclosure 65 Rabbit 76 Rabbit

2015 Cat Enclosure 66 Rabbit 77 Bucket

2016 Cat Enclosure 66 Cat 77 Rabbit

2015 Cat Enclosure 67 Bucket 78 Control

2016 Cat Enclosure 67 Rabbit 78 Rabbit

2015 Cat Enclosure 68 Control 79 Bucket

2016 Cat Enclosure 68 Cat 79 Bucket

2015 Cat Enclosure 69 Rabbit 80 Cat

2016 Cat Enclosure 69 Control 80 Cat

2015 Cat Enclosure 70 Rabbit 81 Bucket

2016 Cat Enclosure 70 Rabbit 81 Control

2015 Cat Enclosure 71 Bucket 82 Cat

2016 Cat Enclosure 71 Control 82 Control

207

Year Expansion ID Treatment ID Treatment

2015 Cat Enclosure 83 Rabbit 95 Cat

2016 Cat Enclosure 83 Cat 95 Rabbit

2015 Cat Enclosure 84 Rabbit 96 Cat

2016 Cat Enclosure 84 Cat 96 Rabbit

2015 Cat Enclosure 85 Bucket 97 Rabbit

2016 Cat Enclosure 85 Control 97 Cat

2015 Cat Enclosure 86 Bucket 98 Bucket

2016 Cat Enclosure 86 Control 98 Bucket

2015 Cat Enclosure 87 Control 99 Control

2016 Cat Enclosure 87 Cat 99 Bucket

2015 Cat Enclosure 88 Control 100 Rabbit

2016 Cat Enclosure 88 Cat 100 Cat

2015 Cat Enclosure 89 Bucket 101 Bucket

2016 Cat Enclosure 89 Bucket 101 Rabbit

2015 Cat Enclosure 90 Rabbit 102 Cat

2016 Cat Enclosure 90 Bucket 102 Control

2015 Cat Enclosure 91 Control 103 Cat

2016 Cat Enclosure 91 Cat 103 Rabbit

2015 Cat Enclosure 92 Rabbit 104 Bucket

2016 Cat Enclosure 92 Rabbit 104 Cat

2015 Cat Enclosure 93 Rabbit 105 Control

2016 Cat Enclosure 93 Control 105 Bucket

2015 Cat Enclosure 94 Cat 106 Control

2016 Cat Enclosure 94 Bucket 106 Control

208

Year Expansion ID Treatment ID Treatment

2015 Cat Enclosure 107 Cat 116 Control

2016 Cat Enclosure 107 Rabbit 116 Bucket

2015 Cat Enclosure 108 Rabbit 117 Bucket

2016 Cat Enclosure 108 Control 117 Rabbit

2015 Cat Enclosure 109 Cat 118 Cat

2016 Cat Enclosure 110 Bucket 118 Cat

2015 Cat Enclosure 110 Bucket 119 Control

2016 Cat Enclosure 111 Rabbit 119 Rabbit

2015 Cat Enclosure 112 Control 120 Bucket

2016 Cat Enclosure 112 Cat 120 Rabbit

2015 Cat Enclosure 113 Rabbit 121 Control

2016 Cat Enclosure 113 Bucket 122 Control

2015 Cat Enclosure 114 Cat 123 Control

2016 Cat Enclosure 115 Rabbit

2015 Cat Enclosure 115 Cat

209

Table S4.7 Ethogram of burrowing bettong (Bettongia lesueur) behaviour.

Behaviour Description a

Bi-pedal sniff Animal looks and/or sniffs air whilst standing upright on hind limbs

Prone sniff Animal looks and/or sniffs the air and/or ground whilst standing on all four limbs

Slow approach Animal moves slowly towards feeding station

Fast approach Animal moves quickly and directly towards feeding station

Vigilant foraging Animal chews with its head up and observing surroundings

Relaxed foraging Animal forages for food and chews with its head down without observing surroundings

Investigate model Animal investigates model type through smell and/or touch

Chew model Animal attempts to eat the model type

Recoil Animal recoils away from the feeding station and/or model type

Retreat Animal retreats away from feeding station and/or model type

Out of sight Animal seen on camera and retreated back out of sight away from camera

Fighting Animal fights with con-specifics

Scent marking Animal has its rear dropped to the ground and tail lifted and appears to drag cloaca along the ground scent marking

Grooming Animals grooms itself using grooming claw on hind limb and/or mouth Other Animal displays other behaviour not listed within ethogram a Definition of postures associated with particular behaviours

210

CHAPTER 5:

Not so naïve: Endangered mammal responds to olfactory cues

of an introduced predator after less than 150 years of

co-existence

211

5. 1 ABSTRACT

The inability to recognise and/or express effective antipredator behaviour against novel predators as a result of ontogenetic and/or evolutionary isolation is known as

‘prey naiveté’. However, natural selection favours prey species that are able to successfully detect, identify and appropriately respond to predators prior to their attack, increasing their probability of escape and/or avoidance of a predator. For many prey species, learning and experience is necessary to properly develop and perform appropriate antipredator behaviours. In this study we investigate the recognition of predator scents (faeces) by a remnant population of greater bilbies (Macrotis lagotis) that are coexisting with dingoes (Canis lupus) and feral cats (Felis catus) in south-west

Queensland. In particular, we were interested to evaluate whether wild-living bilbies were naïve to the scent of feral cats. Bilbies in Queensland have shared more than

3,000 years of co-evolutionary history with dog/dingoes, less than 140 years with feral cats and less than 130 years with European rabbits (Oryctolagus cuniculus). Wild-living bilbies would be at selective advantage if able to successfully detect, identify and respond to both predators with which they co-exist. Bilbies spent the greatest proportion of time investigating and the least amount of time digging when cat and dog faeces were present. Our results show that wild-living bilbies displayed antipredator recognition toward the olfactory cues of both a long term predator (dingos) and an evolutionary novel predator (cats). Our findings suggest that native species can develop anti-predator recognition towards introduced predators, providing support for the idea that predator naïveté can be overcome through learning and natural selection as a result of exposure to introduced predators.

212

Keywords: Anti-predator behaviour, Evolutionary history, greater bilby, Prey naiveté hypothesis, Ontogenetic experience

5.2 INTRODUCTION

When isolated from predators some mammalian prey species lose their ability to recognise predators. As predator avoidance is costly in terms of reduced time and energy available for foraging, mating and territorial defence (Brown and Chivers,

2005), no longer functional antipredator behaviours may be selected against

(Goldthwaite et al., 1990; Jolly et al., 2018) when prey are isolated from all predators over evolutionary and ontogenetic timescales. The inability to recognise and/or express effective antipredator behaviour against novel predators as a result of this isolation is known as ‘prey naiveté’ and can occur over both ontogenetic and evolutionary time-scales (Goldthwaite et al., 1990; Carthey and Banks, 2014). In environments that have been rapidly altered, such as those where novel predators have been introduced, some prey species become ‘trapped’ by their evolutionary responses (Schlaepfer et al., 2005). A lack of predator recognition to introduced, novel predators is a damaging form of naïveté, as prey are unable to mount effective antipredator responses (Cox and Lima, 2006; Ferrari et al., 2015).

Natural selection favours prey species that are able to successfully detect, identify and appropriately respond to predators prior to their attack, increasing their probability of escape and/or avoidance of a predator (Monclús et al., 2005). However, not all species or even individuals are able to accurately recognise a predator. Anti- predator behaviour may be innate (genetically based), learnt through experience or is a combination of the two (Jolly et al., 2018). In many species of mammals (Owings and

213

Owings, 1979; Fendt, 2006), birds (Göth, 2001) and fish (Berejikian et al., 2003) predator recognition is an innate trait. Despite years, decades or even thousands of years of isolation from predators, some prey species retain predator recognition skills of their ancestral predators (Blumstein et al., 2008; Li et al., 2011; Steindler et al.,

2018). However, for many other prey species, learning and experience is necessary to properly develop and perform appropriate antipredator behaviours (Griffin et al.,

2001; Ferrari et al., 2005). Prey that are able alter their behavioural patterns in accordance with their learnt experiences are expected to be at a selective advantage in response to potential predation threats from introduced predators (Maloney and

McLean, 1995; Kovacs et al., 2012).

The ‘learned recognition’ hypothesis suggests that through lifetime experience with predators, naïve prey may enhance their ability to recognise and respond to predators (Turner et al., 2006; Saul and Jeschke, 2015), but only if they survive an initial encounter (Kelley and Magurran, 2003). The development of learnt antipredator recognition skills towards evolutionary and/or ontogenetically novel predators has been shown in fish (Ferrari et al., 2005; Ferrari, 2014), birds (Maloney and McLean,

1995) and mammals (Griffin et al., 2000; Webb et al., 2008). How long it takes to learn predator recognition of previously novel predators depends on the prey species and how readily adjustable they are to novel interactions (Cox and Lima, 2006). Some studies have found that despite over 200 years of co-existence with introduced predators, some naïve species are yet to evolve the appropriate antipredator risk assessment responses (Hayes et al., 2006; McEvoy et al., 2008). In contrast, other studies have found that 200 years or less is sufficient to learn, develop and select for

214

appropriate predator recognition skills (Maloney and McLean, 1995; Anson and

Dickman, 2013).

The introduction of novel predators has caused significant damage to native prey populations, particularly in areas where prey species may be considered ‘naïve’

(Cox and Lima, 2006) and is believed to be the major contributing factor to failed reintroduction attempts of locally extinct species, particularly in Australia and New

Zealand (Fischer and Lindenmayer, 2000, Salo et al., 2007; Moseby et al., 2011). In a review of global animal relocations it was found that Australia and New Zealand have the highest reintroduction failure rate (56%, n = 25), of these 69 % (n = 11) were a result of releasing prey into areas where novel predators were present (Fischer and

Lindenmayer, 2000). With an increasing reliance on predator free islands and fenced reserves for threatened species recovery programs, we need to develop a better understanding of the role lifetime experience with predators plays on the development of appropriate antipredator responses. Previous studies have suggested that isolation from all predators may prohibit predator driven natural selection processes (Moseby et al., 2016; Jolly et al., 2018), preventing a ‘future beyond the fence’ for threatened species reintroductions.

The bilby (Macrotis lagotis) is an omnivorous, burrowing, medium sized (1.5 –

2.5 kg), nocturnal marsupial that was once widespread in Australia (Moritz et al.,

1997). In the last 150 years, bilbies have undergone a severe range decline, which has been attributed in part to naïveté towards introduced predators, the red fox (Vulpes vulpes) and feral cat (Felis catus) (Burbidge and Woinarski, 2016). A previous study of wild bilbies living within the fenced predator-free Arid Recovery reserve found that bilbies with no ontogenetic exposure to mammalian predators recognized the scent of

215

a native predator, the dingo (Canis lupus), which they have shared over 3,000 years of co-evolutionary history, but did not recognize the scent of a recently introduced predator, the feral cat (Steindler et al., 2018). These findings suggested that bilbies have innate recognition of dingoes, but not feral cats and that prey species abilities to respond to the odours of predators, scales with their period of evolutionary coexistence (Peckarsky and Penton, 1988).

In this study we investigate the recognition of predator scents by a remnant population of greater bilbies that co-exist with dingoes and feral cats in south-west

Queensland. In particular, we were interested to evaluate whether wild-living bilbies were naïve to the scent of feral cats, like the ontogenetically predator naïve population at Arid recovery, or had developed recognition of the scent of feral cats. If the latter, we expected that bilbies should be more wary when both cat and dog faeces are present compared to a herbivore (rabbit, Oryctolagus cuniculus) and procedural control (no odour). Wild-living bilbies would be at selective advantage if able to successfully detect, identify and respond to both predators with which they co-exist.

Predator recognition could be due to either learned recognition of the threat posed by cats through their lifetime or strong natural selection imposed by cats over evolutionary time. However, if bilbies recognized dog faeces but not cat faeces it would suggest that bilbies responses towards predators are constrained by their period of evolutionary co-existence and bilbies remain ‘naïve’ to introduced feral cats.

216

5.3 MATERIALS AND METHODS

5.3.1 Study Area

We studied wild bilbies across 21 nights in October 2016 within Astrebla Downs

National Park, Queensland (Fig. 5.1, -24° 12' 24.60" S, 140° 34' 5.39" E). Astrebla

Downs National Park is located in the channel country, consisting of flat to undulating erosion plains dissected by minor drainage lines (Department of National Parks, 2013).

The region is dominated by barley Mitchell grass (Astrebla pectinate), with other herbs and grass growing during the wetter periods (Department of National Parks, 2013).

The climate is typically arid, with low annual rainfall and high summer temperatures

(Gibson, 2001).

Dingos, cats and rabbits are present within Astrebla Downs National Park and threaten the wild bilby population through predation and competition for resources, such as food and shelter (Department of National Parks, 2013). Feral cats have been recognised as a significant predatory threat to the bilby population at Astrebla Downs

National Park (Department of National Parks, 2013). Predator control is conducted within the National Park, with a particular focus on feral cats. Cat numbers within the park are monitored on an annual basis by Queensland Parks and Wildlife staff, approximately every 6 weeks from April to September, with appropriate control methods implemented as needed (McLaughlin, 2016). The population of cats tend to fluctuate in response to plagues of native long-haired rats (Rattus villosissimus), as was the case in 2009 to 2011, with the primary control method being shooting, augmented by 1080 baiting. Between May 2012 and late 2015, approximately 3000 feral cats were shot within Astrebla Downs National Park (McLaughlin, 2016). Bilbies within Astrebla

217

Downs National Park are also predated by avian and reptilian predators (per personal observation).

(b)

(a)

Figure 5.1 (a) Map of Australia showing the approximate location of Astrebla Downs

National Park. (b) Map of Queensland with the exact location of Astrebla Downs

National Park (1,740 km²). (c) Map of Astrebla Downs National Park. The green circles indicate the locations of the burrows where odour recognition studies on wild greater bilbies (Macrotis lagotis) were conducted.

218

5.3.2 Sources and storage of treatment odours

We used faecal samples as they are a useful indicator of predator presence (Hayes et al., 2006) and provide prey with information regarding predation risk, even when a predator is absent at the time of detection (McEvoy et al., 2008). We used faeces from two placental predators, with which wild bilbies have shared varying periods of co- evolutionary history. Bilbies in Queensland have shared more than 3,000 years of co- evolutionary history with dog/dingoes (Savolainen et al., 2004) and less than 140 years with feral cats (Abbott, 2002). We also used a procedural control, with whom bilbies have shared less than 130 years of co-evolutionary history (European rabbits) (Zenger et al., 2003). We also had a control which was no faeces present.

We collected fresh faeces from domestic dog, cat and rabbit sources. We used domestic dog faeces as previous studies have showed that faeces from dogs are chemically indistinguishable from those of dingoes (Carthey, 2013). Although it would have been preferential to source dingo faeces and test whether they chemically indistinguishable from dogs, it was not feasible within the timeframe of this study. All faecal samples were collected fresh from private pet owners and boarding kennel facilities, and stored and sealed in airtight zip lock bags, and frozen at minus 20°C. By collecting samples fresh and freezing, it overcame the issue of decomposition of faecal odours after deposition (Carthey, 2013). Disposable gloves were worn at all times when handling faeces to prevent cross contamination of odours. As faecal samples were collected from private pet owners and boarding kennel facilities, the total number and gender of donor individuals is unknown, however it may be estimated to be between two and fifteen individuals, from both sexes. Faeces allocation, based on source location, was randomized amongst the treatment sites throughout the

219

experiment, reducing the chance of potential donor effects. We did not consider diet to be a potential confounding source during analysis, since the diet of domestic cats, dogs and rabbits were consistent between individuals, with cats and dog fed a mix of raw meats and pet foods and rabbits fed pet food (Carthey, 2013).

5.3.3 Bilby Behaviour

As we were unable to track individual bilbies, we conducted a population-level evaluation of bilby behaviour by conducting our experiments adjacent to bilby burrows and treating each burrow individually. As studying wild populations of bilbies can often be problematic due to their cryptic nature (McGregor and Moseby, 2014), placing faecal odour treatments outside the entrances of active burrows was the most effective way to test population level predator odour recognition and behavioural responses. Although seasonally and across years, bilby burrow use is in a constant state of change, bilbies are known to use two to three burrows per night (Lavery and

Kirkpatrick, 1997). Based on population estimates developed by Lavery and Kirkpatrick

(1997), we estimate from the 128 burrows we examined in October 2016, that there was at least 30 bilbies present within our study area.

Bilby burrows are distinct based on their size, shape and characteristic markings and diggings (Fig. 5.2) (Lavery and Kirkpatrick, 1997). As wild bilbies are known to be site selective (Lavery and Kirkpatrick, 1997), we targeted active bilby burrows located within the centre of Astrebla Downs National Park, where bilby presence was known to be highest (pers. comm. with Chris Mitchell, Qld National

Parks). Active burrows were identified by the presence of fresh bilby faeces and/or fresh diggings around the entrances of bilby burrows (Southgate, 1990). Signs of an

220

active burrow may be blown away within 24 hours as a result of wind (Lavery and

Kirkpatrick, 1997).

We used a repeated measures design, in which each faecal odour treatment was presented once at each burrow, according to a predetermined balanced order. We controlled for order effects experimentally and assessed these effects statistically.

Faecal odour treatments were presented on consecutive nights. Faeces were presented on the surface of the ground, within 20 cm of the burrow entrance. In cases where there were multiple burrow entrances present, faeces were only presented at the burrow entrance that showed most recent signs of activity. One piece of cat and dog faeces of similar size and weight (approximately 25 – 30 g) and 20 pellets of rabbit faeces were presented outside the burrow accordingly. Faeces and all faecal traces, including a fine layer of sand on which the faeces were placed, were removed the following day post treatment. Faecal odours were replaced with fresh faecal samples as per the predetermined balanced order for the duration of the experiment per burrow.

At each burrow, a metal post was positioned approximately 1 to 4 m from the burrow entrance (Fig. 5.2). A Bushnell Trophy Cam (Bushnell; America), Scoutguard

SG550V or Scoutguard Zeroglow (Scoutguard; Australia), infrared motion sensor video camera was mounted to the metal post, 20 – 100 cm off the ground. Cameras were programmed to take 60-s video, when triggered, to enable species identification and observe behavioural responses to the odour treatments, with a 0-s interval between possible triggers, from dusk until dawn (1800 – 0600h).

221

Figure 5.2 Experimental set-up for predator odour discrimination study of wild greater bilbies (Macrotis lagotis) at Astrebla Downs National Park. Infrared motion sensor video camera mounted on a metal post outside the burrow entrance of a wild bilby, where odour treatments (cat, dog and rabbit faeces and procedural control – no odour) were presented.

To test whether burrow location influenced bilby behavioural responses to the odour treatments, we tested for spatial autocorrelation in the residuals of the fitted values for each behaviour, using Moran’s index (i), calculated in the spatial analyst module of ArcGis v10.2. Spatial autocorrelation occurs when the value of a variable at any one location in space can be predicted by the values of nearby locations. The existence of spatial autocorrelation indicates that sampling units are not independent from one another (Fortin and Dale, 2005). Analyses showed experimental sites were independent (Table S5.1 in the Supplementary information).

222

5.3.4 Behavioural scoring

(Table 5.1) (Mazzamuto et al., 2018). With the EB approach the researcher defines groups of behaviours, with each group related to specific behavioural responses. We constructed an ethogram of behaviours (Table 5.1) based upon the initial observations of experimental videos. All behaviours were treated as mutually exclusive (Blumstein &

Daniel, 2007). We scored video recordings ≤ 60-s using the event recorder JWatcher

(Blumstein & Daniel, 2007), only quantifying the first 60-s video footage from each burrow location, with scoring commencing at the start of each 60-sec video. We did this because our study focused on quantifying bilbies’ initial behavioural responses to the presence of predator faeces and we wanted to eliminate the potential for our observations to be influenced by habituation to the presence of faeces. We calculated the proportion of time in sight allocated to each behaviour. As multiple bilbies can share a burrow and bilbies were unmarked, we were unable to differentiate between individuals and only scored one video per night, per burrow.

For analysis we combined behaviours in which bilbies were digging outside the burrow entrance and digging within the field of view of the camera, to create a new category ‘digging’ (Table 5.1). We combined behaviours in which bilbies moved slowly: slow approach (slow movement towards odour treatment and/or burrow), slow entrance (individual enters burrow slowly), slow exit (individual exits burrow slowly) and slow retreat (slow movement away from odour treatment and/or burrow) to form a new category ‘walk’ (Table 5.1). We combined behaviours in which bilbies moved

223

rapidly: fast approach (rapid movement towards odour treatment and/or burrow), fast entrance (individual enters burrow quickly), fast exit (individual exits burrow rapidly) and fast retreat (rapid movement away from odour treatment and/or burrow) to create the new category ‘run’ (Table 5.1). In most cases videos were scored blind with respect to treatment, unless it was possible to visually identify the type of faeces that was deployed.

Table 5.1 Ethogram of greater bilby (Macrotis lagotis) behaviour.

Behaviour Description a

Investigate Odour Animal investigates odour treatment (faeces) through smell.

Digging Animal digging outside burrow entrance and/or digging within focal view of camera.

Bi-Pedal Stance Animal looks and/or sniffs air whilst standing upright on hind limbs.

Walk Animal moves slowly towards and/or retreats from odour and/or burrow entrance. Animal moving slowly when entering and/or exiting burrow.

Run Animal moves rapidly towards and/or retreats from odour and/or burrow entrance. Animal moving rapidly when entering and/or exiting burrow.

a Definition of postures associated with particular behaviours

224

5.3.5 Analysis of behavioural data

Mazzamuto et al. (2018) found that Principal Component Analysis (PCA) scores correlated significantly with EB behaviours. To validate our EB approach with a statistical method, we ran a PCA of the five behaviours measured within this study for bilbies. The initial factorial solutions were rotated by the Varimax procedure (Sokal &

Rohlf 1995) (Fig. 5.3).

We fitted a series of linear mixed effects models in SPSS-25 (IBM Corp. Armonk,

NY, U.S.A) with Diagonal error structure to test whether faecal odour treatment caused wild bilbies to allocate different proportions of time to composite behaviours.

Because the response variables were not normally distributed, we log transformed

(log10 [behaviour +1]) each variable prior to analysis (Quinn and Keough 2002).

We had two fixed effects: treatment and presentation order in our models. To account for the possibility of non-independence between observations, we included burrow ID (1 to 128) as a random effect. We did not include the date of experimental trials in the analysis, as it was assumed that new individual bilbies were present at each burrow, based upon spatial autocorrelation tests that showed experimental sites to be independent. As such, bilbies initial response to the order treatments should not in theory decline over time. In no case was presentation order significant, however, we retained it as a blocking factor in the analysis to control for its effect statistically

(Quinn & Keough, 2002). Because we wished to understand the pattern of responses, in instances where the effect of odour was significant (P < 0.05), we used Fisher’s Least

Significant Difference (LSD) post-hoc analysis to examine planned comparisons (Fisher,

1935, Quinn and Keough, 2002) (cat vs. dog, cat vs. rabbit, cat vs. control, dog vs.

225

rabbit, dog vs. control and rabbit vs. control) for differences in response to each odour treatment. We chose not to use Bonferroni post hoc as there is an increased chance of

Type II errors (rejecting an incorrect null hypothesis) due to the substantial reduction in statistical power (Nakagawa, 2004). In the field of behavioral ecology and animal behavior it is difficult to obtain large sample sizes due to practical and ethical reasons.

This results in low sample sizes which than unacceptably increases the probability of

Type II errors if Bonferroni post hoc analysis is applied (Nakagawa, 2004). It is for these reasons that we employed LSD post hoc analysis, rather than Bonferroni.

5.4 RESULTS

The PCA for behavioural measurements of bilbies in response to odour treatments produced two components that together accounted for 52.72% of the variance. The first component of the PCA (component 1, Eigen value = 1.426; % variance explained =

28.514) was positively correlated with variables describing bipedal (rotated PCA score

= 0.626, Fig.5.3) and digging behaviours (rotated PCA score = 0.797, Fig.5.3).Investigate odour behaviour was found to be negatively correlated with component 1 (rotated

PCA score = -0.565, Fig.5.3) The means derived from the factor analysis for component

1 for predator odours (cat and dog) were also negatively correlated with PCA component 1 Fig.5.3). The means derived from the factor analysis for component 1 for the non-predator treatment odours (rabbit and procedural control – no odour) were also found to be positively correlated with component 1 Fig.5.3). Component 2 (Eigen value = 1.210; % variance explained = 24.207) was characterised primarily by a positive correlation of run – fast locomotion (rotated PCA score = 0.626, Fig.5.3) and walk – slow locomotion (rotated PCA score = 0.540, Fig.5.3) behaviour. Based on the PCA it appears that individuals are trading off investigate against bipedal and digging

226

behaviours. Based on the PCA it would appear that bilbies are investing more time in investigate behaviours when cat and dog odours are present, compared to rabbit and the control. There appears to be a good correspondence between the grouping of behaviours from our EB method and that obtained with the PCA.

0.8

0.6 Run Walk 0.4 Rabbit

0.2 Control

-0.2 Digging Cat

Component 2 Component Dog -0.4 Bi-Pedal Stance -0.6

Investigate Odour -0.8 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Component 1

Figure 5.3 Relationship between component scores derived (diamonds) from a principal component analysis of bilby behaviour (PCA) and the mean factor scores

(circles ± SEM) representing the behaviours displayed by bilbies that were filmed at burrows for each of the odour treatments.

There was a significant effect of treatment on the proportion of time that bilbies spent investigating faecal odours (F3, 141.361= 7.073; P ≤ 0.005; Table 5.2, Fig. 5.4a).

Planned post-hoc comparisons (Table S5.2 in the Supplementary information) revealed that bilbies spent more time investigating predator faecal odours compared to the procedural control (no faeces) (Fishers LSD, cat vs. control, P ≤ 0.005 and dog vs.

227

control, P ≤ 0.005; Fig. 5.4a) and herbivore (rabbit) (Fishers LSD, cat vs. rabbit, P =

0.025 and dog vs. rabbit, P ≤ 0.005; Fig. 5.4a). There was no significant difference in the time spent investigating cat and dog faeces (Fishers LSD, cat vs. dog, P = 0.277; Fig.

5.4a) and rabbit faeces and the control (Fishers LSD, rabbit vs. control, P = 0.403; Fig.

5.4a).

There was a significant effect of treatment on the proportion of time that bilbies allocated to digging (F3, 131.405 = 2.715, P = 0.047; Table 5.2, Fig. 5.4b). Bilbies spent less time digging outside the burrow entrance and within the vicinity of the burrow when predator faeces were present compared to the procedural control (no faeces) (Fishers

LSD, cat vs. control, P = 0.038 and dog vs. control, P = 0.008; Fig. 5.4b). There was no difference between the time spent digging when predator faeces were present (Fishers

LSD, cat vs. dog, P = 0.462; Fig. 5.4b). There was no difference in the proportion of time spent digging when cat and rabbit faeces (Fishers LSD, cat vs. rabbit, P = 0.427; Fig.

5.4b), dog and rabbit faeces (Fishers LSD, dog vs. rabbit, P = 0.144; Fig. 5.4b) and rabbit faeces and the control were present (Fishers LSD, rabbit vs. control, P = 0.215; Fig.

5.4b) (Table S5.2 in the Supplementary information).

There was a significant effect of treatment on the proportion of time that bilbies engaged in bi-pedal stance (F3, 108.206 = 4.572; P = 0.005; Table 5.2, Fig. 5.4c). Bilbies spent less time in a bi-pedal stance when faecal odour treatments were present compared to the procedural control (no faeces) (Fishers LSD, cat vs. control, P = 0.018, dog vs. control, P = 0.006 and rabbit vs. control, P = ≤ 0.005; Fig. 5.4c). There was no difference in the proportion of time allocated to bi-pedal stance when predator (cat and dog) and harmless herbivore (rabbit) faeces were present (Fishers LSD, cat vs. dog,

P = 0.604, cat vs. rabbit, P = 0.162 and dog vs. rabbit, P = 0.380; Fig. 5.4c).

228

There was no effect of treatment on the proportion of time that bilbies allocated to walking (F 3, 143.811= 0.694, P = 0.557; Fig. 5.4d) and running (F 3, 110.065 = 0.403, P =

0.751; Table 5.2, Fig. 5.4e). There was no spatial autocorrelation in the residuals of the fitted values for any of the analysed behaviours for bilbies (Table S5.1 in the

Supplementary information). These results indicate that the burrows and treatment sites were independent for the purpose of our analysis.

229

Table 5.2 Results from a series of linear mixed effects models testing for differences between odour treatments (cat, dog, rabbit, and control – no odour) and effect of order, on the mean log proportion of time spent (log10[behaviour +1]) on each behaviour by wild greater bilbies (Macrotis lagotis).

Behaviour df F p-value

Investigate Odour 3 7.073 ≤0.005**

Order (Investigate Odour) 1 0.016 0.899

Digging 3 2.715 0.047**

Order (Digging) 1 0.690 0.408

Bi-Pedal Stance 3 4.572 0.005**

Order (Bi-Pedal Stance) 1 0.030 0.862

Walk 3 0.694 0.557

Order (Walk) 1 2.263 0.135

Run 3 0.403 0.751

Order (Run) 1 1.735 0.190 Values in italic** indicate significant differences

230

Figure 5.4 The mean (± SEMs) proportion of time in sight (PIS) that wild greater bilbies

(Macrotis lagotis) allocated to the behaviours in response to faecal odour treatments

(cat, n = 44, procedural control, n = 35, dog, n = 38 and rabbit n = 36) outside bilby burrows. Similar letters (e.g., a or b) above bars identify pairwise comparisons that are not statistically distinguishable (P > 0.05). No letters indicate there are no significant differences.

231

5.5 DISCUSSION

Our results show that wild-living bilbies displayed anti-predator recognition toward the

olfactory cues of both a long term predator (dingos) and an evolutionary novel

predator (cats). Previous research found that bilbies isolated from all mammalian

predators within predator free reserves respond to the faecal odours of their long

term historical predator, the dingo but not cats (Steindler et al., 2018). These

Proportion of time Proportion (seconds) of time contrasting findings suggest that native species can develop anti-predator recognition

towards introduced predators. However, we cannot be certain whether the

behavioural responses displayed by wild bilbies towards cat odour were the result of

life time learning (Turner, Turner, & Lappi, 2006; Saul & Jeschke, 2015) or selection for

individuals that have learnt and developed appropriate anti-predator recognition and

responses over evolutionary time (Kovacs et al., 2012; Anson & Dickman, 2013).

Bilbies spent the greatest proportion of time investigating and the least amount

of time digging when cat and dog faeces were present. These findings may be due to

bilbies making a trade-off between the costs and benefits of these behaviours (Lima &

Dill, 1990). Recognition of predator odour cues allows prey to perform antipredator

responses that will increase their chances of survival (Chivers et al., 1995). However,

prey animals require information to make these decisions (Bouskila & Blumstein, 1992)

and often exploit the chemosensory cues found in faeces to provide information on

predator activity level and diet (Ferrero et al., 2011). Thus, approaching and

investigating predator cues may allow prey individuals to assess the situation and

modify their behaviour according to the perceived predatory threat (Lima & Dill, 1990;

Magurran, 1990; Cremona et al., 2014). In the case of bilbies, investigating predator

232

faeces may have enabled individuals to assess the likelihood for a potential lethal encounter with a cat and/or dingo within the area.

Bilbies invested the least proportion of time to standing bi-pedal when predator and herbivore faeces were present compared to the control (no odour).

Bilbies typically adopt the upright bi-pedal posture when entering or leaving a burrow, or when foraging (Johnson & Johnson, 1983). That bilbies equally reduced the proportion of time they stood bi-pedal when rabbit and predator odours were present suggests that this behaviour may be a response to the presence of heterospecifics. It may be possible that bilbies adopt this bi-pedal stance in order to gauge more information regarding whether these hetrospecifics pose a threat or not.

Previous studies have suggested that naïve species generalise their response to predators, irrespective of their evolutionary and/or lifetime experience, as a result of the common constituents (kairomones) found in carnivore odours (Ferrero et al.,

2011). Although bilbies responded to both cat and dog odour through increased investigation and decreased digging compared to the control (no odour), we do not believe that this is a result of generalisation. Aversion to all carnivore smells may be costly in terms of missed opportunities, such as foraging and mate selection (Powell &

Banks, 2004). Naïve prey are at a selective advantage if they are able to learn and respond to specific predatory smells, rather than respond to all carnivorous smells

(Blumstein et al., 2002; Powell & Banks, 2004). Barrio et al. (2010) suggested that the common constituent’s hypothesis may only apply when taxonomic levels are closely related. Since cats and dogs diverged between 52 – 57 million years ago (Hedges et al.,

2006), the differences between the two families could be too great for bilbies to generalise the odours. In order to validate this, it would be preferential to chemically

233

assess cat, dog and dingo faeces in order to see whether they are chemically indistinguishable and whether this may influence the level of recognition observed by bilbies within this study. Within the study area, both dingoes and cats are known to predate on bilbies (Paltridge, 2002; Lollback et al., 2015; Burbidge & Woinarski, 2016).

The exposure to cat and dingo predation over evolutionary time and throughout their lifetime is likely to be a greater driver for the predator response behaviour displayed by wild bilbies in this study, rather than a generalised response to predator odours per se.

The results in this study support the idea that ‘naïve’ prey will not remain eternally naïve and have the ability to respond and develop appropriate antipredator responses towards introduced predators (Kiesecker & Blaustein, 1997; Anson &

Dickman, 2013). However, it is unclear whether these predator recognition abilities have become hard-wired, and a heritable trait that will be passed down through the generations, or whether they are experience dependent and individuals will have to learn appropriate predator recognition throughout their lifetime. For example, phenotypic plasticity and learning may provide a valuable short-term response to change, but hinder the potential for long term adaptation (Schlaepfer et al., 2005). As such, in order to successfully manage bilbies and other predator ‘naïve’ species towards introduced predators, we need to better understand the heritability of antipredator behaviours and whether introduced predator recognition abilities are lost and/or gained through lifetime experience.

A further caveat of this study is that it is unclear whether bilbies have developed the appropriate behavioural strategies to avoid predation by predators such as feral cats and dingoes, or are they simply responding to the presence of their

234

odours. Further field tests would need to be conducted that discern recognition, response and survival, so that we can develop a deeper understanding of the behavioural and physical characteristics that enable bilbies to survive the presence of predation pressure. In order to avoid predation by cats and dingoes, bilbies may simply avoid areas where these predators are present, such as abandoning burrows. This behaviour was found in a study of European rabbits (Rouco et al. 2011), in which it was found that wild rabbits responded to the odour of predator faeces (fox) through warren abandonment. This was assessed through the rate of rabbit defecation per day outside warrens that were treated with fox odour and a control (no odour). It was also suggested that some prey animals may have abandoned the treatment warrens due to a greater perceived predation risk, with the effect of intimidation by predators more important than the consumption of prey during predator-prey interactions (Preisser et al. 2005). The perceived predation risk also appeared temporal, with prey animals relocating to abandoned warrens post removal of predator cues, suggesting that response to predator cues was flexible and adaptive (Rouco et al. 2011).

In a study of captive versus wild caught quolls Jolly et al. (2018) found that wild caught quolls, which had ontogenetic and evolutionary experience with predators, recognized and avoided predator scents, while those isolated from predators for only

13 generations showed no recognition or aversion of predators. Taken together, the findings of Jolly et al. 2018, and the contrasting responses to cat odours displayed by wild bilbies reported in this study versus those from a predator-free reserve (Steindler et al. 2018), suggest that predator aversion can be lost when prey populations are isolated from predators. Loss of anti-predator behaviour in predator-free environments has implications for the management of threatened species. In regions

235

where invasive predators pose a threat to native species one commonly used strategy to mitigate predators’ impacts is to create predator-free refuges. This is often achieved by establishing populations of native species on predator-free islands or within predator free fenced reserves (Legge et al., 2018). However, completely isolating populations from predators runs the risk of creating predator naïve populations, which may lack the anti-predator recognition and responses necessary for reintroductions into environments with predators (Moseby et al., 2016). Our findings, which suggest that native species can develop anti-predator recognition towards introduced predators, provides support for the idea that predator naïveté can be overcome through learning and natural selection resulting from exposure to introduced predators (Anson & Dickman, 2013; Moseby et al., 2016; Saxon-Mills et al., 2018).

However, the challenge with such approaches will be providing the conditions necessary for anti-predator skills to be retained and/or developed and ensuring that prey populations are not driven extinct by predation (West et al., 2017; Saxon-Mills et al., 2018)

236

5.6 ACKNOWLEDGMENTS

We thank private pet owners (Edwina Ring), Calabash Kennels & Cattery (Iris Bleach) and the Sydney Cat Protection Society for supply of odours and Sam Fischer, Mike

Letnic and Chris Mitchell (Queensland National Park staff) for their assistance with the study.

5.7 REFERENCES

Abbott, I. 2002. Origin and spread of the cat, Felis catus, on mainland Australia, with a

discussion of the magnitude of its early impact on native fauna. Wildlife

Research 29(1):51-74. doi: 10.1071/WR01011

Anson, J. R., and C. R. Dickman. 2013. Behavioral responses of native prey to disparate

predators: naiveté and predator recognition. Oecologia 171(2):367-377. doi:

10.1007/s00442-012-2424-7

Barrio, I. C., C. G. Bueno, P. B. Banks, and F. S. Tortosa. 2010. Prey naiveté in an

introduced prey species: the wild rabbit in Australia. Behavioral Ecology

21(5):986-991. doi: 10.1093/beheco/arq103

Berejikian, B. A., E. Tezak, and A. L. LaRae. 2003. Innate and enhanced predator

recognition in hatchery-reared chinook salmon. Environmental Biology of

Fishes 67(3):241-251. doi: 10.1023/A:1025887015436

Blumstein, D. T., L. Barrow, and M. Luterra. 2008. Olfactory Predator Discrimination in

Yellow-Bellied Marmots. Ethology 114(11):1135-1143. doi: 10.1111/j.1439-

0310.2008.01563.x

237

Blumstein, D. T., M. Mari, J. C. Daniel, J. G. Ardron, A. S. Griffin, and C. S. Evans. 2002.

Olfactory predator recognition: wallabies may have to learn to be wary. Animal

Conservation 5(2):87-93. doi: 10.1017/s1367943002002123

Bouskila, A., and D. T. Blumstein. 1992. Rules of thumb for predation hazard

assessment: predictions from a dynamic model. American Naturalist

139(1):161-176. doi: 10.1086/285318

Brown, G. E., and D. P. Chivers. 2005. Learning as an adaptive response to predation.

In: P. Barbosa and I. Castellanos, editors, Ecology of predator/prey interactions.

Oxford University Press, Oxford. p. 34-54.

Burbidge, A. A., and J. Woinarski. 2016. Macrotis lagotis. The IUCN Red List of

Threatened Species 2016. http://dx.doi.org/10.2305/IUCN.UK.2016-

2.RLTS.T12650A21967189.en (Accessed 11 July 2018).

Carthey, A. J., and P. B. Banks. 2014. Naïveté in novel ecological interactions: lessons

from theory and experimental evidence. Biological Reviews 89(4):932-949. doi:

10.1111/brv.12087

Chivers, D. P., G. E. Brown, and R. J. F. Smith. 1995. Acquired recognition of chemical

stimuli from Pike, Esox lucius, by Brook Sticklebacks, Culaea inconstans

(Osteichthyes, Gasterosteidae). Ethology 99(3):234-242. doi: 10.1111/j.1439-

0310.1995.tb00897.x

Cox, J. G., and S. L. Lima. 2006. Naiveté and an aquatic–terrestrial dichotomy in the

effects of introduced predators. Trends in Ecology & Evolution 21(12):674-680.

doi: 10.1016/j.tree.2006.07.011

238

Cremona, T., M. S. Crowther, and J. K. Webb. 2014. Variation of prey responses to cues

from a mesopredator and an apex predator. Austral ecology 39(7):749-754. doi:

10.1111/aec.12138

Department of National Parks, R., Sport and Racing, Queensland. 2013. Astrebla Downs

National Park Management Statement 2013, Queensland Government.

Fendt, M. 2006. Exposure to urine of canids and felids, but not of herbivores, induces

defensive behavior in laboratory rats. Journal of chemical ecology 32(12):2617.

doi: 10.1007/s10886-006-9186-9