ECOLOGY AND STATUS OF A NEW OF CARNIVOROUS , THE BLACK-TAILED DUSKY (A. ARKTOS) AND ITS RELATIONSHIP WITH A SYMPATRIC CONGENER, THE BROWN

ANTECHINUS (A. STUARTII )

Emma Gray B.App.Sci. (Hons)

Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

School of Earth, Environmental and Biological Sciences Faculty of Science and Engineering Queensland University of Technology 2017

Keywords

Activity patterns, Antechinus, , autecology, breeding biology, conservation, , detection, diet, geographic distribution, GLMM, mark-recapture, population dynamics, rarity, sympatric species, threatened species.

Ecology and status of a new species of carnivorous marsupial, the black-tailed (A. arktos) and its relationship with a sympatric congener, the (A. stuartii ) i

Abstract

Antechinus are small, carnivorous endemic to Australia. Recently, the has been revised, seeing four new species named and one existing subspecies raised to species status. Therefore, the primary aim of the present thesis was to investigate the ecology and conservation status of one of these new species, the black- tailed dusky antechinus (Antechinus arktos), about which almost nothing was known. This included collecting baseline ecological data on breeding biology, diet, diel activity, distribution and relative abundance and exploring alternative detection methods. Museum specimens indicated the species once occurred at a range of sites on the slopes of the eroded Tweed Shield Volcano caldera, which straddles the border of Queensland and . However, recently the species has only been confirmed from two proximate locations at the summit of the caldera at Springbrook National Park, where they occur at low apparent density along with populations of the brown antechinus (Antechinus stuartii). To date, few studies have been conducted on the northern A. stuartii clade that co-occurs with A. arktos and so a further aim of the thesis was to collect concurrent data on this species and comment on the degree of competition / niche overlap between the two species in sympatry.

Fieldwork was carried out between 2014 and 2016 within the Tweed Shield Volcano caldera. First, an intensive mark-recapture study was conducted at two sites (Best of All and Bilborough Court Lookouts) within Springbrook National Park in order to determine the autecology and relative abundance of the two antechinus species. Live trapping was conducted monthly, between April and October, for two years (2014-2015), to monitor the biology of antechinus before, during and immediately after breeding. In total, 103 A. arktos and 2, 125 A. stuartii were captured and released over 16, 630 trap nights. In general, the ecology of A. arktos and A. stuartii was similar and conformed to the established patterns of the genus. Both species were sexually dimorphic for size and exhibited a semelparous life history strategy, with synchronous reproduction and a short breeding period in winter / spring (A. arktos: mid-September; A. stuartii: late August to mid-September) culminating in the death of all males. Variability in trap success and body mass were all strongly associated with this reproductive cycle. It has been suggested that when two species of antechinus co-occur, the larger congener will breed first. However, A. arktos was ii Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

significantly larger than A. stuartii and mated later, giving birth to a maximum of six young during mid-October.

Over the course of the mark-recapture study, scat samples were also collected from both species for diet analyses. A total of 252 scat samples were subsequently examined, 80 from A. arktos and 172 from A. stuartii. Overall, both species were found to consume a broad range of prey, consistent with diet studies on congeners. However, the composition of taken differed significantly between species, suggesting they may be foraging in different areas or at different times. There was no difference in diet composition between sites; however, there was a significant difference in diet composition of both species between years. There was considerably less diversity of prey items in scats of both species in 2014 compared to 2015. Lower rainfall in 2014 may have reduced abundance and diversity of arthropod prey causing both species to supplement their diet with soft-bodied prey items such as earthworms, which are rarely detected in scats. Additionally, in 2015, invertebrate prey availability was assessed via pitfall trapping in autumn (April), winter (August) and spring (October) to permit estimates of prey preference. Comparison of prey in scats with invertebrate captures from pitfall traps showed both species to be dietary generalists, despite exhibiting preference and avoidance of several prey categories.

Finally, in 2016, 11 infrared cameras were deployed at the Best of All Lookout field site and left in position for five consecutive deployments (ranging from 11-16 days in duration) to evaluate the effectiveness of camera traps for detecting and monitoring antechinus. The camera traps were fixed to wooden stakes set ~50 m apart along established live trapping transects and oriented vertically toward a bait container on the ground surface. In total, 8, 273 image and video pairs were recorded over 725 camera trap nights, with 5, 168 detecting fauna from 10 taxonomic groups. A. arktos accounted for 2.1 % off all observations, while A. stuartii accounted for 13.2 %. Date and time stamps on each image showed that each species displayed a crepuscular peak in activity during the same two-hour period following sunset. Generalized linear mixed models (GLMMs) of detection probability also showed that both deployment number and days since deployment were important factors influencing their detection probabilities. Both A. arktos and A. stuartii displayed a strong negative linear relationship between detection probability and days since deployment, indicating either a rapid loss of interest in the bait used or a decline in its attractiveness with time.

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) iii

Thus, while infrared digital camera traps can be used to detect A. arktos and other small at a rate comparable to live trapping, to consistently achieve high detectability baits would need to be replenished every two-to-three days. In such cases, a standard three night live trapping survey may be a more practical detection method than camera trapping.

Overall, considerable niche differentiation was observed between the two sympatric congeners. A. arktos consumed a higher frequency and volume of Diptera larvae than A. stuartii, indicating they likely forage predominantly in topsoil and subsurface leaf litter. In comparison, A. stuartii consumed Coleoptera and adult Lepidoptera in higher frequency and volume and were often observed to climb trees, indicating they likely forage regularly above ground. However, interestingly, in allopatry A. stuartii are usually terrestrial. Although there is no experimental confirmation of competition between the two species, video footage obtained from camera trapping showed A. stuartii fleeing from A. arktos before direct contact on multiple occasions. Plausibly, the significantly larger A. arktos is able to exclude A. stuartii from otherwise preferred habitat via interference competition. There is compelling evidence for such interactions between other sympatric antechinus species. Additionally, as antechinus breeding strongly correlates with peak prey availability, the observed reproductive phase differences between A. arktos and A. stuartii may have arisen from their dietary preferences for terrestrial or arboreal invertebrates that peak in abundance at different times.

Further, the results of the present study suggest that the newly described black- tailed dusky antechinus is both a rare and threatened species. A. arktos occurs only within the Tweed Shield Volcano caldera and targeted surveys confirmed their presence at only two fragmented locations > 950 m in high rainfall cloud forest. This represents one of the most restricted ranges of any mainland Australian . The absence of the species from previous capture sites at lower elevations strongly suggests the geographic range of A. arktos has contracted, most likely due to climate change. Continued climate warming may result in further core habitat loss for the species, while increasingly unpredictable rainfall and more frequent drought events predicted under climate change scenarios may also negatively impact A. arktos through declines or changes in the availability or seasonality of invertebrate prey. It is therefore strongly recommended the species be listed federally as critically endangered and steps be taken

iv Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

to ensure its long-term viability. The detailed ecological information on breeding biology, diet, diel activity and optimal camera and live trapping methods detailed in this thesis will assist in the planning of future conservation initiatives.

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) v

Outcomes of the Research

Journal articles

Gray EL., Baker AM., and Firn J. 2017. Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii). Mammal Research 62: 47-63.

Gray EL., Burwell CJ., and Baker AM. 2016. Benefits of being a generalist when threatened by climate change: the comparative dietary ecology of two semelparous marsupials, including a new endangered species (Antechinus arktos). Australian Journal of 64: 249-261.

Gray EL., Dennis TE., and Baker AM. 2017. Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black- tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. PLoS ONE 12 (8): e0181592.

Conservation recommendations

Antechinus arktos is currently listed as Endangered at both Queensland and New South Wales state level (Nature Conservation Act 1992; Threatened Species Conservation Act 1995). In Chapter 5, new and more detailed ecological data collected during the present study was evaluated against Australian national legislation (Environment Protection and Biodiversity Conservation Act) and International Union for Conservation of Nature (IUCN) criteria. The species was found to be eligible for listing as Critically Endangered under Criterion 2. The result of this evaluation is currently being used to inform an Australian federal threatened species listing for A. arktos. If the present recommendation of a Critically Endangered listing is upheld, the species state listings should also be reassessed.

A recovery plan should be developed for A. arktos that focuses on:

1. Establishing an integrated monitoring program

2. Developing models to predict suitable habitat for A. arktos under present and future climates

vi Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

3. Trialling a range of detection techniques (including live-trapping, camera trapping, detection dogs and species-specific pheromone lures) for use during presence / absence surveys

4. Investigating the feasibility of establishing a captive breeding program

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) vii

Table of Contents

Keywords ...... i Abstract ...... ii Table of Contents ...... viii List of Figures ...... x List of Tables ...... xiii List of Abbreviations ...... xv Statement of Original Authorship ...... xvi Acknowledgements ...... xvii Chapter 1: General introduction ...... 1 1.1 Rarity and commonness ...... 1 1.2 The genus Antechinus ...... 8 1.3 The present study ...... 14 1.4 Thesis structure ...... 16 1.5 References ...... 18 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) ...... 29 2.1 Abstract ...... 29 2.2 Introduction ...... 30 2.3 Materials and methods ...... 31 2.4 Results ...... 36 2.5 Discussion ...... 48 2.6 Acknowledgements ...... 57 2.7 Compliance with ethical standards ...... 57 2.8 References ...... 58 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) ...... 67 3.1 Abstract ...... 67 3.2 Introduction ...... 68 3.3 Materials and methods ...... 70 3.4 Results ...... 74 3.5 Discussion ...... 89 3.6 Acknowledgements ...... 94 3.7 References ...... 95 viii Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 103 4.1 Abstract ...... 103 4.2 Introduction ...... 104 4.3 Methods ...... 106 4.4 Results ...... 112 4.5 Discussion ...... 121 4.6 Acknowledgements...... 126 4.7 References ...... 126 Chapter 5: General discussion ...... 135 5.1 Sympatric relationship between A. arktos and A. stuartii ...... 135 5.2 Rarity and conservation status of A. arktos ...... 138 5.3 Overall status of the genus antechinus ...... 153 5.4 References ...... 155 Supplementary material ...... 165

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) ix

List of Figures

Figure 1.1. Structure of the IUCN red list categories (taken from IUCN 2016) ...... 5 Figure 1.2. Bayesian phylogeny of the genus Antechinus showing the grouping of four distinct species lineages (L1-L4) (adapted from Mutton 2016)...... 11 Figure 1.3. Photograph of live A. arktos paratype specimen (a-b); diagnostic features include long guard hairs contributing to its shaggy appearance and orange-brown fur on the upper and lower eyelid, cheek and in front of the ear. (c) Underside hindfoot of the holotype specimen; note the dark pigmentation around the granules and long claws. Photographs (a-b) are Copyright Gary Cranitch (Queensland Museum); photograph (c) is Copyright Harry Hines. Adapted from Baker et al. 2014...... 12 Figure 2.1. Location of the two trapping sites within Springbrook National Park, 100 km south of Brisbane, Queensland’s capital, situated on the east coast of Australia ...... 32 Figure 2.2. Trap success over the 2-year trapping period. a Trap success of A. arktos at Best of All Lookout. b Trap success of A. arktos at Bilborough Lookout. c Trap success of A. stuartii at Best of All Lookout. d Trap success of A. stuartii at Bilborough Lookout ...... 39 Figure 2.3. Minimum number of A. stuartii known to be alive (KTBA) at Best of All and Bilborough ...... 40 Figure 2.4. Minimum number of A. arktos known to be alive (KTBA) at Best of All and Bilborough ...... 42 Figure 2.5. Body mass of A. arktos and A. stuartii between April and October (2014 and 2015) at both sites ...... 46 Figure 2.6. Total rainfall in each month from 2013 to 2015 at Springbrook National Park, compared to the average mean rainfall per month for the last 34 years (Bureau of Meteorology 2015) ...... 54 Figure 3.1. Shade plot showing the relative contribution (by volume) of each prey category identified in scats of A. arktos and A. stuartii on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent)...... 78 Figure 3.2. Shade plot denoting the relative contribution (by volume) of each prey category identified in scats of A. arktos and A. stuartii males and females on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent)...... 79 Figure 3.3. Shade plot denoting the relative contribution (by volume) of each prey category identified in scats of A. arktos between years of study

x Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

(2014 and 2015) on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent)...... 81 Figure 3.4. Shade plot denoting the relative contribution (by volume) of each prey category identified in scats of A. stuartii between years of study (2014 and 2015) on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent)...... 82 Figure 3.5. Frequency of Occurrence (%) of each prey category identified in scats of (a) A. arktos and (b) A. stuartii at Springbrook and Lamington National Parks...... 84 Figure 3.6. Frequency of Occurrence (%) of each prey category identified in A. arktos (AA) and A. stuartii (AS) scats and in pitfalls (PF) during A) autumn, B) winter, and C) Spring...... 88 Figure 4.1. Schematic image of the vertical camera set-up showing camera positioning, with essential components and distances labelled ...... 109 Figure 4.2. Model estimates of the effect of camera deployment period on detection probabilities of our study’s four small-mammal target species: A. arktos (AA), A. stuartii (AS), M. cervinipes (MC) and R. fuscipes (RF). Vertical bars represent 95% confidence intervals...... 116 Figure 4.3. Relationships between detection probabilities estimated by GLMM and days since camera deployment (averaged across all five deployments) for our four target species: A. arktos (AA), A. stuartii (AS), M. cervinipes (MC) and R. fuscipes (RF). Bars represent 95% confidence intervals...... 117 Figure 4.4. Cumulative detection probability curves calculated for A. arktos (AA), A. stuartii (AS), M. cervinipes (MC) and R. fuscipes (RF) from GLMM minimal adequate detection models (averaged across all five deployments)...... 118 Figure 4.5. Estimates of the relative daily activity patterns for each pair of sympatric species pooled across all five deployments. On the x axis time is shown in 24-hour time. In each separate plot, the dashed and solid lines represent the kernel density estimates for the indicated species. The degree of activity overlap between the two species is the area under the minimum of the two density estimates, as indicated by the shaded area in each plot. The estimate of overlap and confidence intervals are given in Table 4.4 ...... 120 Figure 5.1. Distribution map of A. arktos indicating all museum specimen locations (dots) and all present day populations (asterisks). Adapted from Baker et al. (2014)...... 138 Figure 5.2. Criterion one of EPBC Act, population size reduction. Measured over the longer of 10 years or 3 generations based on any of A1 to A4. ... 148 Figure 5.3. Criterion two of EPBC Act. Geographic distribution as indicators for either extent of occurrence and/or area of occupancy...... 149 Figure 5.4. Criterion three of EPBC Act. Population size and decline...... 150

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) xi

Figure 5.5. Criterion four of EPBC Act. Number of mature individuals...... 150 Figure 5.6. Criterion Five of EPBC Act. Quantitative analyses...... 151

Supplementary Material

Supplementary Figure 4.1. Example of the type of images obtained, and the ease of identifying diagnostic features (e.g., body size, body and head shape, and ear morphology) from vertically oriented cameras. Species include (clockwise from top left); A) A. stuartii, B) A. arktos, C) R. fuscipes, and D) M. cervinipes. Note antechinus are smaller in size and have more pronounced pointed snouts compared to the Muridae. A. arktos is larger, with a more rounded rump than A. stuartii; while, M. cervinipes has a shorter face than R. fuscipes and is smaller in size. R. fuscipes also has distinctive large rounded ears and coarser looking fur compared to M. cervinipes…………………………...... 166

Supplementary Figure 5.1. Extent of occurrence (EOO) and area of occupancy (AOO) of A. arktos (all records). Provided by Department of the Environment and Energy (2017)…………………………………...170

Supplementary Figure 5.2 Extent of occurrence (EOO) and area of occupancy (AOO) of A. arktos 1997-2017. Provided by Department of the Environment and Energy (2017)………………………………………………………………171

xii Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

List of Tables

Table 1.1. Seven forms of rarity model (adapted from Rabinowitz 1981; Rabinowitz et al. 1986) ...... 3 Table 2.1 Summary of antechinus and non-target species trapping data gathered at Best of All and Bilborough between 2014 and 2015 ...... 37 Table 2.2. Model extracted mean weights of antechinus subdivided by year, site, species, sex and age class ...... 44 Table 2.3. Results of model selection for A. stuartii ...... 45 Table 2.4. Results of model selection for A. arktos ...... 47 Table 2.5. Distribution, trap success, body weight and reproductive timing and synchrony in all 15 known species of antechinus ...... 49 Table 3.1. Overall dietary composition of A. arktos and A. stuartii at Springbrook National Park between 2014 and 2015 expressed as frequency of occurrence (%) and average volume (%)...... 76 Table 3.2. Results of permutational multivariate analysis of variance (PERMANOVA) and post hoc pair-wise tests comparing the relative volumes of food categories between species, as well as between years, seasons and sexes within species ...... 77 Table 3.3. Exact test of goodness of fit results comparing the occurrence of each prey category in pitfall traps and in the diets of A. arktos and A. stuartii...... 89 Table 4.1. Explanatory variables used for modelling detection probabilities by remote camera traps of four species of small mammals at Springbrook National Park...... 110 Table 4.2. Summary of all camera trap image pairs recorded at Springbrook National Park over five successive deployments during 2016...... 114 Table 4.3. Results of the minimal adequate model explaining detection of A. arktos, A. stuartii, M. cervinipes and R. fuscipes ...... 115 Table 4.4. Estimates of activity pattern overlap (0= no overlap, 1=complete overlap) between four co-occurring small mammal species, with sample size and approximate 95% bootstrap confidence intervals ...... 119

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) xiii

Supplementary Material

Supplementary Table 1.1 IUCN criteria for listing species in a threatened category (taken from IUCN 2016). ……………………...... 165

Supplementary Table 5.1. Geographic location of all A. arktos captures including year, habitat, proximity to type locality, and altitude (m) information…………………………………………………………167

Supplementary Table 5.2 Additional live-trapping field data at Lamington, Border Ranges and Nightcap National Parks, 2014-2016………….169

xiv Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

List of Abbreviations

EPBC Act Environment Protection and Biodiversity Conservation Act

GLMs General Linear Models

GLMMs Generalized Linear Mixed Models

IUCN International Union for Conservation of Nature

KTBA Minimum Number of Known to be Alive

PERMANOVA Permutational Multivariate Analysis of Variance.

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) xv Statement of Original Authorship

The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

Signature: QUT Verified Signature

13/10/2017 Date: ______

xvi Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

Acknowledgements

I would firstly like to thank my principal supervisor, Dr. Andrew Baker for introducing me to antechinus and supporting and encouraging me throughout both my honours and PhD studies. His advice and constant enthusiasm made this project possible. My thanks also go to my associate supervisor Dr. Ian Williamson for his valued feedback during the initial stages of this project. I would also like to thank Dr. Ian Gynther for his expert help during the Lamington National Park survey. I will always remember your cry of joy seeing Antechinus arktos for the first time! Many thanks also to Harry Hines for sharing his extensive knowledge (and Elliott traps) with me, without which the mark- recapture project may not have been so successful. I would also like to thank my wonderful collaborators, Dr. Jennifer Firn, Dr. Chris Burwell and Dr. Todd Dennis for their guidance and valued feedback on data chapters. Thanks also to the rangers of Springbrook, Lamington and Border Ranges National Parks who provided accommodation and assistance during this project. I would also like to thank the Queensland Museum for providing workspace for invertebrate identification during the dietary component of the project. I also thank Dr. Aila Keto and the Australian Rainforest Conservation Society for providing ideal accommodation next to my trapping sites at Springbrook National park. I also appreciate the help of the administrative, technical and academic staff of the School of Earth, Environmental and Biological Sciences at QUT, especially Karina Pyle and Scott Allberry. I also thank Australia Geographic, Holsworth Wildlife Research Endowment and Gold Coast City Council for providing critical funding necessary to conduct this research. My special thanks and gratitude also go to the members of the Baker Mammal Ecology Lab: Thomas Mutton, Coral Pearce, and Eugene Mason, for their always capable assistance in the field and friendship during these last three (and a bit) years. Welcome and thanks to new members Morgan Thomas and Caitlin Riordan who will be continuing work on A. arktos in my stead.

My heartfelt thanks go to all those who assisted me in the field, either during my two- year mark recapture study, surveys in other parts of the Tweed Volcano caldera, or the many day trips to collect and set cameras. My thanks go to: Isaac Towers, Paul Finn, Paul O’Callaghan, Kirsten Wallis, Jesse Rowland, Cley North, Scott Matthews, Laura McCallion, Valerie How, Gabrielle O’Kane, Pablo Tapia-Vergara, Kenton Young,

Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii ) xvii

Manuela Cascini, Patrick Green, Ceris Ash, Diana Uang, Leah Gustafson, Matthew Cobb, Rachael Anne Collett, Brenda Keough, Lily Buck and Todd Landers.

My deepest thanks go to my parents, Stephen and Cheryl Gray, who have supported me throughout this journey. They both cleaned Elliott traps with me when they could (and probably should) have been doing anything else. Stephen accompanied me to Springbrook many times, ruining several pairs of shoes in the process, while Cheryl painstakingly read many drafts of this thesis and was always willing to be a sounding board for ideas. Finally, I would like to thank the antechinus (in no particular order): Drogo, Melisandre, Brienne, Arya, Margaery, Ramsay, Tommen, ‘The Mountain’, Jaimie, Samwell, Podrick, Sansa, Cersei, ‘The Hound’, Cedric, Nigel, Bruce, Shrieky, Hermione, Neville, Severus, Ginny, Minerva, Harry, Drako, Sirius and Fleur.

xviii Ecology and status of a new species of carnivorous marsupial, the black-tailed dusky antechinus (A. arktos) and its relationship with a sympatric congener, the brown antechinus (A. stuartii )

Chapter 1: General introduction

1.1 RARITY AND COMMONNESS

1.1.1 What is rarity? Specific definitions and viewpoints on what constitutes rarity abound in the scientific literature. However, generally most regard rare species as in some way delimited on the basis of abundance or geographic range size (Gaston 1994; Kunin and Gaston 1997). Because both abundance and geographic range size are continuous measurements, rarity naturally exists on a spectrum, with extreme rarity and extreme commonness lying at opposite ends, but where rarity ends and commonness begins is left undefined (Kunin and Gaston 1997). Therefore, there are no precise limits to which species are and are not considered rare and in many cases, rarity appears to be assigned as ‘an intuitive concept’ (Usher 1986). Indeed, the term is regularly used by ecologists and conservation biologists alternately as a word to describe low abundance, small range size, or as the converse of common with little justification for its use (Gaston 1994).

Rarity, then, is perhaps more often understood as a relative concept. According to Reveal (1981) “rarity is merely the current status of an extant organism which, by any combination of biological or physical factors, is restricted either in numbers or area to a level that is demonstrably less than the majority of other organisms of comparable taxonomic entities”. Therefore, a species can be described as rare if its distribution and / or abundance is considered lower relative to the distribution and / or abundance of other taxonomically or ecologically similar species (Flather and Sieg 2007). Thus it is quite common to classify rare species within a community using an arbitrary proportion of the assemblage with the lowest abundances, geographic range sizes or both (Flather and Sieg 2007). For example, Gaston (1994) proposed defining as rare that 25 % of species with the lowest abundances and / or smallest geographic ranges. This “quartile definition” of rarity is particularly useful in compiling lists of rare species for conservation planning (Grenyer et al. 2006; Benkendorff and Przeslawski 2008). However, one weakness of this definition is that, over time, a species may move out of the rare category despite maintaining a stable abundance and geographic distribution. This is because changes in the abundance and range size of

Chapter 1: General introduction 1

the other species determine whether it falls above or below the designated cut off (Gaston 1994; Flather and Sieg 2007).

Regardless of which definition of rarity one uses, classification of rare species is conditioned on the geographic scale of interest. For example, a species may be designated rare on a local scale (e.g., a management unit within a national park) but common at a regional or global scale (Gaston 1994). Such scale dependence can greatly influence the number of species classified as rare (e.g., see Butchart and Dunn 2003). Therefore, it is important to acknowledge the particular spatial scale at which a study is performed. An additional caveat related to the classification of rare species is the prevalence of sampling artefacts. Rare species are often cryptic, sparsely or patchily distributed in space or have specialised life history strategies that can reduce detectability (McDonald 2004; Molina and Marcot 2007). If sampling is not carried out in the right place, or applied at the right time, the distribution and abundance of species may therefore be substantially underestimated, leading to a greater number of species being considered rare (Flather and Sieg 2007). There is also the distinct possibility of not detecting a rare species at all (despite it being present), resulting in its exclusion from occupancy and rarity lists within a survey area entirely (Flather and Sieg 2007).

1.1.2 Rare and common species differences The term ‘rare’ may be used to describe a wide variety of distribution and abundance patterns. Rabinowitz (1981; Rabinowitz et al. 1986) categorized species according to three attributes: geographic range (wide or narrow), local abundance (somewhere large or everywhere small / low) and habitat specificity (broad or restricted) to reflect these different forms of rarity and commonness. She argued that habitat specificity should be included in classifications of rarity because habitat specialists may have different ecological properties to uncommon habitat generalists. These attributes are continuous variables, but can be dichotomized to form an eight- celled model (Table 1.1). Seven of the eight cells contain rare species in some sense of the word, while only one is considered common. The bottom right cell contains species rare in all three dimensions, while the remaining cells contain species rare in two dimensions and species rare in one dimension. This ‘seven forms of rarity model’ was adopted by Rabinowitz (1981; Rabinowitz et al. 1986) to categorize species of

2 Chapter 1: General introduction

British and North American flora, but has since been applied to bird species (Kattan 1992), mammals (Yu and Dobson 2000) and primates (Harcourt 2002).

Table 1.1. Seven forms of rarity model (adapted from Rabinowitz 1981; Rabinowitz et al. 1986)

Geographic distribution Wide Narrow Habitat specificity Broad Restricted Broad Restricted Abundance Somewhere high Common Rare Rare Rare Everywhere low Rare Rare Rare Rare

In addition to the above three attributes, rare / common species have also been differentiated based on one or more of the following characteristics (reviewed by Kunin and Gaston 1997): life history strategy (growth rate, litter size, generation time, number of reproductive episodes in a lifetime); dispersal ability; genetic diversity; competitive ability; taxon age (length of time a species exists); trophic group; and body size.

In general, rare species are considered more likely to occur in low density and / or occupy small geographic ranges (Brown 1984; Gaston 1998), exhibit a narrow habitat specificity (Davies et al. 2004; Fattorini et al. 2013), possess slower life histories (e.g., small litter sizes, fewer reproductive episodes; Paine 1990; Olden et al. 2008), use less common resources and / or a narrower range of resources (Hanski et al. 1993; Fattorini et al. 2013), and display poorer dispersal ability (Murray et al. 2002; Reinhardt et al. 2005). However, it is clear that rarity is not expressed in the same way by all organisms (Kunin and Gaston 1997; Murray et al. 2002). Some species are naturally or intrinsically rare due to inherent biological or ecological characteristics (such as those listed above), while others become rare due to extrinsic or anthropogenic activities (such as habitat loss, pollution or exploitation), irrespective of their biology (Partel et al. 2005; Flather and Sieg 2007). In addition, it is often difficult to determine whether differences observed in rare and common species are genuine or an artefact of differences in sample size (Kunin and Gaston 1997).

1.1.3 Rarity and risk of extinction Rare species are not necessarily threatened, nor are all threatened species rare. Threat and threatened are collective words that refer to species that face a significant

Chapter 1: General introduction 3

risk of decline or extinction (Lindenmayer and Burgman 2005). Rare species are potentially more vulnerable to extinction than common ones by virtue of their low abundance and / or small geographic range. For example, low local abundance increases the likelihood that demographic, genetic and / or environmental stochasticity will wipe out populations (Soulé 1987; Lande 1993), while species with small geographic ranges have fewer alternatives for surviving disastrous events (such as habitat destruction or fire, etc.) than more widespread species, which may survive in unaffected areas of their range (Soulé 1987). If a species has both a low abundance and restricted geographic range, all (or most) populations will likely experience adverse conditions simultaneously and they are said to experience a ‘double jeopardy’ (Gaston 1998). However, despite the intrinsic risks associated with rarity, many naturally rare species are able to persist over long periods of evolutionary time (see Simberloff 1998 for examples). Such species are believed to possess characteristics that allow them to become rare and enable them to persist in a rare state (Kunin and Gaston 1997; Johnson 1998). It would therefore be a contradiction to list naturally rare species at risk of extinction if there is no evidence for any existing or future threatening process causing declines in range or abundance. In contrast, a common species may be considered threatened if, for example, its numbers are declining at a rapid rate or predicted to do so due to threatening processes operating in the near future (IUCN version 12 2016). Thus, threatening processes causing decline and not rarity itself is considered a greater contributor to extinction risk (Munton 1987; Mace and Lande 1991; Mace et al. 2008). This view is reflected in the current criteria used to list threatened species on International Union for Conservation of Nature (IUCN) red lists (IUCN 2016).

The IUCN red list of threatened species is the world’s most comprehensive inventory on the conservation status of biological species (IUCN 2016). The categories: critically endangered, endangered, and vulnerable are collectively described as ‘threatened’ (Figure 1.1).

4 Chapter 1: General introduction

Figure 1.1. Structure of the IUCN red list categories (taken from IUCN 2016)

There are five criteria (A-E) used to determine a species’ threatened status (see Supplementary Table 1.1 for a more detailed breakdown of these criteria):

A. Population reduction (past, present or future); B. Geographic range size and decline, fragmentation, or extreme fluctuation; C. Small population size and decline (present or future) or extreme fluctuation; D. Very small or restricted distribution; E. Quantitative analysis of extinction risk.

To qualify for listing in any of the threatened categories, a species need only meet one of the five criteria. Criteria A through D are qualitative, whereas category E is an extinction risk probability that results from undertaking a quantitative analysis, such as a population viability analysis (Mace et al. 2008). Assessors are asked to adopt a precautionary but realistic attitude when applying the criteria so that a taxon is classified as threatened unless it is certain that it is not threatened (IUCN 2016). The category near threatened is intended for species that only just fail to qualify as vulnerable, while the category least concern is used for species that do not meet any of the criteria (Mace et al. 2008). Under this current system, only criterion D allows rare species to be listed as threatened without evidence of recent past, present or future

Chapter 1: General introduction 5

declines. Species with fewer than 50, 250 or 1,000 mature individuals can be listed as critically endangered, endangered or vulnerable, respectively, while species with very restricted geographic distributions (of <20 km2 or ≤ five locations) can be listed as vulnerable, the lowest threat level (Supplementary Table 1.1). The scaling of these values reflects the relationship between the intrinsic risks associated with rarity and extinction time. While criterion D is justified under the precautionary attitude of the IUCN (2016), it remains one of the most inconsistently applied elements of the criteria (Mace et al. 2008).

1.1.4 Era of global extinctions The world is currently facing biodiversity losses of a magnitude described by many as constituting a sixth mass extinction (Barnosky et al. 2011; Pimm et al. 2014; McCallum 2015). The present event is of potentially greater magnitude than the Cretaceous-Paleogene (K-Pg) extinction, which led to losses of more than 70 % of species and exterminated the dinosaurs some 66 mya (McCallum 2015). Indeed, since the year 1500, at least 320 vertebrate species have gone extinct and approximately 20 % of all extant vertebrate species are considered threatened (IUCN 2016b). However, unlike past extinction events, the major processes now driving species extinctions are of anthropogenic origin (Dirzo et al. 2014). Diamond (1989) classically investigated recent extinctions and defined the ‘evil quartet’ of human drivers, which included: 1) overkill, referring to over hunting and direct human exploitation; 2) habitat destruction and fragmentation, including climate change and alteration of disturbance regimes (e.g., fires); 3) impact of introduced species to new ecosystems, and 4) chains of extinction, referring to when the extinction of one species leads to the extinction of another upon which it depends (e.g., predators, scavengers, pollinators). As a result, most species are now subjected to multiple threats, often operating synergistically in ways that amplify their impacts (Carroll 2007; Brook et al. 2008). This has led to patterns of extinction that differ from those observed in the geological past (McCallum 2015). Generally, common species are least likely to go extinct. However, many recently extinct or currently impaired species were once considered common (e.g., passenger pigeon Ectopistes migratorius, greater glider Petauroides volans and saiga antelope Saiga tatarica) (Gaston and Fuller 2007; Lindenmayer et al. 2011).

6 Chapter 1: General introduction

1.1.5 Extinctions in Australia The vertebrates of Australia have also suffered severe declines and extinction rates in modern times (IUCN 2016b). In particular, the decline and extinction of mammals has been greater than that experienced by any other taxonomic group, and has given Australia the unfortunate distinction of having the world’s worst mammal extinction record since European settlement (Woinarski et al. 2015). Since 1788, 30 (9.5 %) of our 315 terrestrial mammal species have become extinct, the most recent casualty being the Bramble Cay melomys, Melomys rubicola, in 2014 (Woinarski et al. 2015). A further 56 terrestrial mammal species (and 33 subspecies) meet IUCN red list criteria for listing as threatened (Woinarski et al. 2015). This includes 11 species and one sub species that are categorised as critically endangered (Woinarski et al. 2014). The two most commonly cited causes of decline and extinction are changes in land use (vegetation clearing, pastoralism, altered fire regimes and grazing by introduced herbivores) and the introduction of exotic predators (predominantly foxes Vulpes vulpes, and Felis catus) (Burbidge and McKenzie 1989; Johnson 2006; Doherty et al. 2017). In other parts of the world, larger mammals are generally more vulnerable to extinction (Cardillo et al. 2005). Although one study of body size and extinction risk in Australian mammals discovered a similar pattern (Cardillo and Bromham 2001), most studies have found that mid-sized mammals of a ‘critical weight range’ (CWR) between 35 and 5,500 g, that live or forage predominantly on the ground surface and / or occur in low rainfall (primarily semi-arid) and less structurally complex habitat are disproportionately vulnerable to extinction (Burbidge and McKenzie 1989; McKenzie et al. 2007; Johnson and Isaac 2009; Chisholm and Taylor 2010; Fisher et al. 2014). This CWR is consistent with the preferred prey size range of feral cats and foxes, while a ground- dwelling habit in more open habitat likely further increases these species’ susceptibility and exposure to predation and land use changes (McKenzie et al. 2007; Johnson and Isaac 2009). Small-bodied species (below and at the lower end of the CWR) and those occurring in regions with relatively high rainfall and / or dominated by forest have therefore previously been considered ‘safe’ (Fisher et al. 2014). However, recent work suggests that many such species are also in decline. For example, Bilney et al. (2010), compared sooty owl Tyto tenebricosa sub-fossil deposits with their contemporary diet and found that small mammal (10 - 1,500 g) decline in forests of south-east Australia had been severely underestimated. Species occurring in elevationally restricted mountain ecosystems are also inherently

Chapter 1: General introduction 7

vulnerable to present and future climate change, with Williams et al. (2014) already documenting significant declines in abundance and range size for several mammal species occurring in the Wet Tropics World Heritage Area of north-east Queensland. Nevertheless, rodents, insectivores, and most small marsupials are severely underrepresented in single-species conservation studies (Amori and Gippoliti 2000). They also tend to attract less conservation action and funding than larger, more ‘charismatic’ species (Wilson et al. 2003; Fisher 2011). In part, the present thesis aims to help remedy this situation by undertaking a detailed ecological study on Antechinus arktos, a newly discovered, yet potentially rare and / or threatened species of small carnivorous marsupial, endemic to the cool temperate cloud forests of south-east Queensland (Baker et al. 2014).

1.2 THE GENUS ANTECHINUS

The genus Antechinus Macleay 1841 belongs within the Family Dasyuridae, a highly speciose group of carnivorous marsupials, which collectively occur in all major habitat types in Australia and New Guinea (Van Dyck and Strahan 2008). The Dasyuridae is currently divided into two subfamilies, each containing two tribes (Westerman et al. 2015):

Family Dasyuridae (76 extant species)

Subfamily:

Tribe: Dasyurini – , , , , etc.

Tribe: Phascogalini – , murexias,

Subfamily:

Tribe: Planigalini –

Tribe Sminthopsini – , , , etc.

1.2.1 General ecology Antechinuses are small (15-180 g) dasyurids endemic to Australia (Van Dyck et al. 2013). They are typically described as generalist , which take a broad range of predominantly invertebrate prey in accordance with their availability (Hall 1980; Statham 182; Dickman 2014). However, some studies have shown that southern A. stuartii and A. agilis may be more selective, predating on specific taxa in different

8 Chapter 1: General introduction

proportions to which they occur in the landscape (Fox and Archer 1984; Yuncken 1997). Unlike most dasyurids, which are arid-adapted, antechinuses occur broadly in temperate and tropical near-coastal regions of the continent where they assume either ground-dwelling or semi-arboreal habits (Van Dyck et al. 2013). They are generally nocturnal (Van Dyck and Strahan 2008). However, A. mimetes and mainland populations of A. minimus have been described as being active throughout the diel cycle (Hall 1980; Sanecki et al. 2006) and diurnal (Sale and Arnould 2009), respectively. Habitat preferences vary between species (and subspecies) from those that prefer closed rainforest / vine forest to those that favour open forest; however, the majority of antechinus species still occur in sympatry with a congener in at least part of their range (Dickman 1986; Watt 1997; McAllan et al. 2006). As sympatric antechinus species utilize similar resources, interspecific competition likely has and continues to influence population dynamics, behaviour and evolution in the genus (Dickman 1986; Crowther and Blacket 2003).

1.2.2 Life history and breeding biology Breeding studies have been a central focus of antechinus research, because like only a few other mammal genera worldwide, males are semelparous (breed only once) (Woolley 1966; McAllan et al. 2006; Fisher et al. 2013). Females are synchronously monoestrous and mating is confined to a short (1-3 week) period, at the end of which all males in the population undergo complete die-off (Woolley 1966; Wood 1970; Lee et al. 1982; Williams and Williams 1982; McAllan et al. 2006). Male death is a result of stress-induced suppression of the immune system, as elevated testosterone ensures failure of the ‘turn off’ switch for the stress hormone cortisol (Bradley 2003). Antechinus are able to breed in the first year of life. Thus, in the wild, male antechinus live for a maximum of 11.5 months, while females may survive and reproduce in two, or very rarely three, years (Woolley 1966; Wood 1970; Naylor et al. 2008).

The timing of reproduction in each species of antechinus differs little from year to year at the same geographic location; however, breeding timing has been found to vary by up to four months between localities within a species (Wood 1970; Lee et al. 1982). In addition, where two (or more) antechinus species occur in sympatry, they have been recorded to breed at different times (Dickman 1986; McAllan et al. 2006), with the larger congener typically breeding first (Dickman 1982). In all species, breeding timing has therefore been at least partially linked to different rates of change

Chapter 1: General introduction 9

of photoperiod (McAllan et al. 2006). It has been hypothesised that cueing of reproductive periods with specific changes in photoperiod may allow antechinus to time lactation and weaning of offspring with spring / summer insect flushes and thus take advantage of prey availability at this critical time (McAllan et al. 2006; Fisher et al. 2013). However, this alone does not adequately explain male semelparity. Fisher et al. (2013) concluded that the ‘suicidal’ reproductive strategy is an adaptive response to the shortened breeding season and synchronous ovulation, coupled with mate promiscuity imposed on males by females as a means of increasing sperm competition. Since response to photoperiod appears to be species-specific, breeding timing may also function as a reproductive barrier and / or method of ecological segregation between sympatric species (Dickman 1986; McAllan et al. 2006). Crowther and Blacket (2003) believe that, although uncommon, such allochronic isolation in sympatry may have resulted in the speciation of at least one antechinus species.

1.2.3 Recent molecular and morphological assessments of the genus by Van Dyck (2002) recognised ten extant antechinus species: , Antechinus minimus; yellow-footed antechinus, A. flavipes; brown antechinus, A. stuartii; dusky antechinus, A. swainsonii; , A. bellus; rusty antechinus, A. adustus; , A. godmani; cinnamon antechinus, A. leo; , A. agilis and , A. subtropicus. However, since 2012, four new species have been discovered and named: the buff-footed antechinus, Antechinus mysticus (Baker et al. 2012) and the silver-headed antechinus, Antechinus argentus (Baker et al. 2013), both formerly considered as A. flavipes; the black-tailed dusky antechinus, Antechinus arktos (Baker et al. 2014) and the Tasman Peninsula dusky antechinus, Antechinus vandycki (Baker et al. 2015), both formerly grouped under A. swainsonii. An additional species, the mainland dusky antechinus, Antechinus mimetes (formally A. swainsonii mimetes) has also been elevated to species status (Baker et al. 2015). This represents a considerable (50 %) increase in known species diversity and creates many gaps in our knowledge of this ostensibly well-studied genus. Collecting fundamental autecological data on such recently discovered and little understood species is particularly urgent considering Australia’s record of mammal extinctions (Woinarski et al. 2015).

10 Chapter 1: General introduction

1.2.4 The black-tailed dusky antechinus, A. arktos One of the newly identified species, Antechinus arktos, is strikingly different both genetically and morphologically to all other antechinus (Baker et al. 2014). A. arktos belongs within the dusky antechinus species complex, which along with A. minimus is one of the oldest antechinus lineages (Figure 1.2). Although A. arktos is most similar to members of this complex, the four species are still deeply genetically divergent (between 7.6-10.6 %) (Mutton 2016). Further, the black-tailed dusky antechinus has distinctive orange-brown fur on the upper and lower eyelid, cheek and in front of the ear, and a fuscous black tail and hindfeet, which distinguish it from all other congeners (Figure 1.3).

*

*

Figure 1.2. Bayesian phylogeny of the genus Antechinus showing the grouping of four distinct species lineages (L1-L4) (taken from Mutton 2016).

Chapter 1: General introduction 11

Figure 1.3. Photograph of live A. arktos paratype specimen (a-b); diagnostic features include long guard hairs contributing to its shaggy appearance and orange-brown fur on the upper and lower eyelid, cheek and in front of the ear. (c) Underside hindfoot of the holotype specimen; note the dark pigmentation around the granules and long claws. Photographs (a-b) are Copyright Gary Cranitch (Queensland Museum); photograph (c) is Copyright Harry Hines. Adapted from Baker et al. 2014.

A. arktos is known from ten museum specimens (then classified as A. swainsonii) collected between 1966 - 1989 from high rainfall areas (at or above 780m) at the rim of the Tweed Shield Volcano caldera, encompassing Springbrook National Park, eastern Lamington National Park and Border Ranges National Park, on the border of

12 Chapter 1: General introduction

Queensland (Qld) and New South Wales (NSW) in mid-eastern Australia (Baker at al. 2014). It was a further 24 years (from 1989 to May 2013) until another specimen was collected, when Baker et al. (2014) caught four female A. arktos at Best of All Lookout and one male A. arktos, 2.5km distant, at Bilborough Lookout. Both sites are located at the summit of Springbrook National Park (average elevation 950 m) in cool subtropical rainforest, where high rainfall (>3000 mm per year) is augmented by fog- drip (Bureau of Meteorology 2017). This is notably higher (170 m) than the previous capture site at Springbrook National Park. Baker et al. (2014) hypothesised that A. arktos may have retracted to the wettest, highest and coolest part of the caldera in response to climate warming. However, they noted that more intensive and targeted surveys would be required to confirm this limited distribution. Additionally, because so few animals have been captured very little is known of A. arktos ecology. Of the 15 individuals captured thus far, females have ranged in size from 44-59 g and males from 60-120 g, signifying a large size range (Baker et al. 2014). Male A. arktos have been captured in early August (1966 and 1974) at Lamington National Park, indicating that male die-off extends at least to the first week of August in this region. At both present day sites at Springbrook, they were found to occur at low density along with abundant sympatric populations of A. stuartii, Rattus fuscipes and Melomys cervinipes.

1.2.5 The brown antechinus, A. stuartii Antechinus stuartii (Macleay 1841) was one of the first antechinus described and is generally considered one of the most well studied species in the genus. However, since description A. stuartii taxonomy has been repeatedly revised and found to consist of a complex of three genetically distinct species: the brown (A. stuartii), agile (A. agilis) and subtropical (A. subtropicus) antechinuses (Dickman et al. 1998; Van Dyck and Crowther 2000; Figure 1.2). Consequently, the range of A. stuartii proper has been substantially reduced over time and the majority of early studies on what was then known as A. stuartii in fact concern other members of the complex (e.g., Woolley 1966; Wood 1970; Dickman 1986). Even more recently, Mutton (2016) identified distinctive northern and southern clades within A. stuartii (and excluding A. agilis and A. subtropicus) by sequencing two mitochondrial and four nuclear genes.

As it is currently known, A. stuartii is widespread in a variety of forest and heathland habitats within its range (Van Dyck and Strahan 2008). The northern clade occurs north of New England National Park in southern Qld, while the southern clade

Chapter 1: General introduction 13

occurs south of New England National Park in northern and southern NSW (Mutton 2016). A. stuartii is predominantly terrestrial; however, when the southern clade occurs in sympatry with A. mimetes the species has been observed to be partially arboreal (Van Dyck and Strahan 2008). A. stuartii captured alongside A. arktos have been genetically confirmed as the northern clade (Mutton 2016). Therefore, it appears that A. stuartii (north and south) each co-occur with a member of the dusky antechinus species complex in part of their range. The principle of competitive exclusion (Gause 1934) states that two different species competing for access to the same limiting resources (e.g., food, space, or breeding sites) cannot stably coexist if other ecological factors are held constant. Therefore, in order to reduce or avoid interspecific competition in sympatry, ecologically similar species often partition resources (Schoener 1974, 1986; Chase and Leibold 2003). This can lead to differential habitat use (Higgs and Fox 1993; Churchfield et al. 1999), dietary shifts (Goodyear 1992; Luo and Fox 1996), and / or temporal shifts e.g., in foraging or mating timing (Harrington and Macdonald 2008) among competing species. Thus, one way of inferring whether competition has or continues to occur within communities is by examining patterns of resource overlap (Morin 2011).

To date, few studies have been conducted on the northern A. stuartii clade (but see Parra Faundes 2014), and none on populations sympatric with the newly described A. arktos. It would therefore be interesting to compare aspects of northern A. stuartii breeding biology, habitat use and diet with A. arktos and compare their relationship with the closely related southern A. stuartii and A. mimetes from NSW.

1.3 THE PRESENT STUDY

The primary aim of this thesis was to collect and analyse fundamental ecological data for a newly described species of mammal, A. arktos, about which almost nothing is known. Baseline ecological data on relative abundance and distribution will be used to determine the species’ conservation status, while additional data on breeding biology (timing, litter size, etc.), diet, diel activity and alternative detection methods will be outlined in order to facilitate management efforts.

During the study, there was also an opportunity to collect concurrent data for the sympatric northern A. stuartii. Therefore, a further aim of the thesis was to qualitatively estimate and comment on the degree of competition / niche overlap

14 Chapter 1: General introduction

between the two species in sympatry and compare their relationship with those of other members of the genus. Specifically, the study addressed the following research questions:

1. What is the breeding timing of A. arktos and A. stuartii in sympatry?

Do the species breed at different times consistent with observations of other sympatric antechinus? Additionally, does the larger congener breed first, as hypothesised by Dickman (1982)?

2. What is the dietary composition of A. arktos and A. stuartii?

Are the species dietary generalists, like the majority of their congeners? Or do the species show preferences for certain prey taxa regardless of their availability in the environment? Further, is there significant dietary overlap between the two species? Or is diet an important niche difference in sympatry?

3. What is the diel activity of A. arktos and A. stuartii?

Are the species predominantly nocturnal, consistent with the majority of their congeners? Additionally, is there significant activity overlap between the two species? Or is temporal activity an important niche difference in sympatry?

4. Can remote infrared digital cameras detect and differentiate between small mammals, including A. arktos and A. stuartii?

Is camera trapping an alternative way forward for detecting and monitoring A. arktos?

5. Is A. arktos rare or common? Moreover, is the species eligible for listing in a threatened category under EPBC Act and / or IUCN criteria?

Does the species show all three forms of rarity as defined by Rabinowitz (1981; Rabinowitz et al. 1986)? Does A. arktos meet any one of the five criteria (A-E) required for listing in a threatened category as per IUCN criteria (2016), and if so, at what level?

Collecting such fundamental ecological data on breeding biology (timing, litter size etc.), diet, diel activity and detection of recently discovered and little understood species is important because threats cannot be identified, nor enlightened conservation decisions made, without such information (Woinarski et al. 2014; Fleming and Bateman 2016). In addition, since diversity in the genus has increased by 50 % such

Chapter 1: General introduction 15

information will be important to test established theories regarding breeding timing and dietary specialization in antechinus and to place the ecology of A. arktos into perspective within the genus as a whole.

1.4 THESIS STRUCTURE

Due to the diverse subject matter of the PhD, each chapter of the thesis has been written as a standalone research journal article. Although diverse in scope, the components share a common goal: to better understand the ecology and conservation status of the black-tailed dusky antechinus, A. arktos. Research topics in each chapter were chosen based on the dearth of data currently available for the species. It was reasoned that a collection of data chapters encapsulating fundamental subject areas would be the most immediate and useful way of increasing our understanding of A. arktos’ ecology and immediate management requirements, and would thus direct focus in providing a range of crucial launching points for future research.

Because each data chapter (2-4) has been either published or submitted for publication in a different journal, there is necessarily some repetition of species and study area descriptions to establish context. To prevent further repetition, detailed background information concerning diet, diel activity and camera trapping methods are introduced specifically in each data chapter and not presented here in the general introduction. Formatting in each data chapter reflects the requirements of each journal; however, figures and tables have been re-numbered sequentially throughout the whole thesis for internal consistency. For all chapters the present author is the lead author, whose role included conceiving and undertaking all of the fieldwork, performing all of the data analyses and writing all of the chapters. The supervisor of the present project, Andrew Baker (QUT), is co-author on all data chapters as he assisted with the initial study design and reviewed all of the chapters. For chapter 2, Jennifer Firn (QUT) is a co-author as she provided important advice and specific assistance during the data analysis process. For chapter 3, Chris Burwell (Qld Museum) is a co-author as he supervised and assisted with invertebrate identification and reviewed the manuscript. For chapter 4, Todd Dennis (University of Auckland) is a co-author as he provided infrared cameras for the study, advice on study design and data analysis and manuscript review. A brief synthesis of each chapter is provided below.

16 Chapter 1: General introduction

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii).

This chapter documents an intensive, two-year, mark-recapture study of A. arktos and A. stuartii that was conducted at Best of All Lookout and Bilborough Lookout within Springbrook National Park. It includes analyses of trap success, population dynamics, body size and breeding biology, in addition to a discussion of within and between species differences and a summary of conservation implications.

The chapter was submitted as a research paper on 29 March 2016, accepted 30 May 2016, and published online 24 June 2016 in Mammal Research.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos).

This chapter presents the results of a dietary analysis on A. arktos and A. stuartii scats that were collected over the course of the two-year mark-recapture study. The dietary composition and preferences of both species between and within years and sites are analysed and then discussed in comparison to each other and the potential impacts of climate change.

The chapter was submitted as a research paper on 28 June 2016, accepted 28 September 2016, and published online 14 November 2016 in Australian Journal of Zoology.

Chapter 4: Can infrared digital cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

This chapter examines the effectiveness of infrared digital camera traps for accurately detecting and monitoring small, elusive mammals such as A. arktos, especially where abundant co-existing small mammal species may confound identification. The factors influencing trap success and detection of A. arktos, A. stuartii, R. fuscipes and M. cervinipes are analysed in order to establish whether camera traps are an adequate replacement for, or addition to, current live-trapping surveys and monitoring methods. The diel activity of each species is also analysed and the activity overlap and behaviour exhibited by species in the community discussed.

Chapter 1: General introduction 17

The chapter was submitted as a research paper on 13 April 2017, accepted 3 July 2017, and published online 9 August 2017 in the journal PLoS ONE.

Chapter 5: General discussion

This chapter synthesises the above data chapters to provide a general discussion of the relationship between sympatric A. arktos and A. stuartii, including how both species fit (or otherwise) the general patterns of the genus. Data collected in the thesis on A. arktos is then synthesised, interpreted and assessed to classify the species according to International Union for Conservation of Nature (IUCN) and Environment Protection and Biodiversity Conservation (EPBC) Act criteria and discuss the long- term viability of the species in light of potential threats. The thesis concludes by outlining areas for future research and discussing the current conservation status of the antechinus genus as a whole.

1.5 REFERENCES

Amori G., and Gippoliti S. 2000. What do mammalogists want to save? Ten years of mammalian conservation biology. Biodiversity and Conservation 9: 785-793.

Baker A., Mutton T., and Hines H. 2013. A new dasyurid marsupial from Kroombit Tops, south-east Queensland, Australia: the Silver-headed Antechinus, Antechinus argentus sp. nov. (Marsupialia: Dasyuridae). Zootaxa 3746: 201-239.

Baker AM., Mutton TY., Hines H., and Van Dyck S. 2014. The Black-tailed Antechinus, Antechinus arktos sp. nov.: a new species of carnivorous marsupial from montane regions of the Tweed Volcano caldera, eastern Australia. Zootaxa 3765:101– 133.

Baker AM., Mutton TY., Mason ED., and Gray EL. 2015. A taxonomic assessment of the Australian Dusky Antechinus Complex: a new species, the Tasman Peninsula Dusky Antechinus (Antechinus vandycki sp. nov.) and an elevation to species, the Mainland Dusky Antechinus (Antechinus mimetes new status). Memoirs of the Queensland Museum 59: 75-126.

Baker A., Mutton T., and Van Dyck S. 2012. A new dasyurid marsupial from eastern Queensland, Australia: the Buff-footed Antechinus, Antechinus mysticus sp.nov. (Marsupialia: Dasyuridae). Zootaxa 3515: 1-37.

18 Chapter 1: General introduction

Barnosky AD., Matzke N., Tomiya S., Wogan GOU., Swartz B., Quental TB., Marshall C., McGuire JL., Lindsey EL., Maguire KC., Mersey B., and Ferrer EA. 2011. Has the earth’s sixth mass extinction already arrived? Nature 471: 51-57.

Benkendorff K., and Przeslawski R. 2008. Multiple measures are necessary to assess rarity in macro-molluscs: a case study from southeastern Australia. Biodiversity Conservation 17: 2455-2478.

Bilney RJ., Cooke R., and White JG. 2010. Underestimated and severe: small mammal decline from the forests of south-eastern Australia since European settlement, as revealed by a top-order predator. Biological Conservation 143: 52-59.

Bradley A. 2003. Stress, hormones and mortality in small carnivorous marsupials. In: Jones M., Dickman CR., and Archer E (eds). Predators with pouches. CSIRO Publishing, Australia. Pp. 254-267.

Brook BW., Sodhi NS., and Bradshaw CJA. 2008. Synergies among extinction drivers under global change. Trends in Ecology and Evolution 23: 453-460.

Brown JH. 1984. On the relationship between abundance and distribution of species. The American Naturalist 124: 255-279.

Burbidge AA., and McKenzie NL. 1989. Patterns in the modern decline of Western Australia’s vertebrate fauna: causes and conservation implications. Biological Conservation 50: 143-198.

Bureau of Meteorology. 2017. Daily rainfall: Springbrook Road. Australian Government.Available:http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccO bsCode=136&p_display_type=dailyDataFile&p_stn_num=040607&p_startYear=

Butchart S., and Dunn E. 2003. Using the IUCN Red List criteria to assess species with declining populations. Conservation Biology 17: 1200-1203.

Cardillo M., and Bromham L. 2001. Body size and risk of extinction in Australian mammals. Conservation Biology 15: 1435-1440.

Cardillo M., Mace GM., Jones KE., Bielby J., Bininda-Emonds ORP., Sechrest W., Orme DL., and Purvis A. 2005. Multiple causes of high extinction risk in large mammal species. Science 309: 1239-1241.

Chapter 1: General introduction 19

Carroll C. 2007. Interacting effects of climate change, landscape conversion, and harvest on carnivore populations at the range margin: marten and lynx in the northern Appalachians. Conservation Biology 21: 1092-1104.

Chase J., and Leibold M. 2003. Ecological niches: linking classical and contemporary approaches. University of Chicago Press, USA.

Chisholm RA., and Taylor R. 2010. Body size and extinction risk in Australian mammals: an information theoretic approach. Austral Ecology 35: 616-623.

Churchfield S., Nesterenko VA., and Shvarts EA. 1999. Food niche overlap and ecological separation among six species of coexisting forest (Insectivora: Soricidae) in the Russian Far East. Journal of Zoology 248: 349-359.

Crowther MS., and Blacket MJ. 2003. Speciation and biogeography within the Dasyuridae, or why are there so many species of dasyurids? In: Jones M, Dickman CR, and Archer E (eds). Predators with pouches: the biology of carnivorous marsupials. CSIRO Publishing, Australia.

Davies KF., Margules CR., and Lawrence JF. 2004. A synergistic effect puts rare, specialised species at greater risk of extinction. Ecology 85: 265-271.

Diamond J. 1989. Overview of recent extinctions. In: Western D., and Pearl C (eds) Conservation for the twenty-first century. Oxford University Press, USA.

Dickman CR. 1982. Some ecological aspects of seasonal breeding in antechinus (Daysuridae Marsupialia). In: Archer M (ed) Carnivorous marsupials Vol 1. Sydney royal zoological society of New South Wales, Sydney, pp. 139–150

Dickman CR. 1986. An experimental study of competition between two species of dasyurid marsupials. Ecological Monographs 56: 221-241.

Dickman CR. 2014. Micro-carnivores: the ecological role of small dasyurid predators in Australia. In: Glen AS., and Dickman CR (eds) Carnivores of Australia: past, present and future. CSIRO Publishing, Australia.

Dickman CR., Parnaby HE., Crowther MS., and King DH. 1998. Antechinus agilis (Marsupialia: Dasyuridae), a new species from the A. stuartii complex in south- eastern Australia. Australian Journal of Zoology 46: 1-26.

20 Chapter 1: General introduction

Dirzo R., Young HS., Galetti M., Caballos G., Isaac NJB., and Collen B. 2014. Defaunation in the Anthropocene. Science 345: 401-406.

Doherty TS., Dickman CR., Johnson CN., Legge SM., Ritchie EG., and Woinarski JCZ. 2017. Impacts and management of feral cats Felis catus in Australia. Mammal Review 47: 83-97.

Fattorini S., Sciotti A., Tratzi P., and Giulio A. 2013. Species distribution, ecology, abundance, body size, and phylogeny originate interrelated rarity patterns at regional scale. Journal of Zoological Systematics and Evolutionary Research 51: 279- 286.

Fisher DO. 2011. Cost, effort and outcome of mammal rediscovery: neglect of small species. Biological Conservation 144: 1712-1718.

Fisher RA., Dickman CR., Jones M., and Blomberg SP. 2013. Sperm competition drives the evolution of suicidal reproduction in mammals. PNAS 110: 17910–17914.

Fisher DO., Johnson CN., Lawes MJ., Fritz SA., McCallum H., Blomberg SP., VanDerWal J., Abbott B., Frank A., Legge S., Letnic M., Thomas CR., Fisher A., Gordon IJ., and Kutt A. 2014. The current decline of tropical marsupials in Australia: is history repeating? Global Ecology and Biogeography 23: 181-190.

Flather CH., and Sieg CH. 2007. Species rarity: definition, causes, and classification. In: Raphael MG and Molina R (eds) Conservation of rare or little known species: biological, social, and economic considerations. Island Press, USA.

Fleming PA., and Bateman PW. 2016. The good, the bad, and the ugly: which Australian terrestrial mammal species attract the most research? Mammal Review 46: 241-254.

Fox BJ and Archer E. 1984. The diet of Sminthopsis murina and Antechinus stuartii in sympatry. Australian Wildlife Research 11: 235–248.

Gaston KJ. 1994. Rarity. Chapman and Hall, London, UK.

Gaston KJ. 1998. Rarity as double jeopardy. Nature 394: 229-230.

Gaston KJ., and Fuller RA. 2007. Commonness, population depletion and conservation biology. Trends in Ecology and Evolution 23: 14-19.

Chapter 1: General introduction 21

Gause G. 1934. The struggle for existence. Williams and Wilkins, USA.

Goodyear NC. 1992. Spatial overlap and dietary selection of native rice rats and exotic black rats. Journal of Mammalogy: 73, 186–200.

Grenyer R., Orme C., Jackson SF., Thomas GH., Davies RG., Davies TJ et al. 2006. Global distribution and conservation of rare and threatened vertebrates. Nature 444: 93-96.

Hall 1980. The Diets of Two Coexisting Species of Antechinus (Marsupialia: Dasyuridae). Wildlife Research 7: 365-378.

Harcout AH., Coppeto SA., and Parks SA. 2002. Rarity, specialisation and extinction in primates. Journal of Biogeography 29: 445-456.

Harrington LA., and Macdonald DW. 2008. Spatial temporal relationships between invasive American mink and native European polecats in the southern United Kingdom. Journal of Mammalogy 89: 991-1000.

Higgs P., and Fox BJ. 1993. Interspecific competition: a mechanism for rodent succession after fire in wet heathland. Australian Journal of Ecology 18: 193–201.

Hanski I., Kouki J., and Halkka A. 1993. Three explanations of the positive relationship between distribution and abundance of species. In: Ricklefs R., and Schluter D (eds) Species diversityin ecological communities: historical and geographical perspectives. University of Chicago Press, USA.

IUCN. 2016. Guidelines for using the IUCN red list categories and criteria. Version 12.

IUCN. 2016b. International Union Conservation of Nature Species Survival Commission. http://www.iucnredlist.org

Johnson CN. 1998. Species extinction and the relationship between distribution and abundance. Nature 394: 272-274.

Johnson CN. 2006. Australia’s mammal extinctions: a 50 000 year history. Cambridge University Press, Australia.

Johnson CN., and Isaac JL. 2009. Body mass and extinction risk in Australian marsupials: The ‘critical weight range’ revisited. Austral Ecology 34: 35-40.

22 Chapter 1: General introduction

Kattan GH. 1992. Rarity and vulnerability: the birds of the Cordillera Central of Colombia. Conservation Biology 6: 64-70.

Kunin WE., and Gaston KJ. 1997. The biology of rarity: causes and consequences of rare-common differences. Chapman and Hall, London, UK.

Lande R. 1993. Risks of population extinction from demographic and environmental stochasticity and random catastrophes. The American Naturalist 142: 911-927.

Lee AK., Woolley PA., and Braithwaite R. 1982. Life history strategies of dasyurid marsupials. In: Archer M (ed). Carnivorous marsupials. Sydney Royal Zoological Society of New South Wales Vol 1: 1-11.

Lindenmayer D., and Burgman M. 2005. Practical conservation biology. CSIRO Publishing, Australia.

Lindenmayer DB., Wood JT., McBurney L., MacGregor C., Youngentob K., and Banks SC. 2011. How to make a common species rare: A case against conservation complacency. Biological Conservation 144: 1663-1672

Luo J., and Fox BJ. 1996. Seasonal and successional dietary shifts of two sympatric rodents in coastal heathland: a possible mechanism for coexistence. Australian Journal of Ecology 21: 121–32.

Mace GM., and Lande R. 1991. Assessing extinction threats: toward a reevaluation of IUCN threatened species categories. Conservation Biology 5: 148-157.

Mace GM., Collar NJ., Gaston KJ., Hilton-Taylor C., Akçakaya HR., Leader- Williams N., Milner-Gulland EJ., and Stuart SN. 2008. Quantification of extinction risk: IUCN’s system for classifying threatened species. Conservation Biology 22: 1- 19.

Macleay WS. 1841. Notice of a new genus of Mammalia discovered by J. Stuart, Esq., in New South Wales. Annals and Magazine of Natural History VIII: 241–243.

McAllan BM., Dickman CR., and Crowther MS. 2006. Photoperiod as a reproductive cue in the marsupial genus Antechinus: ecological and evolutionary consequences. Biological Journal of the Linnean Society 87: 365-365.

Chapter 1: General introduction 23

McCallum ML. 2015. Vertebrate biodiversity losses point to a sixth mass extinction. Biodiversity Conservation 24: 2497-2519.

McDonald LL. 2004. Sampling rare populations. In: Thompson W (ed) Sampling rare or elusive species: concepts, designs, and techniques for estimating population parameters. Island Press, USA.

McKenzie NL., Burbidge AA., Baynes A., Brereton RN., Dickman CR., Gordon G., Gibson LA., Menkhorst PW., Robinson AC., Williams MR., and Woinarski JCZ. 2007. Analysis of factors implicated in the recent decline of Australia’s mammal fauna. Journal of Biogeography 34: 597-611.

Molina R., and Marcot BG. 2007. Definitions and attributes of little-known species. In: Raphael MG and Molina R (eds) Conservation of rare or little known species: biological, social, and economic considerations. Island Press, USA.

Morin PJ. 2011. Community ecology (2nd edition). John Wiley & Sons, UK.

Munton P. 1987. Concepts of threat to the survival of species used in red data books and similar compilations. In: Fitter R., and Fitter M (eds) The road to extinction: problems of categorizing the status of taxa threatened with extinction. IUCN, Switzerland and Cambridge, UK.

Murray BR., Thrall PH., Gill MA., and Nicotra AB. 2002. How plant life-history and ecological traits relate to species rarity and commonness at varying spatial scales. Austral Ecology 27: 291-310.

Mutton TY. 2016. Evolutionary biology of the Australian carnivorous marsupial Genus Antechinus. PhD thesis, Queensland University of Technology, Australia.

Naylor R., Richardson SJ., and McAllan BM. 2008. Boom and bust: a review of the physiology of the marsupial genus Antechinus. Journal of Comparative Physiology 178: 545-562.

Olden JD., LeRoy N., and Bestgen KR. 2008. Trait synergisms and the rarity, extirpation, and extinction risk of desert fishes. Ecology 89: 847-856.

Paine MD. 1990. Life history tactics of darters (percidae: Etheostomatiini) and their relationship with body size, reproductive behaviour, latitude and rarity. Journal of Fish Biology 37: 473-488.

24 Chapter 1: General introduction

Parra Faundes D. 2014. Sex allocation, terminal investment and the effects of environmental constraints on maternal investment in subtropical antechinus. PhD thesis, University of Queensland, Australia.

Pärtel M., Kalamees R., Reier Ü., Tuvi E., Roosaluste E., Vellak A., and Zobel M. 2005. Grouping and prioritization of vascular plant species for conservation: combining natural rarity and management need. Biological Conservation 123: 271- 278.

Pimm SL., Jenkins CN., Abell R., Brooks TM., Gittleman JL., Joppa LN., Raven PH., Roberts CM., and Sexton JO. 2014. The biodiversity of species and their rates of extinction, distribution, and protection. Science 344: 987-997.

Rabinowitz D. 1981. Seven forms of rarity In: Synge H (ed) The biological aspects of rare plant conservation. John Wiley & Sons Ltd, USA.

Rabinowitz D., Cairns S., and Dillon T. 1986. Seven forms of rarity and their frequency in the flora of the British Isles. In: Soulé ME (ed) Conservation biology, the science of scarcity and diversity. Sinauer Associates INC, USA.

Reinhardt K., Köhler G., Maas S., Detzel P., and Spence JR. 2005. Low dispersal ability and habitat specificity promote extinctions in rare but not in widespread species: the Orthoptera of Germany. Ecography 28: 593-602.

Reveal JL. 1981. The concepts of rarity and population threats in plant communities. In: Morse LE., and Henefin MS (eds) Rare plant conservation. The New York Botanical Garden, USA.

Sale MG., and Arnould JP. 2009. Spatial and Temporal Organization in the Swamp Antechinus: Comparison between Island and Mainland Populations. Journal of Mammalogy 90: 347-355.

Sanecki GM., Green K., Wood H., and Lindenmayer D. 2006. The influence of snow cover on home range and activity of the Bush Rat (Rattus fuscipes) and the Dusky Antechinus (Antechinus swainsonii). Wildlife Research 33: 489-496.

Schoener TW. 1974. Resource partitioning in ecological communities. Science 185: 27-39.

Chapter 1: General introduction 25

Schoener TW. 1986. Resource partitioning. In: Kikkawa J., and Anderson DJ (eds) Community ecology: pattern and process. Blackwell Scientific Publications, Australia.

Simberloff D. 1998. Small and declining populations. In: Sutherland WJ (ed) Conservation science and action. Blackwell Science Ltd.

Soulé ME. 1987. Where do we go from here? In: Soulé ME (ed) Viable populations for conservation. Cambridge University Press, UK.

Statham HL. 1982. Antechinus stuartii (Dasyuridae, Marsupialia) diet and food availability at Petroi, north-eastern New South Wales. In: Archer M (ed) Carnivorous Marsupials Vol 1. Sydney Royal Zoological Society of New South Wales.

Usher MB. 1986. Wildlife conservation evaluation: attributes, criteria and values. In: Usher MB (ed) Wildlife Conservation evaluation. Chapman and Hall, London, UK.

Van Dyck S. 2002. Morphology-based revision of and Antechinus (Marsupialia: Dasyuridae). Memoirs of the Queensland Museum 48: 239-330.

Van Dyck S., and Crowther M. 2000. Reassessment of northern representatives of the Antechinus stuartii complex (Marsupialia: Dasyuridae): A subtropicus sp. nov. and A. adustus new status. Memoirs of the Queensland Museum 45: 611-635.

Van Dyck S., Gynther I., and Baker AM. 2013. Field companion to mammals of Australia. New Holland Publishers, Sydney, Australia.

Van Dyck S., and Strahan R. 2008. The mammals of Australia. New Holland Publishers, Sydney, Australia.

Watt A. 1997. Population ecology and reproductive seasonality in three species of antechinus (Marsupialia: Dasyuridae) in the Wet Tropics of Queensland. Wildlife Research 24: 531-539.

Westerman M, Krajewski C., Kear BP., Meehan L., Meredith RW., Emerling CA., and Springer MS. 2015. Phylogenetic relationships of dasyuromorphian marsupials revisited. Zoological Journal of the Linnean Society 176: 686-701.

Williams SE., Scheffers BR., VanDerWal J., Roslan N., Anderson A., Storlie C., Brunner R., Welbergen J., and Parsons S. 2014. Species resilience: the key to

26 Chapter 1: General introduction

understanding biodiversity in the rainforests of the Australian Wet Tropic. Report to the National Environment Research Program. Reef and Rainforest Research Centre Limited, Cairns.

Williams R., and Williams A. 1982. The life cycle of Antechinus swainsonii (Dasyuridae, Marsupialia). In: Archer M (ed). Carnivorous Marsupials. Sydney Royal Zoological Society of New South Wales Vol 1 Pp 89-95.

Wilson BA., Dickman CR., and Fletcher TP. 2003. Dasyurid dilemmas: problems and solutions for conserving Australia’s small carnivorous marsupials. In: Jones M., Archer M., and Dickman CR (eds) Predators with Pouches: the biology of carnivorous marsupials. CSIRO Publishing, Australia.

Woinarski JCZ., Burbidge AA., and Harrison PL. 2014. The action plan for Australian mammals 2012. CSIRO Publishing, Australia.

Woinarski JCZ., Burbidge AA., and Harrison PL. 2015. Ongoing unravelling of a continental fauna: decline and extinction of Australian mammals since European settlement. PNAS 112: 4531-4540.

Wood DH. 1970. An ecological study of Antechinus stuartii (Marsupialia) in a south-east Queensland rain forest. Australian Journal of Zoology 18: 185-207.

Woolley PA. 1966. Reproduction in antechinus spp. And other dasyurid marsupials. Symposium of the Zoological Society of Australia 15: 281-294.

Yu J., and Dobson SF. 2000. Seven forms of rarity in mammals. Journal of Biogeography 27: 131-139.

Yuncken D. 1997. Diet and prey availability of Antechinus stuartii (Marsupialia: Dasyuridae). Honours Thesis, Monash University, Melbourne.

.

Chapter 1: General introduction 27

28 Chapter 1: General introduction

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Emma L. Gray & Andrew M. Baker & Jennifer Firn

Received: 29 March 2016 /Accepted: 30 May 2016 /Published online: 24 June 2016 in Mammal Research.

2.1 ABSTRACT

The carnivorous marsupial Antechinus are one of the few mammal genera known to exhibit the phenomenon of semelparous reproduction, where all males die at the end of a frenetic annual mating period. The genus Antechinus has recently been revised, seeing four new species named and one existing subspecies raised to species status. Here, we present the first ecological assessment of one new species, the endangered black-tailed dusky antechinus (Antechinus arktos), based on a 2-year mark-recapture study of two proximate sites within the cloud forest of Springbrook National Park, southeast Queensland, Australia. We also present comparative ecological data from a sympatric congener, Antechinus stuartii. In total, 103 A. arktos (49 male; 54 female) and 2125 A. stuartii (1229 male; 896 female) captures were made over 16,630 trap nights at the two sites. The occurrence and synchrony of reproductive events observed in A. arktos closely parallel the semelparous reproductive strategy exhibited by all congeners studied to date. The A. arktos populations mate during mid-September, with birth of up to six young occurring during mid-October. The low abundance and fecundity, together with a limited and apparently retracting distribution, suggest A. arktos is under threat of climate-induced extinction in the coming decades.

Keywords Breeding biology. Conservation. Dasyuridae. Mark-recapture. Population dynamics

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 29

2.2 INTRODUCTION

The dasyurid marsupials of the genus Antechinus have long been a source of ecological interest. Antechinus are small (15– 180 g) carnivorous marsupials, with both semi-arboreal and ground-dwelling species that occur broadly in temperate and tropical regions of Australia (Van Dyck and Strahan 2008). They are predominantly nocturnal, insectivorous and sexually dimorphic for size, with males larger than females (Wood 1970; Friend 1985; Dickman 1988; Watt 1997; Leung 1999; Marchesan and Carthew 2004). Both sexes occupy overlapping home ranges, with male ranges typically larger than those of females (Marchesan and Carthew 2004).

Breeding studies have been a central focus of antechinus research (Woolley 1966; McAllan et al. 2006; Fisher et al. 2013) because, like only a few other mammal genera worldwide, they are semelparous reproducers. Females are synchronously monoestrous and mating is confined to a short (1–3week) period, at the end of which all males in the population die (Woolley 1966; Wood 1970; Lee et al. 1982; Williams and Williams 1982; McAllan et al. 2006). Recent molecular and morphological assessments of the genus recognised ten extant species (Krajewski et al. 2007). However, in the past 4 years, antechinus taxonomy has been revised. This has resulted in five new species, including the black-tailed dusky antechinus, Antechinus arktos (Baker et al. 2012, 2013, 2014, 2015).

A. arktos once inhabited several altitudinal zones and a range of sites on the slopes of the eroded Tweed Shield Volcano caldera that straddles the border of New South Wales (NSW) and Queensland (QLD) on Australia’s mid-east coast (Baker et al. 2014). Today, however, A. arktos is only known from several proximate sites at the summit of the caldera in World Heritage Listed cloud forest, where populations occur in low density (see Baker et al. 2014 for a distribution map).

To date, no ecological studies of the endangered A. arktos have been undertaken. Virtually nothing is known about this endangered species’ basic ecology or relationship with its only known sympatric congener, the brown antechinus, Antechinus stuartii. As currently understood, A. stuartii is widespread in a variety of forest and heathland habitats within its range. However, recent genetic work has identified two forms of A. stuartii (a northern and southern clade) within this range, with proposed parapatry along a geographic barrier in New England National Park, north-east NSW (Mutton 2016). A recent study (Parra Faundes 2014) was completed

30 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

on aspects of northern clade A. stuartii (then recognised as Antechinus subtropicus) breeding, parentage and recruitment at Springbrook National Park. All other previous research on A. stuartii has been conducted on the southern clade (Marlow 1961; Statham 1982).

The present paper thus presents results of the first detailed ecological assessment of A. arktos based on mark-recapture trapping data from two proximate sites. We also present comparative ecological data from the little-studied northern form of sympatric congener, A. stuartii. Because of its unique mating strategy, the genus Antechinus is a model for studies in evolutionary ecology, requiring a clearer understanding of species boundaries and dynamics. The present study will help fill this knowledge gap. Fundamental autecological data will also be compared and contrasted within and between the two species, then considered in relation to published work on congeners in order to: (1) test established theories regarding breeding timing of antechinus within and between species, sites and years; (2) place the ecology of the recently described and endangered A. arktos into perspective: and (3) facilitate conservation recommendations.

2.3 MATERIALS AND METHODS

2.3.1 Study areas Our mark-recapture study was conducted at Best of All Lookout (28° 14′ 29.6002″ S, 153° 15′ 50.6002″ E) and Bilborough Lookout (28° 14′ 03.0001″ S, 153° 17′ 22.9999″ E) within Springbrook National Park (Fig. 2.1). The site at Best of All Lookout (‘Best of All’) is characterised by complex notophyll vine forest on Cainozoic igneous rocks (Regional Ecosystem 12.8.5; Queensland Herbarium 2013) at an average elevation of 950 m. The site contains a moderately steep headwater gully with flax lily, Helmholtzia glaberrima, and a small stand of Regional Ecosystem 12.8.6 (simple microphyll fern forest with Antarctic beech, Nothofagus moorei, on Cainozoic igneous rocks; Queensland Herbarium 2013).

The site at Bilborough Lookout (‘Bilborough’) is approximately 2.5-km straight- line distance east of Best of All at an average elevation of 950 m and is split into two adjacent sections on either side of a radio control tower. One side is characterised by regrowth Regional Ecosystem 12.8.6 surrounded by Regional Ecosystem 12.8.5 (Queensland Herbarium 2013) and consists of very dense vines, a low thick canopy

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 31

and rocky substrate. The other side is characterised by regrowth 12.8.6 surrounded by non-remnant vegetation (Queensland Herbarium 2013) and is notably more open.

Figure 2.1. Location of the two trapping sites within Springbrook National Park, 100 km south of Brisbane, Queensland’s capital, situated on the east coast of Australia

2.3.2 Trapping Trapping was conducted monthly between April and October, for 2 years (2014– 2015), to monitor biology of adults before, during and immediately after breeding. Animals were captured using type A aluminium Elliott folding traps (Elliott Scientific Equipment, Upwey Vic) baited with peanut butter, oats and bacon pieces (as standard, see Dickman 1986a).

32 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

At Best of All, four parallel line transects were established, each with a total of 25 tags, positioned at 8-m intervals along the line (total length = 200 m). Lines were separated by approximately 10 m. At Bilborough, due to steep terrain and encroaching property boundaries, two parallel lines (approximately 10 m apart) were necessarily established in each section of the site, each with a total of 25 tags positioned at 8 m intervals along the line (total length = 200 m). During an April 2014 pilot study, a high volume of non-target species (chiefly rodents: Rattus fuscipes and Melomys spp.) were caught. Therefore, two traps were placed at each marker, spaced a minimum distance of 1 m apart, (totalling 200 traps per site; i.e. four lines of 50). This double-trap design has been used successfully in other studies with similarly high captures of non-target species in order to increase the number of available open traps per unit area in the field (e.g. Dickman 1986b in a study of Antechinus agilis and Antechinus mimetes).

In their early taxonomic study of Springbrook A. arktos and A. stuartii, Baker et al. (2014) suggested both species were predominantly nocturnal. Therefore, in the present study, during each month trapping was conducted for a total of six nights (three consecutive nights per site) with traps set just on dusk, checked as close to dawn as possible and kept closed during daylight hours. All Antechinus captures were weighed (to the nearest 0.5 g using Pesola spring balances), sexed, ear clipped (for genetics in a parallel study), assessed for reproductive condition (see below) and micro-chipped with a passive integrated transponder (PIT) tag for recapture ID. Individuals were then immediately released at point of capture. Non-target species were identified and then also released at point of capture.

2.3.3 Assessing reproductive condition The commencement and duration of the breeding season for each species at each site was determined via a variety of direct and indirect measures that have been used in other antechinus studies (e.g., Woolley 1966; McAllan et al. 2006). For female captures, the pouch was closely examined and assigned a development score as per Woolley (1966). Each female was allocated a pouch score from one to four. Stage one being the immature state prior to breeding, when the pouch is small, pale, fully to partially furred and the nipples are pale and inverted; stage two being the mature state occurring at the time of mating, when the pouch is enlarged, becomes pink and almost naked; stage three being the state during early to mid-stage pregnancy, when the covering hairs grow longer and the pouch area further expands; and stage four being

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 33

the state of late pregnancy, when the skin takes on a granular appearance and ridges between the nipples and sides of the pouch area develop and thicken. After birth, when the adult females were carrying offspring, crown-rump measurements of three (randomly chosen) young were made to the nearest 0.1 mm, using Mitutoyo manual callipers. Birth dates of A. arktos young were then estimated by referring crown-rump lengths to the known growth curve of the related and similar sized congener, A. mimetes (Williams and Williams 1982; Baker et al. 2015). Once this information was determined, the probable time of ovulation was then extrapolated using an estimation of the gestation period (as per McAllan et al. 2006). Note that this was an estimate of the timing of ovulation rather than mating timing, since some antechinus species are known to store sperm for up to 2 weeks (Selwood 1985). Because gestation times are not currently known for A. arktos, the A. mimetes gestation period of 30 days was assumed to apply most closely, since both species are members of the dusky antechinus complex (Williams and Williams 1982; Dickman 1985; Baker et al. 2015). The same method was used to calculate A. stuartii breeding using the growth curve and gestation times determined for that species by Marlow (1961).

2.3.4 Data analysis General linear models (GLMs) were constructed to test the effects of site (Best of all, Bilborough), sex (male, female), year (2014, 2015) and month (April–October) on trapping success, KTBA, survival and sex ratios for both species. GLMs were constructed by first loading the data sets into R studio and graphically checking the data for errors. An error distribution was then chosen for each dataset according to the type of data being examined (as per Crawley 2002) and the model formula, including the dependent variable and fixed effects inputted. After the model was run, Chi-square or Wald Z, F or t statistics were used to assess the significance of different fixed effects within the GLM by conducting likelihood ratio tests of the full model with the effect in question against the model without the effect in question (as per Bolker et al. 2009). Which statistic was used depended on whether the model was overdispersed, i.e. when the residual variance in the model was greater than the residual degrees of freedom (Crawley 2002). If overdispersion occurred, Wald Z or Chi square statistics were utilised and if overdispersion did not occur, Wald t or F statistics were adopted (Bolker et al. 2009). p values <0.05 were taken to be statistically significant.

34 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

GLMMs with body mass as the dependent variable were used to model the influence of fixed effects (sex, site and time) and random effects (individual ID) on body mass for each species. Because the data was found to be auto-correlated, we ran the same model with spatial correlation structures in the nlme package (Pinheiro et al. 2015). All model comparisons were undertaken using the model selection function in the MuMIn package (Barton 2016). The information-theoretic approach was used to select the “best supported model” as described by Burnham and Anderson (2002). All analyses were carried out using the lme4 (Bates et al. 2015) and nlme (Pinheiro et al. 2015) packages in R studio (v3.1.1 R Development Core Team 2014).

Trapping data For recapture data, residents were classified as those caught in more than one consecutive month, and transients as those caught only once, as per other studies of this type (e.g., Wood 1970; Gilfillan 2001).

Population dynamics Abundance was estimated by the minimum number of animals known to be alive (KTBA), calculated as the number of individuals caught in a given month in addition to previously marked individuals caught later but not at that time (Krebs 1966). Although there are more robust methods for calculating population size for this type of data (e.g., the Cormack-Jolly-Seber model; Williams et al. 2002), the low trap success and lack of recaptures of A. arktos during some trapping periods rendered these estimators unreliable.

Minimum survival between trapping sessions was calculated as the percentage of individuals known to be alive in one trapping session that were recorded subsequently. In this case, minimum survival was defined as the probability of an staying in place for a specified period, as it was not possible to distinguish whether an individual’s absence was due to death or emigration (Leung 1999; Glen 2008).

Breeding Timing of ovulation in A. stuartii was compared between sites and years by numbering the days of the year from January 1st (day 1) to December 31st (day 365) and performing a Kruskal-Wallis test using the mean of the days when the

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 35

phenomenon occurred (as per McAllan et al. 1991). A comparative calculation could not be made for A. arktos due to insufficient data (see ‘Results’ section).

The length of photoperiod prevailing at the time of ovulation in populations of each species was calculated as per McAllan and Dickman (1986) and McAllan et al. (2006), from a standard astronomical formula for the solar day. The rate of change of photoperiod (difference in photoperiod from 1 day to the next) was then calculated by subtracting each successive calculation for the solar day from the preceding calculation. Latitudes, longitudes and elevations for each locality were obtained directly from GPS readings taken during fieldwork.

2.4 RESULTS

2.4.1 Trapping In total, 103 A. arktos (49 males; 54 females) and 2,125 A. stuartii (1,229 males; 896 females) were captured and released over 16, 630 trap nights at the two sites (Table 2.1). This equated to 26 A. arktos individuals (15 males; 11 females) and 496 A. stuartii individuals (290 males; 206 females) pit-tagged and monitored over the 2 years of study.

In general, more individuals of both species were caught and recaptured at Best of All than Bilborough (see Table 2.1). The number of individuals and captures of A. stuartii at Best of All were comparable between years (Table 2.1) but dropped noticeably at Bilborough (from 535–355 captures and 127–88 individuals, respectively; Table 2.1). In comparison, there were increases in both the number of individuals (by more than double) and number of captures of A. arktos at both sites between years (Table 2.1). There was also a strong male bias in terms of both number of individuals and number of captures for A. stuartii in both years. A similar bias was observed in A. arktos during 2014 but there was a female capture bias in 2015 (Table 2.1). Other mammal species captured during the study were the bush rat (Rattus fuscipes), mosaic-tailed rat (Melomys spp.) and the northern brown bandicoot (Isoodon macrourus).

36 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Table 2.1 Summary of antechinus and non-target species trapping data gathered at Best of All and Bilborough between 2014 and 2015

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 37

Recaptures 53.8 % of A. arktos were recaptured at least once (average = four times), while 61.3 % of A. stuartii were recaptured at least once (average = 4.3 times) (Table 2.1). Males of both species always had a higher recapture probability than females, despite females living longer (see ‘Population dynamics’ section).

Trap success Overall trap success for A. arktos was low at 0.6 %, while A. stuartii was much higher, at 12.8 % (Table 2.1). This difference was significant (t = 9.082, d.f = 110, p < 0.001). R. fuscipes constituted the majority of captures with a total trap success of 21 %, with Melomys constituting 9.5 % total trap success; however, even with this high density of rodents and occasional trap disturbance by scrub turkeys (Alectura lathami), the overall percentage of traps remaining open after a night of trapping was high, 53.5 %.

Temporal variation in antechinus trap success was also high, ranging from 3.2 to 26.3 % for A. stuartii and 0 to 3.5 % for A. arktos among sites and trapping sessions (Fig. 2.2). The greatest trapping success occurred during the winter months, corresponding with the lead up to and beginning of the breeding seasons (June, July, August; Fig. 2.2). For A. stuartii, the lowest trapping success occurred after male die-off during female pregnancy and lactation (September, October), while A. arktos exhibited sporadic trap success at Bilborough and outside the breeding period at Best of All (Fig. 2.2).

There was no difference in trap success between sites or years for A. stuartii. However, trap success of this species differed between months (t = −4.322, d.f = 53, p = <0.05), with extracted model parameters showing highest trap success in July and lowest in October. Males showed greater trap success than females (t = 2.565, d.f = 53, p = <0.05), although there was no interaction between sex and month. In contrast, trap success did not differ between months for A. arktos. Instead, variation can mostly be explained by effects of site and year, with greater success recorded at Best of All (t = −3.326, d.f = 53, p = <0.001) and in the second year at both sites (t = 3.042, d.f = 53, p = <0.003).

38 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Figure 2.2. Trap success over the 2-year trapping period. a Trap success of A. arktos at Best of All Lookout. b Trap success of A. arktos at Bilborough Lookout. c Trap success of A. stuartii at Best of All Lookout. d Trap success of A. stuartii at Bilborough Lookout

2.4.2 Population dynamics A. stuartii Mean population abundance did not differ between sites (f = 1.738, d.f = 1, p = 0.193) or years (f = 0.230, d.f = 1, p = 0.634), but did differ between months (f = 19.885, d.f = 6, p < 0.0001).

Monthly KTBA at each site varied from five to 88 (mean = 47.0, Std = 25.4) at Best of All and from six to 70 (mean = 35.9, Std = 23.6) at Bilborough and appeared to exhibit an annual cycle of fluctuation (Fig. 2.3). KTBA increased prior to and during

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 39

the first part of the mating season (June, July, August), corresponding with the appearance of predominantly male transients and the increased mobility of both sexes. There was a sharp decrease after post-mating death of males in late August / early September (Fig. 2.3). Excluding the months of September and October, mean sex ratio did not differ between sites (X2 = 11.903, d.f = 1, p = 0.083) or trapping sessions (X2 = 12.136, d.f = 18, p = 0.96) but more males than females were trapped (f = 7.455, d.f = 1, p = 0.009). The overall sex ratio at Best of All was 1.39, and the ratio at Bilborough was 1.83. However, at Bilborough in June, July and August, there were 2.4, 2.2 and 3.1 males for every female trapped, respectively.

Figure 2.3. Minimum number of A. stuartii known to be alive (KTBA) at Best of All and Bilborough

Mean survival was calculated from KTBA for each interval between months. Survival between August and September was clearly influenced by sex due to post- mating death of males during this interval. However, mean survival over the remaining intervals was not found to be affected by sex (t = 0.239, d.f = 30, p = 0.811). Mean survival between sites was also not different (t = −0.088, d.f = 45, p = 0.93).

40 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Mean monthly survival for A. stuartii was estimated at 65.9 % (range 36–100 %) for females and 46 % (0–91 %) for males. To test for temporal differences in survival, males and females were separated (because of the known interaction between sex and sampling time). The test for differences among the 13 intervals for males was significant (X2 = 5.6621, d.f = 25, p = 0.017). Mean male survival was highest between April–May and lowest between August– September once post-mating male death had occurred. The test for females was also significant (X2 = 6.968, d.f = 25, p = 0.015). Mean female survival was also highest during the first two intervals (April–May) and lowest during and after male die-off (August–September and September–October).

Overall, in the present study, 73.3 % of females at Best of All and 36.4 % of females at Bilborough (total n = 15) alive in October 2014 survived until the next year. Resident females were trapped for a mean of 4.35 months (range 2–14) at both sites and the second-year females survived for a mean of 8.2 months (range 6–11) after their first litters. However, none of these second-year females apparently survived long enough to successfully carry a second litter. In comparison, males were trapped for a mean of 3.3 months (range 2–6)

A. arktos Mean population abundance was larger at Best of All (f = 8.924, d.f = 1, p = 0.006) than Bilborough. At both sites, abundance did not differ between months (Best of All f = 0.002, d.f = 6, p = 0.96; Bilborough f = 0.247, d.f = 6, p = 0.628), but there were more individuals in the second year (Best of All f = 4.809, d.f = 1, p = 0.0487; Bilborough f = 11.33, d.f = 1, p = 0.0056).

Monthly KTBA varied from zero to eight at Best of all (mean = 3.1, Std = 2.09) and from one to three at Bilborough (mean = 1.2, Std = 0.97). Although not statistically significant, KTBA was lowest at both sites before the breeding season and after post- mating death of males, and highest just prior to and during the breeding season (Fig. 2.4).

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 41

Figure 2.4. Minimum number of A. arktos known to be alive (KTBA) at Best of All and Bilborough

Because of particularly poor trap success of female A. arktos at Bilborough, sex ratios were not compared between sites or trapping sessions. However, there was no difference between overall number of trapped males and females (f = 0.071,d.f = 1,p = 0.791).

Because of the low KTBA of A. arktos at both sites, mean survival between months was highly variable. Several times throughout the study, there were no individuals known to be alive at a site, while at others, a single individual captured over several months produced 100 % survival probabilities for a site. For these reasons, no statistical analyses were performed.

Overall, no females captured in the first year were recaptured in the second. Resident females were trapped for a mean of 6.2 months (range 2-7); resident males were trapped for a mean of 2.9 months (range 2-6).

2.4.3 Body mass Between species There was a difference in average body mass between species and between sexes (Z = 1990.572, d.f = 3, p = <0.001). On average, A. arktos males were 47–49 g heavier

42 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

than A. stuartii males (Table 2.2) and A. arktos females were 23–24 g heavier than A. stuartii females (Table 2.2), although A. stuartii males at the maximum of their body mass range did overlap with A. arktos females at the minimum of their range (Table 2.2).

Measurements of second-year female A. stuartii were not included in the model as they could only be accurately identified in the second year of the study; however, on average, they were 6–7 g heavier than first-year females and more similar in size to A. stuartii males (Table 2.2).

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 43

Table 2.2. Model extracted mean weights of antechinus subdivided by year, site, species, sex and age class

Year Site Species Sex Weight (g)

Mean Range Std. D SE N 2014 BA A. arktos Male 87.6 64–104.5 12.04 3.22 14 Female 57.8 46–69 9.55 3.34 8

A. stuartii Male 33.3 16–54 5.68 0.33 294

Female 24.4 16–40 4.81 0.29 268

BC A. arktos Male 75.8 62–113 19.6 8.00 6

Female – – – – –

A. stuartii Male 34.8 18–51.5 6.15 0.35 308

Female 23.2 15.5–36.5 4.36 0.33 179

2015 BA A. arktos Male 79.2 61–101 10.91 2.14 26 Female 43.9 36.5–55 5.09 0.89 33

A. stuartii Male 28.7 18–45 4.10 0.22 335

Female 21.7 16–31.5 3.05 0.20 231

2nd year 28.8 20–42.5 4.18 0.56 56 female BC A. arktos Male 83.5 70–98 14.03 8.10 3

Female 48 41–57 4.87 1.41 12

A. stuartii Male 32.4 22–42 4.57 0.31 224

Female 23.6 15.5–33.5 4.28 0.46 85

2nd year 26.8 23–33.5 2.87 0.52 30 female

44 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Within species (A. stuartii) Thirteen GLMMs were compared (Table 2.3) to determine the influence of fixed effects (sex, site and month) and random effects (individual ID) on body mass for A. stuartii. The best supported model (model 1) was the full model and included five parameters: sex, site, month and the interaction between sex and site and sex and month. On average, males were found to increase in body mass each month they were captured until the breeding period, after which they sharply declined (Fig. 2.5); females increased in mass after the breeding period (Fig. 2.5). Males and females at both sites were lighter in 2014 than in 2015 (Table 2.2).

Table 2.3. Results of model selection for A. stuartii

Model Parameters df LogLik AICc delta Weight 1 1, 2, 3, 4, 5 11 −5437.56 10,897.25 0.00 0.98 2 1, 2, 3, 4 9 −5444.13 10,906.34 9.09 0.01 3 1, 2, 3, 5 10 −5443.90 10,907.91 10.77 0 4 1, 2, 3 8 −5451.19 10,918.46 21.20 0 5 1, 2, 4 7 −5453.33 10,920.72 23.46 0 6 1, 2 6 −5460.48 10,933.00 35.74 0 7 2, 3, 5 9 −5479.95 10,978.00 80.74 0 8 2, 3 7 −5486.66 10,987.38 90.12 0 9 2 5 −5497.82 11,005.68 108.42 0 10 1, 3 7 −5593.13 11,200.31 303.05 0 11 1 5 −5599.23 11,208.49 311.24 0 12 3 6 −5616.34 11,244.71 347.46 0 13 (Null) 4 −5623.45 11,254.92 357.67 0

Parameters: 1 = month, 2 = sex, 3 = site, 4 = month: sex, 5 = month: site.

Marginal R2 value = 0.438. Conditional R2 value = 0.742

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 45

Figure 2.5. Body mass of A. arktos and A. stuartii between April and October (2014 and 2015) at both sites

Within species (A. arktos) Nineteen GLMMs were compared (Table 2.4) to determine the influence of fixed effects (sex, site and time) and random effects (individual ID) on body mass for A. arktos. The best supported model (model 1) was the full model and included seven parameters: sex, site, month and the interaction between sex and site, sex and month, month and site, and month, sex and site. On average, males were found to increase in mass each month they were captured until the breeding period, after which they sharply declined (Fig 2.5); females increased after the breeding period (Fig 2.5). Male and female A. arktos at Best of All had notably lower average body masses in 2014 compared to 2015; average mass of males at Bilborough increased between 2014-2015 (Table 2.2).

46 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Table 2.4. Results of model selection for A. arktos

Model Parameters df LogLik AICc Delta Weight 1 1, 2, 3, 4, 5, 11 −321.01 666.88 0.00 0.99 6, 7 2 1, 2, 3, 4, 5, 10 −327.69 677.75 10.87 0 6 3 1, 2, 3, 5, 6 9 −329.14 678.20 11.32 0 4 1, 2, 3, 4, 5 9 −331.01 681.93 15.05 0 5 1, 2, 3, 5 8 −332.45 682.42 15.54 0 6 1, 2, 3, 4, 6 9 −331.83 683.58 16.70 0 7 1, 2, 3, 6 8 −333.33 684.18 17.30 0 8 2, 3, 6 7 −334.72 684.60 17.72 0 9 1, 2, 3, 4 8 −335.27 688.07 21.18 0 10 1, 2, 3 7 −336.76 688.68 21.79 0 11 2, 3 6 −338.23 689.33 22.45 0 12 1, 2, 4 7 −337.77 690.71 23.82 0 13 1, 2 6 −339.27 691.40 24.52 0 14 2 5 −340.78 692.17 25.29 0 15 1, 3, 5 7 −348.06 711.28 44.40 0 16 3 5 −353.75 718.11 51.22 0 17 1, 3 6 −352.88 718.63 51.75 0 18 (Null) 4 −356.98 722.36 55.47 0 19 1 5 −356.08 722.77 55.88 0

Parameters: 1 =month, 2 = sex, 3 = site, 4 =month: sex, 5 =month: site, 6

= sex: site, 7 = month: sex: site. Marginal R2 value = 0.603. Conditional R2 value = 0.900

2.4.4 Photoperiod In A. arktos, ovulation ranged from 10th–15th September at a prevailing day length between 11 h 43 min and 11 h 50 min. The average rate of change of photoperiod at Best of All was 90 s/day. Only one female was caught with offspring at Bilborough (ovulation 14th September ± 1 day) so photoperiod was not calculated for this site. In A. stuartii, ovulation ranged from 26th August to 9th September at a shorter prevailing day length between 11 h 19 min and 11 h 41 min. The average rate of change of photoperiod at Best of All was 106 s/day and Bilborough 92 s/day.

2.4.5 Reproductive patterns between species The pattern of reproductive development in A. arktos and A. stuartii was generally the same, with female pouch reddening occurring first, followed by the onset of male deterioration, male die-off and finally birth of offspring. However, the timing of these events was found to differ slightly, with A. arktos females having an earlier-

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 47

onset and potentially longer duration of pouch redness before ovulation. Unfortunately, because only three A. arktos mothers were caught, comparison of ovulation timing between species could not be statistically compared. However, in 2015, A. arktos ovulated between 10– 19 days later than A. stuartii at Bilborough and almost simultaneously at Best of All, with ovulation just 1 to 9 days later than A. stuartii.

2.4.6 Reproductive patterns within species Statistical analysis of ovulation timing between sites was not conducted for A. arktos due to low number of captures; however, based on the three individuals caught, ovulation timing was found to overlap. In A. stuartii, ovulation timing was found to vary between sites and years (independent samples median test = 10.711, d.f = 3, p = 0.013), with females at Best of All found to ovulate 4 days later in 2015 than 2014 and females in 2015 found to ovulate 9 days later at Best of All compared to Bilborough.

2.5 DISCUSSION

We found both species conformed to the semelparous breeding model proposed for antechinus (Woolley 1966; Lee et al. 1982), with synchronous reproduction culminating in post-mating death of all males. Observed variability in A. stuartii trappability, population density and size were all strongly associated with this reproductive cycle. A. arktos exhibited similar demographic patterns, albeit within limits imposed by their much lower abundance at each site.

2.5.1 A. stuartii Trapping and population dynamics A. stuartii are highly abundant in Springbrook National Park. The overall trap success of A. stuartii fell within the range recorded for other populations (Marlow 1961; Statham 1982); notably, the maximum monthly success rate of 26.3 % recorded at our Best of All Lookout site in 2014 is among the highest recorded for any antechinus species (see Table 2.5). The population density (KTBA), ranging from five to 88 over the study period, was also high for such a small study area. By definition, however, KTBA is a minimal estimate of population size, which suffers from negative bias, particularly towards the beginning and end of a mark-recapture study (Krebs, 1966; White 2005). The ‘true’ population size of A. stuartii at both sites is therefore likely even higher than our estimates.

48 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Table 2.5. Distribution, trap success, body weight and reproductive timing and synchrony in all 15 known species of antechinus

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 49

Body mass The range in adult body mass of A. stuartii recorded in our study (males 23–50 g; females 15–43 g) is similar to ranges compiled for the species as a whole by Van Dyck and Strahan (2008) (Table 2.5).

Breeding The occurrence and synchrony of reproductive events in A. stuartii at Springbrook closely paralleled those reported by Parra Faundes (2014) on other Springbrook A. stuartii, by Woolley (1966) on NSW and Australian Capital Territory populations of A. agilis (then A. stuartii) and other congeners (Williams and Williams 1982). At Springbrook, the northern clade of A. stuartii cued to a rate of change of photoperiod between 92–106 s/day, which is close to the 99–111 s/day range calculated for the species by McAllan et al. (2006). We also observed temporal trends in density, survival and size of antechinus over the breeding period. Population density increased prior to and during the first part of the mating season, corresponding with the appearance of a large number of predominantly male transients at both sites (2014, n = 25; 2015, n = 15). Our predicted population density decreased upon male deterioration and death, and when females had pouch young, as per other populations of A. stuartii and congeners (Wood 1970; Parra Faundes 2014). Our observed reduced trappability of female A. stuartii is likely because a female’s movement is hampered by presence of pouch young and as such her foraging radius and probability of encountering a trap are reduced (Cuttle 1982). In our study, both sexes in A. stuartii exhibited lower survival during mating and for females after mating, a phenomenon also observed in other studies (Wood 1970; Dickman 1986a; Leung 1999). This is most likely caused by the energy expenditure and stress of mating promiscuously, and for females the added burden of pregnancy and carrying pouch young. Finally, size of A. stuartii also fluctuated over the mating period. On average, males were found to increase in mass each month they were captured until the breeding period, after which they sharply declined (Fig. 2.5); females increased after the breeding period with the attachment and growth of young (Fig. 2.5). A similar trend was also observed by Woolley (1966) in A. agilis.

Breeding dates recorded in the present study for A. stuartii occurred about the same time as for other Springbrook A. stuartii (Parra Faundes 2014; the last 2 weeks

50 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

of August) and approximately 1 month later than those recorded for southern populations of A. stuartii (McAllan et al. 2006; Table 2.5). This is consistent with other findings where breeding timing in antechinus was found to vary by up to 4 months between localities within a species (Van Dyck 1982; Watt 1997; Marchesan and Carthew 2004). Dickman (1982) correlated such changes primarily with latitude and altitude. He observed that mating in a species typically occurred later in populations at lower latitudes and higher altitudes. The pattern with latitude and altitude, coupled with the synchronous nature of the breeding period, led researchers to believe photoperiodic cues may be the primary driver of reproduction timing in the genus (Dickman 1985). This was later corroborated by McAllan and Dickman (1986), who showed rate of change of photoperiod as the only reliable predictor of the timing of mating in most antechinus species. Lee et al. (1982) were first to suggest that such a synchronised mating period is timed at each locality such that lactation and weaning of offspring will coincide with spring insect flushes. More than three decades later, their theory is still best supported, most recently by Fisher et al. (2013), who found that at a continental scale, prey seasonality did indeed increase with latitude in dasyurids. They also found that the last month of lactation and weaning was the month of mean peak prey abundance in all monoestrous species (Fisher et al. 2013). During our research, pitfall trapping was conducted at each site in each season (before, during and after mating) for a parallel diet preference study and once examined will enable us to test whether this theory holds true for Springbrook antechinus.

Interestingly, we also found a significant difference in breeding timing within the Best of All A. stuartii population between years and between populations of A. stuartii at Best of All and Bilborough in 2015. Because the sites are only 2.5 km apart, are of similar elevation (approx. 950 m) and differences in timing also occurred at the same site between years, broad-scale geographical and climatic factors and associated differences in invertebrate peaks cannot fully explain such variation. However, the mating season of antechinus may also be partially influenced by factors such as intraspecific competition (Dickman 1985; Scott 1986) and local climatic and habitat factors. Such localised variation is not considered common in antechinus species (Lee et al. 1977; McAllan et al. 2006), but has been recorded in a few populations of Antechinus flavipes (Marchesan and Carthew 2004) and may be a factor at play in A. stuartii at Springbrook.

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 51

2.5.2 A. arktos Trapping and population dynamics A. arktos occurred in low abundance and returned erratic trap success outside the breeding season at both sites. The trap success in studies on other members of the dusky antechinus complex (A. mimetes, Antechinus swainsonii and Antechinus vandycki) was also low, often not higher than 0.5 % (Baker et al. 2015). However, there are exceptions, most notably in some populations of A. mimetes, which can reach 9.66 % (Table 2.5). Trap success of A. arktos in our study was also comparable to that found in previous studies of Antechinus godmani (0.27 %, Table 2.5), a similar-sized congener occurring at altitudes above 600 m in the wet tropics of north-east Queensland (Laurance 1992; Watt 1997). For A. godmani, low trappability is typically coupled with a low recapture rate of 30 % (Watt 1997), while our A. arktos had a comparably higher recapture rate of 53.8 % (Table 2.1). We also found a strong behavioural effect of trapping in A. arktos, with recapture probability high and residency of many individuals lengthy (particularly females; see ‘Population dynamics’ section), indicating that some individuals became conditioned to traps. This, as well as the 10 A. arktos individuals captured during 600 trap nights at Lamington in August 2015, indicates that the generally low trappability of A. arktos is most likely due to low abundance and not trap shyness, which may be the case for A. godmani (Watt 1997) and, conceivably, other members of the dusky antechinus complex.

Trap success and density of A. arktos was significantly greater at Best of All compared to Bilborough (Figs. 2.2 and 2.4). Site proximity and similar aspect eliminates broad-scale latitudinal, altitudinal and climatic effects from explaining differences in relative abundance. Instead, disparities in population density may be associated with local climatic factors and vegetation differences (Getz et al. 2001; Marchesan and Carthew 2004). Best of All habitat consists of minimally-disturbed old growth Gondwanan cloud forest, while the Bilborough site consists of approximately 100-year regrowth forest and non-remnant vegetation (Hall 1990). The Best of All site also contains a head water gully, where flax lily (Helmholtzia glaberrima) is abundant. H. glaberrima proved one of the most reliable indicators of A. arktos presence in an August 2015 survey of various high altitude habitats at nearby Lamington National Park. Thus, it may be that the Bilborough site, where Helmholtzia is absent, is on the

52 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

fringe of preferred A. arktos habitat and less favourable than the denser, older rainforest at Best of All and Lamington National Park sites. A detailed comparative floral assessment of these sites will allow such differences to be quantified and is currently in progress.

The trapping success and density of A. arktos increased in both our study sites from 2014 to 2015, which, assuming similar fecundity between years, suggests a higher survival rate of the juvenile cohort in late 2014 compared to late 2013. Rainfall has strong impacts on the survival, condition and overall abundance of A. flavipes (Lada et al. 2013), A. minimus (Sale et al. 2008), A. agilis and A. mimetes (Parrott et al. 2007) and A. stuartii at Springbrook (Parra Faundes 2014). High rainfall may potentially result in more favourable growing conditions for plants and also invertebrates (as prey), as has been recorded for wet tropical locations and species (Frith and Frith 1990; Watt 1997; Leung 1999). Our results suggest the months of November to March as critical for suckling, weaning and survival of dispersing A. arktos young. A plot of average monthly rainfall in 2013–2015 in comparison to the average monthly rainfall recorded at Springbrook for 34 years shows no clear pattern to explain our recorded differences in trapping success between years due to rainfall specifically during these times (Fig. 2.6). However, 2014 was a notably dry year overall, with rainfall lower than 2015 (and the 34-year average) in all months except for March and August, which ultimately may have influenced recruitment (Fig. 2.6). Understanding the exact relationship between seasons and invertebrate abundance would be critical in determining the drivers of year to year differences in A. arktos dynamics, which will impact long-term species survival.

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 53

Figure 2.6. Total rainfall in each month from 2013 to 2015 at Springbrook National Park, compared to the average mean rainfall per month for the last 34 years (Bureau of Meteorology 2015)

54Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Body mass The ranges in adult body mass recorded for A. arktos in our study (males 61– 113 g; females 38–67 g) are similar to those made on the first individuals captured at Springbrook National Park in 2013 by Baker et al. (2014) and those captured at Lamington National park in 2015 (ELG and I. Gynther, unpublished data).

Breeding The occurrence and synchrony of reproductive events in A. arktos at Springbrook closely paralleled the type one reproductive strategy exhibited by congeners (Marlow 1961; Woolley 1966; Williams and Williams 1982; McAllan et al. 2006). Temporal trends in the capture rate of A. arktos were not as pronounced as those observed in other species (Woolley 1966; Wood 1970; Reed and Wallis 1975; Wainer 1976; Friend 1985; Leung 1999; Marchesan and Carthew 2004); however, both abundance and trap success were generally higher during the breeding season than at all other times (Figs. 2.2 and 2.4). As per norms for the genus, A. arktos males were also found to increase in mass until the breeding period, after which they sharply declined (Fig. 2.5), while female mass increased after the breeding period with the attachment and growth of young (note n = three adult females, Fig. 2.5).

The A. arktos populations at Springbrook were estimated to mate during mid- September. There was no perceived difference in breeding times between sites, with ovulation estimated to occur at Best of All between the 10th and 14th of September and at Bilborough between the 13th and 15th of September. The population at Best of All appears to ovulate at a time when photoperiod is changing at a rate of 90 s/day.

However, because this information is based on only three individual females, we have little understanding of variance. Ovulation times are therefore expected to take place over a slightly longer time period than calculated in our study. Other members of the dusky antechinus complex (A. vandycki and A. mimetes, the two species then A. swainsonii) have shown no obvious photoperiodic correlation with time of ovulation (McAllan et al. 2006). More populations of A. arktos would need to be discovered and their ovulation times determined before any clear correlations can be made.

2.5.3 Between species comparisons A. arktos was significantly larger than A. stuartii at both sites and displayed clear-cut morphological differences (as described in Baker et al. 2014). A. stuartii was

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 55

significantly more abundant and accounted for 95.4 % of all antechinus captures at Springbrook. The type one reproductive pattern exhibited in both species was very similar; however, the timing of these events between sites in the second year of our study (when A. arktos young were recorded) was found to differ. In 2015, A. arktos ovulated between 10–19 days later than A. stuartii at Bilborough but only 1 to 9 days later at Best of All. Ovulation in A. arktos occurred during longer absolute day lengths and at a lower rate of change of photoperiod than A. stuartii. Reed and Wallis (1975) found that A. mimetes mated around the same time as co-occurring A. agilis. However, such similarity in mating timing between sympatric species is not the norm for the genus (Dickman 1982; McAllan et al. 2006). Where two species of antechinus co- occur, they should breed at different times (McAllan et al. 2006). Typically, the larger of two sympatric antechinus species mates first (Dickman 1982; McAllan et al. 2006). Dickman (1986a, b) predicted that in a species which breeds once annually, the earlier breeder may gain access to the first flush of invertebrates in spring, thus promoting larger size over time. This was not the case in the present study, where the significantly smaller A. stuartii was found to breed earlier than A. arktos. It is plausible that reproductive phase differences may arise from dietary preferences for different insect groups that peak in abundance at different times. That is, relative timing of breeding may in part be a reflection of temporal prey availability that the earlier breeder, in this case A. stuartii, prefers (Wainer 1976; Baker et al. 2012). Clearly, there is more happening in regard to competitive interactions between sympatric antechinus species than we currently understand or can easily predict. We are currently in the process of recalculating the rate of change of photoperiods for the genus in light of new species discoveries and as boundaries between several species are refined. Prey preference studies may help determine possible drivers of mating timing, since it has recently been demonstrated that access to predictable prey resources is a major factor determining both the synchronous and short, intensive mating periods characteristic of the group (Fisher et al. 2013).

2.5.4 Conservation of A. arktos We have now confirmed the presence of A. arktos at three proximate locations in Queensland. These populations are each a few kilometres apart, in high altitude (900–1150 m) ancient Gondwanan remnant cloud forest, situated at the summit of Tweed Caldera (just 40 km in diameter), the eroded peaks of which form the ‘Scenic

56 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Rim’ at the border of Queensland and New South Wales in eastern Australia. The region has a large and growing human population (Office of Economic and Statistical Research 2012) and high elevation sites with minimal habitat disturbance are rare and scattered (Cohen 2012). Even so, there are several other locations in the Scenic Rim which exhibit small pockets of appropriate, high elevation habitat that may support A. arktos.

Climate trends suggest recent decades have been the warmest in recorded human history (CSIRO and Bureau of Meteorology 2010; Hughes 2011). A combination of human disturbance, predation by feral cats (Felis catus) and foxes (Vulpes vulpes) and global warming may have caused these elusive mammals to retreat to the highest, coolest and wettest parts of their range (Baker et al. 2014). Our research suggests that unfortunately this newly discovered mammal species faces a serious risk of extinction in the coming decades.

2.6 ACKNOWLEDGEMENTS

This study was supported by grants from Australian Geographic, Holsworth Wildlife Research Endowment and Gold Coast City Council. The School of Earth, Environmental and Biological Sciences at the Queensland University of Technology provided access to vehicles and field equipment. ELG was financially supported by an Australian Postgraduate Award for doctoral research. We thank Ian Gynther (EHP) and Harry Hines (QPWS) who loaned us Elliott traps and freely shared their expertise. We thank Aila Keto (Australian Rainforest Conservation Society) and the rangers at Springbrook National Park (QPWS) who provided support and accommodation. Finally, we thank the residents of Springbrook for their interest and the many volunteers who ventured out with us to Springbrook to help collect this data. Comments provided by reviewers improved the manuscript.

2.7 COMPLIANCE WITH ETHICAL STANDARDS

All aspects of the study were carried out following the American Society of Mammologists’ guidelines (Sikes et al. 2011) and with approval from the Queensland University of Technology ethics department (approval number: 1400000005) and the Queensland Parks and Wildlife service (approval number: WITK14454114).

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 57

2.8 REFERENCES

Baker A, Mutton T, Hines H (2013) A new dasyurid marsupial from Kroombit Tops, south-east Queensland, Australia: the silver-headed antechinus, Antechinus argentus sp. nov. (Marsupialia: Dasyuridae). Zootaxa 3746:201–239.

Baker A, Mutton T, Mason E, Gray EL (2015) A taxonomic assessment of the Australian Dusky Antechinus Complex: a new species, the Tasman Peninsula Dusky Antechinus (Antechinus vandycki sp. nov.) and an elevation to species, the Mainland Dusky Antechinus (Antechinus mimetes new status). Memoirs of the Queensland Museum 59:75–126.

Baker A, Mutton T, Hines H, Van Dyck S (2014) The black-tailed Antechinus, Antechinus arktos sp. nov.: a new species of carnivorous marsupial from montane regions of the Tweed Volcano caldera, eastern Australia. Zootaxa 3765:101–133.

Baker A, Mutton T, Van Dyck S (2012) A new dasyurid marsupial from eastern Queensland, Australia: the buff-footed antechinus, Antechinus mysticus sp.nov. (Marsupialia: Dasyuridae). Zootaxa 3515:1–37.

Baker A, Van Dyck S (2013a) Taxonomy and redescription of the Atherton antechinus, Antechinus godmani (Thomas) (Marsupialia: Dasyuridae). Zootaxa 3670:401–439.

Baker A, Van Dyck S (2013b) Taxonomy and redescription of the fawn antechinus, Antechinus bellus (Thomas) (Marsupialia: Dasyuridae). Zootaxa 3613:201–228.

Baker A, Van Dyck S (2013c) Taxonomy and redescription of the yellow-footed antechinus, Antechinus flavipes (Waterhouse) (Marsupialia: Dasyuridae). Zootaxa 3649:1–62.

Baker A, Van Dyck S (2015) Taxonomy and redescription of the swamp antechinus, Antechinus minimus (Geoffroy) (Marsupialia: Dasyuridae). Mem Queensland Mus 59:127–170.

Barton K (2016) MuMIn: multi-model inference. R package version 1.15.6.

Bates D, Maechler M, Bolker B, Walker S (2015) Fitting linear mixed effects models using lme4. J Stat Softw 67:1–48.

58 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Bolker B, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MH, White JS (2009) Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol Evol 24:127–135.

Bureau of Meteorology (2015) Daily rainfall, Springbrook Road. Australian Government. Available at: http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=136&p_displa y_ type=dailyDataFile&p_stn_num=040607&p_startYear=

Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a practical information-theoretic approach. Springer Verlag, New York.

Calaby JH, Taylor JM (1981) Reproduction in two marsupial mice, Antechinus bellus and A. bilarni (Dasyuridae), of tropical Australia. J Mammal 62:329–341.

Cohen BE (2012) The scenic rim of south-eastern Queensland, Australia: a history of mid Cenozoic intraplate volcanism. Episodes 35:103– 109.

Crawley MJ (2002) Statistical computing, an introduction to data analysis using s-plus. John Wiley & Sons, Ltd, West Sussex.

CSIRO and Bureau of Meteorology (2010) State of climate, technical report. Retrieved from http://www.csiro.au/resources/State-of-theClimate.html. Commonwealth Scientific and Industrial Research Organization and Australian Bureau of Meteorology. Melbourne, Australia.

Cuttle P (1982) Life history strategy of the dasyurid marsupial tapoatafa. In: Archer M (ed) Carnivorous marsupials Vol 1. Sydney royal zoological society of New South Wales, Sydney, pp. 13–21.

Dickman CR (1980) Ecological studies on Antechinus stuartii and Antechinus flavipes (Marsupialia: Dasyuriade) in open forest and woodland habitats. Austral Zool 20:433–446.

Dickman CR (1982) Some ecological aspects of seasonal breeding in antechinus (Daysuridae Marsupialia). In: Archer M (ed) Carnivorous marsupials Vol 1. Sydney royal zoological society of New South Wales, Sydney, pp. 139–150.

Dickman CR (1985) Effects of photoperiod and endogenous control on timing of reproduction in the marsupial genus Antechinus. J Zool (Lond) 206:509–524.

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 59

Dickman CR (1986a) An experimental study of competition between two species of dasyurid marsupials. Ecol Monogr 56:221–241.

Dickman CR (1986b) An experimental manipulation of the intensity of interspecific competition: effects on a small mammal. Oecologia 70: 536–543.

Dickman CR (1988) Body size, prey size and community structure in insectivorous mammals. Ecology 69:569–580.

Fisher DO, Dickman CR, Jones ME, Blomberg SP (2013) Sperm competition drives the evolution of suicidal reproduction in mammals. PNAS 110:17910–17914.

Friend GR (1985) Ecological studies of a population of Antechinus bellus (Marsupialia: Dasyuridae) in tropical northern Australia. Aust Wildl Res 12:151–162.

Frith D, Frith C (1990) Seasonality of litter invertebrate populations in an Australian upland tropical rain forest. Biotropica 22:181–190.

Getz LL, Hofmann JE, McGuire B, Dolan TW (2001) Twenty-five years of population fluctuations of Microtus ochrogaster and M. pennsylvanicus in three habitats in east-central Illinois. J Mammal 82:22–34.

Gilfillan SL (2001) An ecological study of a population of Pseudantechinus macdonnellensis (Marsupialia: Dasyuridae) in central Australia II. Population dynamics and movements. Wildl Res 28:481–492.

Glen AS (2008) Population attributes of the spotted-tailed (Dasyurus maculatus) in north-eastern New South Wales. Austral J Zool 56:137–142.

Gray EL (2013) A field study of breeding biology in a new species of carnivorous marsupial, the Buff-footed Antechinus, A. mysticus. Honours thesis, Queensland University of Technology, Brisbane, Queensland, Australia.

Green RH (1972) The murids and small dasyurids in .5. Antechinus swainsonii (Waterhouse 1840). Rec Queen Vic Mus 46:1–13.

Hall P (1990) Springbrook, where the clouds touch the earth. Published by P. Hall, Queensland, Australia.

Hughes L (2011) Climate change and Australia: key vulnerable regions. Reg Environ Chang 11:189–195.

60 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Krajewski C, Torunsky R, Sipiorski T, Westerman M (2007) Phylogenetic relationships of the dasyurid marsupial genus Murexia. J Mammal 88:696–705.

Krebs CJ (1966) Demographic changes in fluctuating populations of Microtus californicus. Ecol Monogr 36:239–273.

Lada H, Thomson JR, Cunningham SC, Mac Nally R (2013) Rainfall in prior breeding seasons influences population size of a small marsupial. Austral Ecol 38:581– 591.

Laurance WF (1992) Abundance estimates of small mammals in Australian tropical rainforest: a comparison of four trapping methods. Wildl Res 19:651–655.

Lee AK, Bradley AJ, Braithwaite RW (1977) Corticosteroid levels and male mortality in Antechinus stuartii. In: Stonehouse. B, Gilmore D (eds) The biology of marsupials. Macmillan Press, London, pp. 209–220.

Lee AK, Woolley PA, Braithwaite R (1982) Life history strategies of dasyurid marsupials. In: Archer M (ed) Carnivorous marsupials Vol 1. Sydney Royal Zoological Society of New South Wales, Sydney, pp. 1–11.

Leung L (1999) Ecology of Australian tropical rainforest mammals. I. The Cape York antechinus, Antechinus leo (Dasyuridae: Marsupialia). Wildl Res 26:287–306.

Marchesan D, Carthew SM (2004) Autecology of the yellow-footed antechinus (Antechinus flavipes) in a fragmented landscape in southern Australia. Wildl Res 31:273–282.

Marlow B (1961) Reproductive behaviour of the marsupial mouse, Antechinus flavipes (Marsupialia) and the development of the pouch young. Austral J Zool 9:203– 220.

McAllan BM, Dickman CR (1986) The role of photoperiod in the timing of reproduction in the dasyurid marsupial Antechinus stuartii. Oecologia 68:259–264.

McAllan BM, Joss JM, Firth BT (1991) Phase delay of the natural photoperiod alters reproductive timing in the marsupial Antechinus stuartii. J Zool 225:633–646.

McAllan BM, Dickman CR, Crowther MS (2006) Photoperiod as a reproductive cue in the marsupial genus Antechinus: ecological and evolutionary consequences. Biol J Linn Soc 87:365–365.

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 61

Mutton T (2016) Evolutionary biology of the Australian carnivorous marsupial Genus Antechinus, with particular focus on two species, A. mysticus and A. subtropicus. Submitted Ph.D. dissertation. Queensland University of Technology.

Office of Economic and Statistical Research (2012) Population and dwelling profile, Scenic Rim Regional Council. Retrieved from http://www.qgso.qld.gov.au/products/reports/pop-housing-profileslga/pop-housing- profile-scenic-rim.pdf

Parra Faundes DV (2014) Sex allocation, terminal investment and the effects of environmental constraints on maternal investment in Subtropical Antechinus. Ph.D. dissertation, University of Queensland, Queensland, Australia.

Parrott ML, Ward SJ, Temple-Smith PD, Selwood L (2007) Effects of drought on weight, survival and breeding success of agile antechinus (Antechinus agilis), dusky antechinus (A. swainsonii) and bush rats (Rattus fuscipes). Wildl Res 34:437–442.

Pinheiro J, Bates D, DebRoy S, Sarkar D, and R Core Team (2015) Nlme: linear and nonlinear mixed effects models. R package version 3.1– 122.

Queensland Herbarium (2013) Regional Ecosystem Description Database (REDD). Version 6.1 (February 2013). Queensland Department of Science, Information Technology, Innovation and the Arts, Brisbane.

R Development Core Team (2014) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna www.R-project.org/.

Reed GF, Wallis RL (1975) Studies of Antechinus swainsonii and other small mammals in an area of Sherbrooke Forest Park. Vic Nat 92: 84–90.

Sale MG, Wilson BA, Arnould JY (2008) Factors influencing population dynamics in island and mainland populations of the Swamp Antechinus (Antechinus minimus). Austral J Zool 56:187–194.

Scott MP (1986) The timing and synchrony of seasonal breeding in the marsupial Antechinus stuartii: interaction of environmental and social cues. J Mammal 67:551– 560.

Selwood L (1980) A timetable of embryonic development of the dasyurid marsupial Antechinus stuartii (Macleay). Austral J Zool 28:649–668.

62 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Selwood L (1985) Synchronization of oestrus, ovulation and birth in female Antechinus stuartii (Marsupialia: Dayuridae). Austral Mammal 8:91–96.

Sikes RS, Gannon WL, and the Animal Care and Use Committee of the American Society of Mammalogists (2011) Guidelines of the American Society of Mammalogists for the use of wild mammals in research. J Mammal 92:235–253.

Smith GC (1984) The biology of the yellow-footed antechinus, Antechinus flavipes (Marsupialia: Dasyuridae), in a swamp forest of Kinaba Island, Cooloola, Queensland. Aust Wildl Res 11:465– 480.

Statham HL (1982) Antechinus stuartii (Dasyuridae, Marsupialia) diet and food availability at Petroi, north-eastern New South Wales. In: Archer M (ed) Carnivorous marsupials Vol 1. Sydney Royal Zoological Society of New South Wales, Sydney, pp. 151–163.

Van Dyck S (1982) The relationships of Antechinus stuartii and A. flavipes (Dasyuridae, Marsupialia) with special reference to Queensland. In: Archer M (ed) Carnivorous marsupials. Sydney Royal Zoological Society of New South Wales, Sydney, pp. 723– 766.

Van Dyck S, Strahan R (2008) The mammals of Australia. New Holland Publishers, Sydney.

Wainer JW (1976) Studies of an island population of Antechinus minimus (Marsupialia: Dasyuridae). Austral J Zool 19:1–7.

Watt A (1997) Population ecology and reproductive seasonality in three species of Antechinus (Marsupialia: Dasyuridae) in the wet tropics of Queensland. Wildl Res 24:531–539.

WhiteGC (2005) Correcting wildlife counts using detection probabilities. Wildl Res 32:211–216.

Williams BK, Nicols JD, ConroyMJ(2002) Analysis and management of animal populations. Academic Press, California.

Williams R, Williams A (1982) The life cycle of Antechinus swainsonii (Dasyuridae, Marsupialia). In: Archer M (ed) Carnivorous marsupials Vol 1. Sydney Royal Zoological Society of New South Wales, Sydney, pp. 89–95.

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 63

Wilson BA (1986) Reproduction in the female dasyurid Antechinus minimus maritimus (Marsupialia: Dasyuridae). Austral J Zool 34: 189–197.

Wood DH (1970) An ecological study of Antechinus stuartii (Marsupialia) in a south-east Queensland rain forest. Austral J Zool 18:185–207.

Woolley PA (1966) Reproduction in Antechinus spp. and other dasyurid marsupials. Symp Zool Soc Aust 15:281–294.

64 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii) 65

66 Chapter 2: Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii)

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Emma L. Gray, Chris J. Burwell, and Andrew M. Baker

Submitted on 28 June 2016, accepted 28 September 2016, and published online 14 November 2016 in Australian Journal of Zoology.

3.1 ABSTRACT

The endangered black-tailed dusky Antechinus, Antechinus arktos, was only described in 2014, so most aspects of its ecology are unknown. We examined diet composition and prey selection of A. arktos and a sympatric congener, the northern form of A. stuartii, at two sites in Springbrook National Park. Overall, taxa from 25 invertebrate orders were identified in the diets from 252 scat samples. Dietary components were similar for each species, but A. arktos consumed a higher frequency and volume of dipteran larvae and Diplopoda, while A. stuartii consumed more Coleoptera, Lepidoptera, Orthoptera and Isopoda. Both species of Antechinus had a higher percentage of ‘empty’ scats (devoid of any identifiable invertebrate material) in 2014 compared to 2015. The former was a drier year overall. Lower rainfall may have reduced abundance and diversity of arthropod prey causing both species to supplement their diet with soft-bodied prey items such as earthworms, which are rarely detected in scats. Comparison of prey in scats with invertebrate captures from pitfall traps showed both species to be dietary generalists, despite exhibiting distinct preference and avoidance of certain prey categories. The ability of an endangered generalist marsupial to switch prey may be particularly advantageous considering the

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 67

anticipated effects of climate change on Gondwanan rainforests during the mid–late 21st century.

Keywords Diet, prey selection, invertebrates, faeces, endangered species

3.2 INTRODUCTION

Knowledge of a species’ diet and dietary strategy is fundamental to understanding its ecology, including life history traits, survival, population fecundity, distribution and activity. A dietary specialist selects (or avoids) specific prey regardless of the proportions at which they occur in the environment (Ross et al. 2010). A specialist dietary strategy is advantageous in terms of increased feeding efficiency and potentially reducing dietary overlap under interspecific competition (McArthur and Pianka 1966; Futuyma and Moreno 1988). However, specialisation may also lead to an increased dependence on preferred prey items, resulting in the species being less adaptable to potential changes in resource abundance and / or availability (Boyles and Storm 2007). Generalist foragers consume a variety of prey in proportions relative to their availability in the environment (Allison et al. 2006). They display more flexible behaviour that allows them to switch foraging strategies more easily and therefore adapt to changes in resource composition with relative ease (O’Donoghue et al. 1998). Thus, the extent of a species’ dietary specialisation may influence its extinction risk (Boyles and Storm 2007).

Small carnivorous marsupials (Family Dasyuridae) are perfect study organisms in which to measure dietary composition, overlap and selection. Currently there are 56 species of small dasyurids (< 500 g), which occur in most terrestrial habitats in Australia and New Guinea, often in sympatry (Crowther and Blackett 2003; Dickman 2014). They can be locally abundant, but also contain several rare, cryptic species such as the white-footed (Sminthopsis leucopus) and brush-tailed Phascogale (Phascogale tapoatafa) (Van Dyck and Strahan 2008). Dietary studies conducted on 25 small dasyurid species have shown them to be generalist carnivores, taking a broad range of predominantly invertebrate prey items (Hall 1980; Statham 1982; Dickman 2014). However, vertebrates such as amphibians (Chen et al. 1998), small lizards (Green 1989; Lunney et al. 2001; Pearce 2016), birds (Chen et al. 1998; Pavey et al. 2016) and mammals (Ward 2000; Lunney et al. 2001; Pavey et al. 2016) have been recorded infrequently in the diet of some dasyurids.

68 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

The dasyurid genus Antechinus has been the focus of numerous dietary studies (e.g., Hall 1980; Lunney et al. 2001; Allison et al. 2006). Although these studies provide a strong foundation for the studies of diet in Antechinus spp. further research concerning dietary composition, prey availability and prey selection are required to fully understand the ecology of individual species and their sympatric relationships. This is especially true given that five new species of Antechinus have been discovered in the past few years (Baker et al. 2012; Baker et al. 2013; Baker et al. 2014; Baker et al. 2015).

Here we present the first study of diet in one of these new taxa, the black-tailed dusky Antechinus, A. arktos, a member of the dusky Antechinus species complex and formerly considered a northern outlier of A. mimetes (Baker et al. 2015). Antechinus arktos is a small (38-120 g) carnivorous marsupial, which once inhabited a range of sites on the slopes of the eroded Tweed Shield Volcano caldera that straddles the border of New South Wales (NSW) and Queensland (Qld) on Australia’s mid-east coast (Baker et al. 2014). Today, however, A. arktos is only known from several proximate sites at the summit of the caldera (in Springbrook and Lamington National Parks), where they occur at low density in sympatry with a little-studied northern phylogenetic clade of brown Antechinus, A. stuartii (Gray et al. 2016). Antechinus arktos has thus been listed as Endangered under the Nature Conservation Act 1992 (Qld) and the Threatened Species Conservation Act 1995 (NSW). One explanation for the species’ apparent range contraction is climate change. In the study region, mean annual maximum temperatures have increased by 1° C and total rainfall has dropped by 76 mm between 1950 and 2003 (Hennessy et al. 2004). Under a changing climate, declines may have been caused by withdrawal of A. arktos to match the upward altitudinal shift of suitable (wetter, colder) habitat (Baker et al. 2014). This warming and drying trend is expected to continue. Climatic changes predicted for the Gondwana Rainforests, of which Springbrook and Lamington National Parks are a part, include: an average temperature increase of 1.3° C ± 0.6° C by 2030, a drop in average rainfall of 3.5% ± 11%, more severe and frequent extreme weather events (e.g., drought and flash flooding), and a rise in the cloud base set to reduce mist and therefore affect the availability and consistency of moisture (Hutley et al. 1997; McJannet et al. 2007; Australian National University 2009). The forecast climate change is of particular concern for many endemic cold-adapted plant species (e.g., Antarctic beech,

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 69

Nothofagus moorei; Taylor et al. 2005) and vertebrates, especially amphibians, which are particularly sensitive to temperature and moisture changes (Australian National University 2009). Hagger et al. (2013) observed that Antechinus spp. would also be particularly vulnerable to climate change if the peak of insect abundance was altered to no longer coincide with timing of lactation. Insect abundance is controlled by climate variables such as rainfall and temperature, while Antechinus breeding is regulated by photoperiod and thereby fixed in time, regardless of food availability (McAllan et al. 2006; Hagger et al. 2013).

In this study, we analysed scats of sympatric A. arktos and A. stuartii collected during a two-year, parallel, mark-recapture project at Springbrook National Park and an intensive single fieldtrip to Lamington National Park (Gray et al. 2016). We also sampled available prey in the environment during three seasons (autumn, winter and spring) via pitfall trapping. Specifically, we aimed to: 1) compare the degree of dietary overlap between and within species among sexes, sites, seasons and years and 2) determine the degree of prey specialisation in each species by comparing the categories of prey consumed with prey available in the environment. We hypothesised that, like their congeners, A. arktos and A. stuartii would be dietary generalists and take prey largely in proportion to their availability in the environment. However, we predicted that there would be quantitative differences in diet between the two Antechinus spp., sites, and seasons. This knowledge will improve our ecological understanding of the endangered A. arktos and the possible impacts of climate change on its food resources. The study will also contribute to our scientific understanding of sympatric relationships within this iconic genus of semelparous marsupials.

3.3 MATERIALS AND METHODS

3.3.1 Study sites The study sites were located at the rim of the Tweed Shield Volcano caldera in World Heritage listed Gondwanan rainforest, which provides the principal habitat for many primitive, relict, endemic and threatened species of flora and fauna (Hunter 2003). Faecal material was primarily sourced from Best of All Lookout (28.2415°S, 153.2640°E) and Bilborough Lookout (28.2341S, 153.2897°E) at altitudes of approximately 950 m. Both sites are located within Springbrook National Park, 100 km south of Brisbane, Queensland’s capital, situated on the east coast of Australia.

70 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

The site at Best of All Lookout consists of complex notophyll vine forest, Regional Ecosystem (RE) 12.8.5, and simple microphyll fern forest with Antarctic beech, Nothofagus moorei (RE 12.8.6). The site at Bilborough Lookout consists of regrowth RE 12.8.5 and 12.8.6 surrounded by non-remnant vegetation and is notably more open. At the commencement of the study, these were the only two known present-day sites of occurrence for A. arktos. Antechinus arktos co-occurs with A. stuartii at both sites. Additional comparative faecal material for both A. arktos and A. stuartii was collected opportunistically during a 600 trap night fieldtrip to Lamington National Park (28.2513°S, 153.1760°E), located 8 km (straight-line distance) west of Springbrook plateau on steep slopes at altitudes of approximately 1,150 m.

3.3.2 Animal trapping Faecal samples were collected throughout the course of a parallel two-year mark- recapture growth and breeding timing study at Springbrook National Park (see Gray et al. 2016). Trapping was conducted monthly between April and October (2014-2015) using Elliott traps (Type A, 33 x 10 x 9 cm; Elliott Scientific, Victoria, Australia). At each site, four parallel line transects were established, each with a total of 25 tags, positioned at 8 m intervals along the line (total length = 200 m). Two traps were placed at each tag along the line, spaced a minimum distance of 1 m apart (totalling 200 traps per site) and lines were separated by approximately 10 m. During each month, trapping was conducted for a total of six nights (three consecutive nights per site), totalling 600 trap nights at each site per month. Antechinus spp. captured were identified to species, weighed, sexed, ear clipped (analysed in growth data [Gray et al. 2016] and genetics [parallel study]), assessed for reproductive condition and micro-chipped with a Passive Integrated Transponder (PIT) tag for recapture identification. Individuals were released immediately after processing, at the point of capture. Scats were collected from the floor of each trap and stored in 70 % ethanol. Due to a high capture rate of Antechinus and non-target rodents each day, traps were unable to be cleaned so scats were taken on the first day of trapping and on subsequent days only if the trap was recorded as being previously unoccupied.

3.3.3 Dietary analysis A sample consisted of a single faecal pellet collected from one individual, from one trap night. Duplicate samples were taken from every individual each night. However, only one randomly selected (unique) sample was ever examined. An effort

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 71

was made to sample as evenly as possible between species and across sites, sexes, years and months; however, because of the large difference in trap success between the two species (A. stuartii accounted for 95.4 % of all Antechinus captures [Gray et al. 2016]) this proved impossible. Therefore, all 80 unique A. arktos samples (42 female; 38 male) collected were examined; while a subset of 172 unique A. stuartii samples (105 female; 67 male) were examined out of a total of 200 samples.

Each faecal pellet was placed in a petri dish, immersed in 70 % ethanol and teased apart using fine forceps. The faecal pellet was then systematically searched for identifiable prey material under a dissecting microscope. Prey fragments were identified to ordinal level using key morphological features and with reference to the invertebrate collection held at the Queensland Museum. However, order Hemiptera was classified into suborders Heteroptera and Auchenorrhyncha. Orders Coleoptera, Lepidoptera and Diptera were also divided into adult and larval life stages, as soft- bodied larvae may be preferentially consumed by some Antechinus spp. (Lunney et al. 2001; Sale et al. 2006). As Formicidae and staphylinid (rove) were found abundantly in pitfall traps, they were given their own category so as not to overestimate the importance of the orders Hymenoptera and Coleoptera, respectively. Identified prey fragments were then grouped together by order, counted and the percentage volume visually estimated to the nearest 5 % as per Pavey et al. (2009) and Mason et al. (2015). Unidentified material was not included in percentage volume estimates. Taxa that contributed less than 5 % by volume were also not included in percentage volume estimates but were recorded as being present, as was material assumed to have only been incidentally ingested (e.g., ticks, mites, lice and small quantities of vegetation and hair).

3.3.4 Food availability Pitfall trapping was used to measure composition of potential prey items available to both species to permit estimates of prey preference. Pitfall traps are the most appropriate method of capturing mobile, ground-dwelling invertebrates (Luff 1975; Work et al. 2002) and are therefore commonly used to estimate prey availability for Antechinus spp. which are assumed to predominately hunt mobile arthropods on the ground surface (Fisher and Dickman 1993). Unfortunately, due to time and labour constraints, pitfall trapping could not be conducted in 2014. However, pitfall trapping was conducted in 2015 at both sites during autumn (April), winter (August) and spring

72 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

(October), coinciding with regular Elliott trapping. A pitfall trap was set at every fifth tag along each transect (five pitfalls per transect, totalling 20 pitfalls per site). Each pitfall trap consisted of a cylindrical plastic jar (6.5 cm wide by 8 cm deep) buried in the soil so that the lip was level with the ground surface. Traps were partially filled with a solution of 70 % ethanol and 30 % glycerol to reduce evaporation. A random subsample of 24 pitfall traps (four traps from each site per season from a possible 120 traps) was then examined. All invertebrates in a pitfall trap were identified to ordinal (or special category) level with the aid of reference texts (CSIRO 1991; Zborowski and Storey 2010). Because we only wished to identify and count potential edible prey, data on invertebrates less than 2 mm in length that were likely ingested incidentally were excluded from analyses (as per Green 1989; Fisher and Dickman 1993).

3.3.5 Statistical analysis We analysed diet in terms of both percent frequency of occurrence and relative volume of each prey category in Antechinus scats, while prey availability was examined in terms of both frequency of occurrence and abundance of each prey category in pitfall traps. Frequency of occurrence of each prey category was calculated as the number of scats / pitfall traps containing a particular prey category divided by the total number of scats / pitfall traps analysed, grouped by species, sex and site as required. Volume was calculated as the mean percent volume of each prey category in the scats, grouped by species, sex and site as required. Abundance was calculated as the total number of individuals of each prey category identified in each pitfall trap, grouped by season and site. Where possible, we also considered the nutritional value of prey (i.e., fat and protein content) using available literature, as consumption of particular prey categories may be influenced by the nutritional requirements of Antechinus at different stages of their life cycle (Lunney et al. 2001; Hume 2003).

To test for differences in the prey composition of scats (between species and among various categories within species) based on relative volume of the various prey categories, we used a permutational multivariate analysis of variance (PERMANOVA) based on a Bray-Curtis similarity matrix. The percent volume values of each prey category were fourth root transformed to reduce the contribution of dominant prey categories in relation to less abundant ones (Anderson et al. 2008). A dummy (prey category) variable was also added to the dataset due to the large number of zeroes recorded in 2014 so that two samples containing no species were not left

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 73

undefined and sparse samples did not fluctuate wildly (Clarke et al. 2014). Because the experimental design was necessarily unbalanced (see dietary analysis) a type III sums of squares partition was used by default as per Anderson et al. (2008). To statistically test for differences in prey category composition of the pitfall traps (between seasons and among sites within each season), we used a PERMANOVA based on the Bray-Curtis similarity matrix. The abundance data were square root transformed. These analyses were conducted in Primer 7 (Clarke and Gorley 2015), with 9, 999 permutations used to calculate pseudo F statistics, and P-values <0.05 taken to be statistically significant. The relative volume of prey categories in scats was also visually compared between species and the factors of sex, site, year and season, using shade plots (Clarke et al. 2014).

In order to determine whether the Antechinus spp. showed some degree of dietary specialisation, the frequency of occurrence of each prey category identified in the diet was compared to its frequency of occurrence in the pitfall traps using one- sided Fisher’s exact tests (as per Allison et al. 2006, for low sample sizes). This was done separately for each species and each season (autumn, winter and spring). The frequencies of prey categories identified within the scats were the observed values and the frequencies of categories identified within pitfall samples were the expected values. Only invertebrates recorded in scats of Antechinus were considered ‘available prey’, as per Green (1989) and Fisher and Dickman (1993). Therefore, annelids which were frequently caught in pitfall traps but not detected in scats were excluded from analysis. Adult Lepidoptera, which could not be sampled using pitfall traps, and rare categories that contributed little to the overall dietary composition of Antechinus, such as Collembola, staphylinid beetles, Ostracoda, Trichoptera and Neuroptera, were also excluded.

3.4 RESULTS

3.4.1 Overall diet In total, 252 faecal pellets (samples) were examined, with 25 invertebrate orders recognised from five subclasses: Hexapoda (springtails and insects), Chelicerata (arachnids), Crustacea (amphipods, isopods, and ostracods), Myriapoda (millipedes and centipedes) and Gastropoda (terrestrial snails and semi-slugs) (Table 3.1). Antechinus arktos most frequently consumed dipteran larvae (33.3 %), Araneae (30.6

74 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

%), Amphipoda (22.2 %) and Diplopoda (16.7 %); A. stuartii most frequently consumed Coleoptera (30.9 %), Araneae (22.2 %), Isopoda (19.1 %) and Amphipoda (21 %) (Table 3.1). Three orders, Neuroptera, Trichoptera and Ostracoda, were only found in A. stuartii scats; however, they all occurred at very low frequencies, with Neuroptera and Trichoptera at less than 5 % volume (Table 3.1). Collembola was only found in A. arktos scats, but always occurred at less than 5 % volume (Table 3.1). Hair was also present in small amounts in 3.1 % of A. stuartii scats and 12.5 % of A. arktos scats, but was assumed to be ingested during grooming and excluded from analysis (Table 3.1). Vertebrate remains (feather fragments) were also evident in one June 2015 sample (A. stuartii), although these may have been ingested incidentally, accounting for less than 5 % volume of the scat. No bones were found in scats of either species. A large number of scats in the first year of our study contained no identifiable arthropod or animal material (A. arktos 63.6 %; A. stuartii 57.8 %). There were markedly fewer such ‘empty’ scats examined in the second year (A. arktos 23.1 %; A. stuartii 22.8 %). Both species also had a higher than expected frequency of plant material (seeds, leaf and stem fragments) in their diet, occurring in 23.6 % of A. arktos samples and 9.9 % of A. stuartii samples across both years (Table 3.1). However, such material was assumed to be ingested incidentally while foraging and was not included in percentage volume estimates.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 75

Table 3.1. Overall dietary composition of A. arktos and A. stuartii at Springbrook National Park between 2014 and 2015 expressed as frequency of occurrence (%) and average volume (%).

Note that the percent volume columns do not add up to 100 % due to the number of empty scats included in the averaged dataset.

3.4.2 Dietary differences between and within species and sexes The composition of invertebrates in scats differed between species (F = 7.79, d.f. = 1, p <0.001). In general, a greater number of invertebrate categories were consumed by A. stuartii (Figure 3.1). Antechinus stuartii consumed a considerably higher volume of Coleoptera, Lepidoptera, Orthoptera and Isopoda while A. arktos consumed a higher volume of dipteran larvae and Diplopoda (Figure 3.1). However, both species consumed Araneae () in high volume (Figure 3.1).

76 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

The composition of invertebrates in the diet also differed between sexes (within and between species) (F = 5.01, d.f. = 3, p <0.001). Post hoc pair-wise comparison found all combinations of species and sex differed in dietary composition (Table 3.2). In general, female scats of both species contained a greater number of invertebrate orders than males; however, only males of both species were found to consume gastropods (Figure 3.2).

Table 3.2. Results of permutational multivariate analysis of variance (PERMANOVA) and post hoc pair-wise tests comparing the relative volumes of food categories between species, as well as between years, seasons and sexes within species

f / t= the pseudo-F or pseudo-T test statistic used, d.f= degrees of freedom, p= p value of the test, unique perms= the number of unique values of the test statistic obtained under permutation. AF= A. arktos female, AM= A. arktos male, SF= A. stuartii female, SM= A. stuartii male, year 1=2014, year 2=2015.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 77

Figure 3.1. Shade plot showing the relative contribution (by volume) of each prey category identified in scats of A. arktos and A. stuartii on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent).

78Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Figure 3.2. Shade plot denoting the relative contribution (by volume) of each prey category identified in scats of A. arktos and A. stuartii males and females on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent).

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 79

3.4.3 Dietary differences between years There was a significant difference in dietary composition between species within and between years (F = 11.92, d.f. = 3, p = 0.001). The shade plots (Figures 3.3 and 3.4), show less diversity in scats of both species in 2014 compared to 2015. There was also a significant difference in dietary composition between years when year was nested within species and sex (F = 6.90, d.f. = 7, p = 0.001). Post hoc pair-wise comparison (Table 3.2) showed dietary composition differed between the majority of groups. However, there was no difference in dietary composition between male A. arktos in the first and second year of study and in A. stuartii males and females in 2015.

80 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Figure 3.3. Shade plot denoting the relative contribution (by volume) of each prey category identified in scats of A. arktos between years of study (2014 and 2015) on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent).

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 81

Figure 3.4. Shade plot denoting the relative contribution (by volume) of each prey category identified in scats of A. stuartii between years of study (2014 and 2015) on a grey scale. Prey categories with high relative volumes are displayed in darker shades, while categories with lower relative volumes are displayed in lighter shades (or white if absent).

82Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

3.4.4 Dietary difference between sites There was no difference in dietary composition in either species between Best of All and Bilborough sites (A. arktos: F = 1.82, d.f. = 1, p = 0.119; A. stuartii: F = 1.64, d.f. = 1, p = 0.138). Therefore, the data for both Springbrook sites in August 2015 were pooled together separately for each species and compared with the data for each species obtained from the more geographically distant Lamington National Park in August 2015. There was a difference in diet composition of species in Springbrook and Lamington National Parks (F = 2.51, d.f. = 1, p = 0.03). The diet of A. arktos differed between sites (t = 1.88, p = 0.013), while A. stuartii diet did not (t = 1.21, p = 0.212). Antechinus arktos at Lamington consumed a greater number of prey orders (Lamington 13; Springbrook eight) (Figure 3.5a). However, dipteran larvae still dominated the diet of A. arktos in terms of both frequency and volume at both sites (Figure 3.5a). In A. stuartii, Coleoptera was consumed less frequently at Lamington but still dominated the diet in terms of volume at both sites (Figure 3.5b).

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 83

Figure 3.5. Frequency of Occurrence (%) of each prey category identified in scats of (a) A. arktos and (b) A. stuartii at Springbrook and Lamington National Parks.

84 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

3.4.5 Dietary difference between seasons There was no difference in scat composition between seasons when examined nested within species and sex (F = 1.54, d.f. = 4, p = 0.0668). There was also no difference in scat composition in either species between seasons for 2015 alone (F = 0.57, d.f. = 4, p = 0.953).

3.4.6 Prey availability Overall, 20 invertebrate categories were recorded from pitfall traps in 2015. There was no significant difference in invertebrate composition between the Best of All and Bilborough sites (F = 1.91, d.f. = 1, p = 0.0583). However, invertebrate composition did significantly differ between seasons (F = 3.07, d.f. = 2, p <0.001). The autumn assemblage differed from both the winter and spring assemblages (t = 2.06, p <0.001; t = 1.91, p <0.001, respectively); however, there was no difference between winter and spring assemblages (t = 1.32, p = 0.1146). The autumn assemblage had the highest number of prey categories. The categories of Collembola, staphylinid beetles, Diptera, Diptera larvae and Coleoptera all showed consistently high frequencies of occurrence across seasons, while the frequencies of Amphipoda, Isopoda and Araneae dropped dramatically in winter (Figure 3.6).

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 85

(a)

urrence urrence (%) Frequency occof

Invertebrate category

Figure 3.6. Frequency of Occurrence (%) of each prey category identified in A. arktos (AA) and A. stuartii (AS) scats and in pitfalls (PF) during A) autumn, B) winter, and C) Spring.

86 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

(b)

Frequency occurrenceof (%)

Invertebrate category

Figure 3.6. Frequency of Occurrence (%) of each prey category identified in A. arktos (AA) and A. stuartii (AS) scats and in pitfalls (PF) during A) autumn, B) winter, and C) Spring.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 87

(c)

Frequency occurrenceof (%)

Invertebrate category

Figure 3.6. Frequency of Occurrence (%) of each prey category identified in A. arktos (AA) and A. stuartii (AS) scats and in pitfalls (PF) during A) autumn, B) winter, and C) Spring.

3.4.7 Relationship between diet and prey availability Overall, there were numerous deviations from observed (in scats) and expected (in pitfalls) prey frequencies (Table 3.3). Both Antechinus species did not consume Coleoptera (beetles) in proportions relative to their high availability in the environment in any season; however, A. stuartii ate Coleoptera more frequently than A. arktos. Diptera (adults) and Formicidae (ants) were also consumed less compared to their availability, except the latter during spring when Formicidae were not sampled in

88 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

pitfalls. Dipteran larvae were frequent in pitfalls in all seasons (particularly spring) and A. arktos consumed them at or above their frequency while A. stuartii consumed them less than expected (especially during spring). Antechinus arktos consumed more Diplopoda than expected in all seasons (particularly autumn), and both species also consumed Chilopoda more frequently than expected.

Table 3.3. Exact test of goodness of fit results comparing the occurrence of each prey category in pitfall traps and in the diets of A. arktos and A. stuartii.

Taxa categories A. arktos A. stuartii Autumn Winter Spring Autumn Winter Spring Isopoda - *** +*** N/S -*** +*** +*** Blattodea N/S N/S N/S N/S N/S N/S Amphipoda -* N/S N/S -*** N/S N/S Orthoptera N/S N/S N/S N/S +*** N/S Heteroptera N/S -*** -* N/S -*** -*** Diptera -* -*** -*** -*** -*** -*** Diptera larvae N/S +** N/S -* N/S -*** Lepidoptera larvae N/S N/S +*** +* N/S +*** Coleoptera -*** -*** -*** -*** -* -** Coleoptera larvae N/S N/S N/S -* N/S -** Formicidae -** -*** +*** -*** -*** +*** Araneae -* N/S N/S -*** N/S -* Chilopoda N/S +*** +*** N/S +*** +*** Diplopoda +** +*** +*** N/S +*** N/S Mollusca N/S +*** N/S N/S +*** +*** + denotes that frequency in the diet is greater than that in pitfalls. – denotes frequency in the diet less than that in pitfalls. N/S denotes no difference between diet and pitfalls. *P<0.05; **P<0.01; ***P<0.001. Only categories that were found at least once in both the diet and pitfall traps were

3.5 DISCUSSION

3.5.1 Overall dietary patterns Scat analysis revealed that A. arktos and A. stuartii at Springbrook National Park consume a broad range of invertebrate prey (Table 3.1), consistent with diet studies on congeners (Hall 1980; Green 1989; Allison et al. 2006). Plant matter was also frequent in the diet of both species across years. Plant material has been documented in the scats of A. minimus (Allison et al. 2006), A. flavipes (Goldingay 2000) and A. stuartii in north-eastern New South Wales (Statham 1982), which could indicate that plants are

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 89

chosen as a supplementary food source by some species. However, because plant matter was only described in terms of frequency of occurrence in our study, its contribution to diet may have been overestimated due to the poor digestibility of such material by the dasyurid gut system (Hume 2003).

3.5.2 Differences between species The range of invertebrate categories consumed by both species was similar. However, the composition of invertebrates taken differed significantly between species. Superficially, the sympatric A. arktos and A. stuartii have access to the same invertebrate assemblage. However, two species utilizing essentially the same food resource would find themselves in competition if they occupied the same microhabitats (Fox and Archer 1984). Differences in diet of A. arktos and A. stuartii therefore suggest they are foraging in different areas or using different behaviour (e.g., foraging at different times). There is evidence of such separation in other Antechinus spp. For example, Dickman (1986 and 1988) found that controlled removal of A. mimetes from a community shared with A. agilis produced marked changes in microhabitat use of the latter species. Before removal, the larger A. mimetes foraged extensively in topsoil and subsurface leaf litter to obtain prey, whereas the smaller A. agilis foraged on the leaf litter surface or in trees (Dickman 1986). However, following removal of A. mimetes, capture rates of A. agilis in trees declined by 83 % (Dickman 1986 and 1988). Braithwaite et al. (1978) also found that when A. minimus co-occurred with A. agilis the former occupied the soil-fossicking niche and the latter the scansorial niche. A similar relationship may exist between congeners in the present study. A. arktos is also a member of the phylogenetic clade containing A. swainsonii and A. minimus (Baker et al.2015; Mutton 2016) and is significantly larger than A. stuartii (Gray et al. 2016). This complex is also notable as containing the only two Antechinus spp. which forage during the day. Dickman (1988) hypothesised that this would allow these larger species to encounter both diurnal and nocturnal prey. It is currently unknown whether A. arktos also displays some degree of diurnal activity.

3.5.3 Differences between sexes Based on the dietary analyses, males of both species select different prey and may therefore forage in different micro-habitats to their female counterparts. Lazenby- Cohen and Cockburn (1991) suggested that female A. agilis foraged in specific areas because of their need for resources to sustain the costs of reproduction, while male

90 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

foraging varied. They theorized that the quality of a male’s diet was not as important as long as it sustained them through the mating period. The same may apply to the A. arktos and A. stuartii populations studied here.

3.5.4 Differences between sites The Best of All and Bilborough sites in our study differ in vegetation composition and structure. Best of All consists of old growth Gondwanan rainforest, whereas Bilborough consists largely of regrowth rainforest (surrounded by non- remnant forest) and is notably more open (Gray et al. 2016). It was assumed that this variation would be reflected in both the invertebrate composition of the sites and in the diets of both Antechinus species. However, this was not the case. When diet of each species was compared between Springbrook and Lamington, A. stuartii was not found to differ between sites, whereas A. arktos at Lamington consumed a greater number of prey categories. However, the same prey categories (dipteran larvae, Diplopoda, etc.) still dominated A. arktos diet in terms of volume at both sites. Several other studies that examined diet in Antechinus spp. have also found a lack of dietary difference between sites. For example, Lunney et al. (2001) found no difference in dietary composition of A. agilis in unlogged and regenerating forest in NSW and more recently Pearce (2016) found no notable difference in dietary composition of A. mysticus at sites 800 km apart in open (riparian) forest versus rainforest.

3.5.5 Differences between seasons and / or years There was no difference in scat composition between seasons; however, there was a large difference in composition between years. There was considerably less diversity of prey items in scats of both species in 2014 (Figures 3.3 and 3.4) with more ‘empty’ scats devoid of any identifiable invertebrate material. This may, at least in part, be attributable to annual environmental variation. In particular, rainfall has been suggested to have a strong impact on the survival, body mass and overall abundance of A. flavipes (Lada et al. 2013), A. minimus (Sale et al. 2008), A. agilis and A. mimetes (Parrott et al. 2007) and A. stuartii at Springbrook (Parra Faundes 2014). Antechinus spp. typically inhabit highly seasonal, predictable habitats and synchronize their short annual mating season such that lactation and weaning of offspring coincides with peak invertebrate abundance in spring / summer that follow the seasonal pattern of rainfall (Lee et al. 1982; Fisher et al. 2013). Overall, invertebrate abundance increases with rainfall and warm temperatures in tropical zones (Frith and Frith 1990). In dry years,

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 91

invertebrate abundance declines and some species may disappear completely, making it difficult for Antechinus spp. to find sufficient food (Bell 2006). Rainfall in 2014 was lower than 2015 (and the 34 year average) in all months except for March and August (Bureau of Meteorology 2015). It is therefore plausible that invertebrate abundance was lower in 2014. If true, this might explain the lower diversity and volume of invertebrate material found in scats during this time.

Regardless of this purported reduction of invertebrate prey availability in 2014, A. arktos and A. stuartii did not supplement their diet with vertebrates, as observed in other congeners (Green 1989; Dickman 1986; Lunney et al. 2001). Instead, both species may have switched to consuming a higher proportion of soft-bodied prey, such as earthworms (Oligochaeta), less affected by variable rainfall and notoriously difficult to detect in faeces (Dickman and Huang 1988). There is some anecdotal evidence to suggest this. Earthworms were present in up to 25 % of pitfall samples at Springbrook in 2015, so we know they occur in abundance at our study sites. In addition, A. arktos at Lamington consumed earthworms preferentially when offered a range of food sources in captivity (Gray pers. obs.). Consumption of earthworms is also not unprecedented in the genus. Hall (1980) found that earthworms, although uncommon in faeces, were found in the stomach contents of A. mimetes. Grossek et al. (2010) also found that earthworms were a dominant taxon in the diets of small carnivorous marsupials from New Guinea. The high protein and moderate fat content of earthworms and the fact that they have no chitinous exoskeleton and are generally less well defended than some other invertebrates (e.g., Chilopoda and Hymenoptera) may make this group a lucrative dietary item, especially if other prey taxa are scarce (Zhenjun et al. 1997). Molecular analysis of scats would be needed to test this idea.

3.5.6 Dietary strategies Overall, A. arktos and A. stuartii consumed a wide variety of invertebrate prey characteristic of a generalist insectivore, which concurs with the general dietary pattern of the genus. However, whether each species can be considered an opportunistic forager is more difficult to establish. Our results show that quantitatively the diet of both species was not simply determined by prey availability and that these Antechinus spp. were more discerning in their avoidance and preference of certain prey categories than expected. For example, despite being consistently available during all seasons, Diptera (adults) and Formicidae (ants) were largely avoided as prey items. Adult

92 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Diptera are flying insects, and so their avoidance by soil fossicking and scansorial Antechinus spp. is not surprising, while ants have a relatively high proportion of chitin compared to other invertebrates that may render them profitable food sources only if found in high densities (Fisher and Dickman 1993). In our study, both species also consumed less adult Coleoptera than available. However, A. arktos in particular, consumed considerably less Coleoptera than expected in view of their high availability in each season. Coleoptera are one of the most easily identified groups in faecal pellets (Dickman and Huang 1988) so it is unlikely coleopteran fragments were overlooked during our examination of scats. This suggests that beetles were avoided as prey items by A. arktos. This is unexpected, as beetles form a major component in all other Antechinus spp. diets, including the closely related A. mimetes and A. minimus (Green 1989; Allison et al. 2006; Sale et al. 2006). Alternatively, A. arktos may prefer to consume predominantly soft-bodied prey. Amphipoda, Araneae, lepidopteran larvae and coleopteran larvae were all consumed at or just below their availability and dipteran larvae were consumed preferentially in winter. Because pitfall traps essentially measure the activity level of invertebrates and not their abundance directly (Luff 1975), there is a possibility that the availability of insect larvae was underestimated due to their low mobility. However, the high frequency and volume of dipteran larvae in the scats of A. arktos indicate they are preferred prey items. The apparent preference of A. arktos for wet slopes supporting the cloud-stripping Flax Lily (Helmholtzia glaberrima) at both Springbrook and Lamington may be linked to their perceived best foraging areas (Gray et al. 2016; ELG pers. obs.). Antechinus stuartii consumed less coleopteran larvae and dipteran larvae relative to their availability during autumn and spring. Such avoidance may be because this potentially more scansorial species did not encounter these leaf litter invertebrates as often. However, if A. stuartii and A. arktos are foraging in different micro-habitats as suggested, pitfall traps which only measure the prey activity of ground-dwelling invertebrates may not accurately reflect the full suite of available prey for A. stuartii. In future, pitfall trapping could be conducted in conjunction with leaf litter sampling, trunk trapping and / or beating sampling (Moeed and Meads 1983; Ausden 1996) to more completely assess the range of available prey below, on and above ground.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 93

3.5.7 Potential impacts of climate change The ability of individual species to cope with key threatening processes, including climate change, may hinge on characteristics such as range size, dispersal ability, predation, fecundity and dietary specialisation (Laurance 1991; Boyles and Storm 2007). The work of Gray et al. (2016) and the present study suggest that A. arktos populations at Springbrook and Lamington National Parks sustain a small number of males and females that must compete for resources with an orders of magnitude larger sympatric population of A. stuartii. The present study demonstrates an ability of both species to switch dietary preferences markedly, putatively in response to drier conditions and concomitant reduction in invertebrate abundance. The generalist and opportunistic strategies employed by these populations of A. arktos and A. stuartii are advantageous in Gondwanan Rainforest subject to continued disturbance by climate change. As dietary generalists, they may be less sensitive to the loss of prey species or changes to the timing of peak insect abundance. Unfortunately, predatory strategy is only one of many factors affecting the durability of endangered species like A. arktos. Any resilience such dietary flexibility in the face of climate change affords may be swamped by the combined negative effects of annually halving the population (via male die-off), predation by cats (Felis catus) and foxes (Vulpes vulpes), and increasing anthropogenic disturbance from an expanding corner of south-east Qld.

3.6 ACKNOWLEDGEMENTS

This study was generously supported by grants from Australian Geographic, Holsworth Wildlife Research Endowment and Gold Coast City Council. The School of Earth, Environmental and Biological Sciences at the Queensland University of Technology provided access to vehicles and field equipment. ELG was financially supported by an Australian Postgraduate Award for doctoral research. We are greatly indebted to Ian Gynther (Queensland Department of Environment and Heritage Protection) and Harry Hines (Queensland Parks and Wildlife Service) who loaned us Elliott traps and freely shared their expertise, both of which proved critical to our success. We thank all volunteers who ventured out with us to Springbrook, rain or shine (usually the former) to help us check traps and dig pitfalls. We also thank Jennifer Firn and Coral Pearce who provided essential insight and advice about using PRIMER to analyse our volume data.

94 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

3.7 REFERENCES

Allison, L., L. Gibson., and J. Aberton. 2006. Dietary strategy of the swamp antechinus (Antechinus minimus maritimus) (Marsupialia: Dasyuridae) in coastal and inland heathland habitats. Wildlife Research 33: 67-76.

Anderson, M.J., R.N. Gorley., and K.R. Clarke. 2008. PERMANOVA+ for PRIMER: guide to software and statistical methods. PRIMER-E Ltd. Plymouth, UK.

Ausden, M. 1996. Invertebrates. Pp 139-177 in Ecological Census Techniques: A Handbook (W.J. Sutherland eds). Cambridge University Press, UK.

Australian National University. 2009. Implications of climate change for Australia’s world heritage properties: a preliminary assessment. A report to the Department of Climate Change and the Department of Environment, Water, Weritage and the Arts by the FennerSschool of Environment and Society, the Australian National University.

Baker, A., T. Mutton., and S. Van Dyck S. 2012. A new dasyurid marsupial from eastern Queensland, Australia: the buff-footed Antechinus, Antechinus mysticus sp. nov. (Marsupialia: Dasyuridae). Zootaxa 3515: 1-37.

Baker, A., T. Mutton., and H. Hines. 2013. A new dasyurid marsupial from Kroombit Tops, south-east Queensland, Australia: the silver-headed Antechinus, Antechinus argentus sp. nov. (Marsupialia: Dasyuridae). Zootaxa 3746: 201-239.

Baker, A., T. Mutton., H. Hines., and S. Van Dyck. 2014. The black-tailed Antechinus, Antechinus arktos sp. nov.: a new species of carnivorous marsupial from montane regions of the Tweed Volcano caldera, eastern Australia. Zootaxa 3765: 101- 133.

Baker, A., T. Mutton., E. Mason., and E.L. Gray. 2015. A taxonomic assessment of the Australian dusky Antechinus complex: a new species, the Tasman Peninsula dusky Antechinus (Antechinus vandycki sp. nov.) and an elevation to species, the mainland dusky Antechinus (Antechinus mimetes new status). Memoirs of the Queensland Museum.Nature 59: 75-126.

Bell, H.L. 2006. Seasonal variation and the effects of drought on the abundance of arthropods in savannah woodland on the Northern Tablelands of New South Wales. Australian Journal of Ecology 10: 207-221.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 95

Boyles, J.G., and J.J. Storm. 2007. The perils of picky eating: dietary breadth is related to extinction risk in insectivorous bats. PLoS ONE 2: e672.

Braithwaite, R.W., A. Cockburn., and A.K. Lee. 1978. Resource partitioning by small mammals in lowland heath communities of south-eastern Australia. Australian Journal of Ecology 3: 423-445.

Bureau of Meteorology. 2015. Daily rainfall: Springbrook Road. Australian Government.http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=1 36&p_display_type=dailyDataFile&pstnnum=040607&pstartYear=.

Chen, X., C.R. Dickman., and M.B. Thompson.1998. Diet of the , Dasycercus cristicauda in the Simpson Desert, central Australia. Wildlife Research 25: 233-242.

Clarke, K.R., R.N. Gorley., P.J. Somerfield., and R.M. Warwick. 2014. Change in marine communities: an approach to statistical analysis and interpretation. PRIMER-E Ltd. Devon, UK.

Clarke, K.R., and R.N Gorley. 2015. PRIMER v7: User Manual/Tutorial. PRIMER-E, Plymouth.

Crowther, M.S., and M.J. Blacket.2003. Speciation and biogeography within the Dasyuridae, or why are there so many species of dasyurids? Pp 124-130 in Predators with pouches: the biology of carnivorous marsupials (M.E. Jones., C.R. Dickman., and M. Archer, eds.). CSIRO Publishing. Collingwood, Australia.

CSIRO. 1991. The insects of Australia 2nd edition. Melbourne University Press. Melbourne.

Dickman, C.R. 1986. An experimental study of competition between two species of dasyurid marsupials. Ecological Monographs 56: 221-241.

Dickman, C.R. 1988. Body size, prey size and community structure in insectivorous mammals. Ecology 69: 569-580.

Dickman, C.R. 2014. Micro-carnivores: the ecological role of small dasyurid predators in Australia. Pp 241-262 in Carnivores of Australia: past, present and future (A.S. Glen and C.R. Dickman, eds). CSIRO Publishing. Collingwood, Australia.

96 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Dickman, C.R., and C. Huang. 1988. The reliability of faecal analysis as a method for determining the diet of insectivorous mammals. Journal of Mammalogy 69: 108-113.

Fisher, D.O., and C.R. Dickman. 1993. Body size-prey relationships in insectivorous marsupials: tests of three hypotheses. Ecology 74: 1871-1883.

Frith, D., and C. Frith. 1990. Seasonality of litter invertebrate populations in an Australian upland tropical rainforest. Biotropica 22: 181-190.

Fisher, D.O., C.R. Dickman., M.E. Jones., and S.P. Blomberg. 2013. Sperm competition drives the evolution of suicidal reproduction in mammals. Proceedings of the National Academy of Sciences 110: 17910-17914.

Fox, B.J., and E. Archer. 1984. The diets of Sminthopsis murina and Antechinus stuartii (Marsupialia: Dasyuridae) in sympatry. Australian Wildlife Research 11: 235- 248.

Futuyma, D.J., and G. Moreno. 1988. The evolution of ecological specialization. Annual Review of Ecology and Systematics 19: 207-233.

Goldingay, R.L. 2000. Small dasyurid marsupials – are they effective pollinators? Australian Journal of Zoology 48: 597-606.

Gray, E.L., AM. Baker., and J. Firn. 2016. Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii). Mammal Research DOI 10.1007/s13364-016-0281-1.

Green, K. 1989. Altitudinal and seasonal differences in the diets of Antechinus swainsonii and A. stuartii (Marsupialia: Dasyuridae) in relation to the availability of prey in the Snowy Mountains. Australian Wildlife Research 16: 581-592.

Grossek, G.L., P.A. Woolley., and J.I. Menzies. 2010. The diet of some New Guinean dasyurid marsupials. Australian Mammalogy 32: 145-155.

Hagger, V., D. Fisher., S. Schmidt., and S. Blomberg. 2013. Assessing the vulnerability of an assemblage of subtropical rainforest vertebrate species to climate change in south-east Queensland. Austral Ecology 38: 465-475.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 97

Hall, S. 1980. The diets of two coexisting species of Antechinus (Marsupialia: Dasyuridae). Wildlife Research 7: 365-378.

Hennessy. K.J., K.L. McInnes., D.J. Abbs., R.N. Jones., J. Bathols., R. Suppiah., J. Ricketts., T. Rafter., D. Collins., and D. Jones. 2004. Climate change in New South Wales Part 2: projected changes in climate extremes. New South Wales Greenhouse Office, Climate Impact Group, CSIRO Atmospheric Research and National Climate Centre, Australian Government Bureau of Meteorology: Canberra.

Hume, I.D. 2003. Nutrition of carnivorous marsupials. Pp 221-228 in Predators with pouches: the biology of carnivorous marsupials (M. Jones., C.R. Dickman., and M. Archer, eds). CSIRO Publishing. Collingwood, Australia.

Hunter, R.J. 2003. World heritage and associative natural values of the Central Eastern Rainforest Reserves of Australia. NSW National Parks and Wildlife Service. Government of New South Wales. Retrieved 13 May 2016.

Hutley, L. B., D. Doley., D.J. Yates., and A. Boonsaner. 1997. Water balance of an Australian subtropical rainforest at altitude: the ecological and physiological significance of intercepted cloud and fog. Australian Journal of Botany 45: 311–329.

Lada, H., J.R. Thompson., S.C. Cunningham., and R.M. Nally. 2013. Rainfall in prior breeding seasons influences population size of a small marsupial. Austral Ecology 38: 581-591.

Laurance, W. F. 1991. Ecological correlates of extinction proneness in Australian tropical rainforest mammals. Conservation Biology 5: 79–89.

Lazenby-Cohen, K.A., and A. Cockburn. 1991. Social and foraging components of the home range in Antechinus stuartii (Dasyuridae: Marsupialia). Australian Journal of Ecology 16: 301-307.

Lee, A.K., P. Woolley., R.W. Braithwaite. 1982. Life history strategies of dasyurid marsupials. Pp. 1-11 in Carnivorous marsupials (M. Archer, eds). Royal Zoological Society of New South Wales. Sydney, Australia.

Luff, M.L. 1975. Some features influencing the efficiency of pitfall traps. Oecologia 19: 345-357.

98 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Lunney, D., A. Matthews., and J. Grigg. 2001. The diet of Antechinus agilis and A. swainsonii in unlogged and regenerating sites in Mumbulla State Forest, south- eastern New South Wales. Wildlife Research 28: 459-464.

Mason, E.D., C.J. Burwell., and A.M. Baker. 2015. Prey of the silver-headed antechinus (Antechinus argentus), a new species of Australian dasyurid marsupial. Australian Mammalogy 37: 164-169.

Moeed, A., and M.J. Meads. 1983. Invertebrate fauna of 4 tree species in the Orongorongo Valley, New Zealand, as revealed by trunk traps. New Zealand Journal of Ecology 6: 39-53.

Mutton, T. 2016. Evolutionary biology of the Australian carnivorous marsupial genus Antechinus, with particular focus on two species, A. mysticus and A. subtropicus. Submitted P.h.D dissertation. Queensland University of Technology.

MacArthur, R.H., and E.R. Pianka. 1966. On optimal use of a patchy environment. American Naturalist 100: 603-609.

McAllan, B.M., C.R. Dickman., and M.S. Crowther. 2006. Photoperiod as a reproductive cue in the marsupial genus Antechinus: ecological and evolutionary consequences. Biological Journal of the Linnean Society 87: 365-365.

McJannet, D., W.P. Fitch., M. Disher., and P. Reddell. 2007. Water balance of tropical rainforest canopies in north Queensland Australia. Hydrological Processes 21: 3473-3484.

O’Donoghue, M.S., S. Boutin., C.J. Krebs., G. Zuleta., D.L. Murray., and E.J. Hofer. 1998. Functional responses of coyotes and lynx to the snowshoe hare cycle. Ecology 79: 1193-1208.

Parra Faundes, D.V. 2014. Sex allocation, terminal investment and the effects of environmental constraints on maternal investment in subtropical antechinus. PhD thesis. University of Queensland, Australia.

Parrott, L.P., S.J. Ward., P.D. Temple-Smith., and L. Selwood. 2007. Effects of drought on weight, survival, and breeding success of agile antechinus (Antechinus agilis), dusky antechinus (A. swainsonii) and bush rats (Rattus fuscipes). Wildlife Research 34: 437-442.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 99

Pavey, C.R., C.J. Burwell., G. Körtner., and F. Geiser. 2009. Vertebrate diet decreases winter torpor use in a desert marsupial. Naturwissenschaften 96: 679-683.

Pavey, C.R., C.J. Burwell., G. Körtner., and F. Geiser. 2016. Why is the marsupial kaluta, Dasykaluta rosamondae, diurnally active in winter: foraging advantages or predator avoidance in arid northern Australia? Journal of Arid Environments 133: 25-28.

Pearce, C., C.J. Burwell., and A.M. Baker. 2016. Diet, breeding timing and growth in a new carnivorous marsupial, the Buff-footed Antechinus, Antechinus mysticus, at the northern and southern limits of its range. Master’s thesis, Queensland University of Technology, Australia.

Ross, S., B. Munkhtsog., and S. Harris. 2010. Dietary composition, plasticity, and prey selection of Pallas’s cats. Journal of Mammalogy 91: 811-817.

Sale, M.G., S.J. Ward., and J.P.Y. Arnould. 2006. Aspects of the ecology of swamp antechinus (Antechinus minimus maritimus) on a Bass Strait island. Wildlife Research 33: 215-221.

Sale, M.G., B.A. Wilson., and J.P.Y. Arnould. 2008. Factors influencing population dynamics in island and mainland populations of the swamp antechinus (Antechinus minimus). Australian Journal of Zoology 56: 187-194.

Statham, H.L. 1982. Antechinus stuartii (Dasyuridae, Marsupialia) diet and food availability at Petroi, north-eastern New South Wales. Pp 151-163 in Carnivorous Marsupials Vol 1 (M. Archer, eds) Sydney Royal Zoological Society of New South Wales.

Taylor, K.J., A.J. Lowe., R.J. Hunter., T. Ridgway., P.M. Gresshoff., and M. Rossetto. 2005. Genetic diversity and regional identity in the Australian remnant Nothofagus moorei. Australian Journal of Botany 53: 437–444.

Van Dyck, S., and R. Strahan. 2008. The Mammals of Australia. New Holland Publishers. Sydney, Australia.

Ward, S.J. 2000. Possible predation of feathertail gliders (Acrobates pygmaeus) by agile antechinus (Antechinus agilis). Australian Mammalogy 21: 173-176.

100 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Work, T.T., C.M. Buddle., Korinus L.M., and Spence, J.R. 2002. Pitfall trap size and capture of three taxa of litter-dwelling arthropods: implications for biodiversity studies. Environmental Entomology 31: 438-448.

Zborowski, P., and R. Storey. 2010. A field guide to insects in Australia 3rd edition. Reed New Holland Publishers. NSW.

Zhenjun, S., Xianchun, L., Lihui, S., & Chunyang, S. 1997. Earthworm as a potential protein resource. Ecology of Food and Nutrition 36: 221–236.

Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos) 101

102 Chapter 3: Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two sympatric semelparous marsupials, including a new endangered species (Antechinus arktos)

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Emma L. Gray, Todd E. Dennis and Andrew M. Baker

Submitted on 13 April 2017, accepted 3 July 2017, and published online 9 August 2017 in the journal PLoS ONE.

4.1 ABSTRACT

The black-tailed dusky antechinus (Antechinus arktos) is an endangered, small carnivorous marsupial endemic to Australia, which occurs at low population density along with abundant sympatric populations of other small mammals: Antechinus stuartii, Rattus fuscipes and Melomys cervinipes. Using A. arktos as a model species, we aimed to evaluate the effectiveness of infrared digital camera traps for detecting and differentiating small mammals and to comment on the broad applicability of this methodology. We also sought to understand how the detection probabilities of our target species varied over time and characterize their activity patterns. We installed 11 infrared cameras at one of only three known sites where A. arktos occurs for five consecutive deployments. Cameras were fixed to wooden stakes and oriented vertically, 35 cm above ground, directly facing bait containers. Using this method, we successfully recorded and identified individuals from all four species of small mammal known previously in the area from live trapping, including A. arktos. This validates the effectiveness of the infrared camera type and orientation for small mammal studies. Periods of activity for all species were highly coincident, showing a strong peak in activity during the same two-hour period immediately following sunset. A. arktos, A.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 103

stuartii and M. cervinipes also displayed a strong negative linear relationship between detection probability and days since deployment. This is an important finding for camera trapping generally, indicating that routine camera deployment lengths (of one- to-two weeks) between baiting events may be too long when targeting some small mammals.

Keywords Antechinus arktos, Antechinus stuartii, camera trapping, detection, diel activity, GLMM, Melomys cervinipes, Rattus fuscipes

4.2 INTRODUCTION

Earth is currently experiencing biodiversity losses of a magnitude described by many as constituting a sixth mass extinction [1, 2]. Conservation of threatened cryptic species often hinges on our ability to identify where they occur, so that critical populations can be monitored and managed appropriately. Field surveys involving the capture of individuals by live traps is a common method employed by wildlife researchers to establish the occurrence of small terrestrial mammal species and to monitor populations (e.g., their abundance, survival and recruitment) over time [3, 4]. These direct sampling methods provide immediate and generally unambiguous species identifications and enable additional information to be collected from individuals, such as genetic material and their sex, age, body mass and condition [5]. However, recently, the proliferation of commercial wildlife camera traps has led to a sharp increase in the use of camera traps for small-medium sized mammal occurrence surveys and monitoring [6-8]. Camera traps are generally defined as remotely triggered cameras that automatically take images and / or videos of passing animals [6]. Remotely deployed cameras eliminate the need for researchers to trap and physically handle wild animals (thus avoiding often severe regulatory constraints) and can be deployed for long periods (up to several months), thereby reducing operational costs, time and effort [5, 7]. Although camera traps may be seen and / or heard by animals to some degree [9, 10], they still provide an opportunity to detect and monitor rare and / or trap-shy species that may otherwise be missed or under-detected by direct census methods [11, 12], as well as collect potentially valuable information about behaviour and activity [13, 14].

When deployed in the field, camera traps are generally mounted horizontally on trees, oriented passively outward or toward one or more bait holders [5, 15]. Such a

104 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

setup has proved highly successful for detecting a wide variety of medium-large mammals, such as cats [16], ungulates [17] and foxes [18]. For detecting smaller mammal species, having the camera positioned closer to the subject is advantageous [19]. In such cases, whether the best camera orientation is horizontal or vertical is an open question [see 5, 15, 20, 21]. De Bondi et al. [5] demonstrated that small-medium sized mammals could be successfully detected by vertically orienting camera lenses toward a bait holder placed a standard height above ground. This alternative method has since been used to successfully detect the critically endangered central rock-rat (Zyzomys pedunculatus) [22] and invasive black rat (Rattus rattus) [12, 23]. However, the accurate identification of small mammals from camera traps is still challenging, especially where morphologically similar looking species co-exist [5, 24, 25]. If camera trapping is going to continue to be used for small threatened species surveys and monitoring, then further ‘proof of concept’ evidence is required in challenging environments to ensure that individual species can be consistently detected and identified where they occur, and that efforts to reduce field time and cost do not compromise estimates of species occurrence and persistence [24, 25].

The black-tailed dusky antechinus, Antechinus arktos, is one of 15 species of Antechinus, a genus of small (16-170 g) carnivorous marsupials endemic to Australia [26-28]. Antechinus are predominantly nocturnal insectivores renowned for their semelparous reproductive system, which features a short, promiscuous mating period, concluded by the abrupt death of all males in a population [26, 29]. Collectively, members of the genus occur in coastal / near coastal forest across all of Australia’s states and mainland territories. The geographic distributions of several species, including A. arktos, are severely limited [27]. The species is known only from three sites located a maximum 8 km (straight line distance) apart within cloud forest at the summit of the Tweed Shield Volcano caldera (900-1200 m elevation), which straddles the border of Queensland (Qld) and New South Wales (NSW) in mid-eastern Australia [30]. Therefore, A. arktos has been classified as Endangered in both states [31, 32] and is currently being considered for federal threatened species listing. Currently, the most important conservation priorities for A. arktos are to ensure the continued persistence of the three known populations and locate and protect previously unknown populations, should they exist. However, Antechinus arktos exemplifies the challenges associated with detecting and monitoring small, elusive / rare mammals.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 105

Recently, a two-year mark-recapture study of A. arktos was undertaken using Elliott (metal box) traps [30]. This method proved effective at capturing A. arktos; however, trap success (number of captures divided by number of trap nights) was consistently low throughout its limited range, especially outside their short pre- breeding / breeding period, never exceeding 3.5 % (or 7 individuals per 200 trap nights). Additionally, A. arktos co-occurs with much larger populations of the brown antechinus (Antechinus stuartii), bush rat (Rattus fuscipes) and fawn-footed melomys (Melomys cervinipes), which may restrict access of A. arktos to live traps. Thus, to be 95 % confident of detecting the species at other sites if they are present requires 600 trap nights conducted within this short enhanced detection-timing window [30, 33]. Deploying such a large number of traps at remote sites limits the number of sites that can be surveyed.

The camera trapping approach described by De Bondi et al. [5] may provide an alternative means of detecting A. arktos for occurrence surveys and monitoring programs. However, first camera trapping trials within known A. arktos habitat must be conducted. In contrast to many previous camera trapping studies, A. arktos co- occurs with a morphologically similar congener A. stuartii and other highly abundant small mammal populations, which may confound accurate identification. Therefore, using A. arktos as a model, we aimed to: 1) assess the utility of infrared digital camera traps for detecting and distinguishing A. arktos from other co-occurring small mammals; 2) identify factors that influence temporal variation in detection probability for each species; 3) examine diel activity patterns and their extent of overlap among species’ and 4) provide recommendations concerning the applicability of camera trapping as a survey method for A. arktos and other similar small mammal species.

4.3 METHODS

All aspects of the study were carried out with approval from the Queensland University of Technology ethics department (approval number: 1400000005) and the Queensland Parks and Wildlife service (approval number: WITK14454114).

4.3.1 Study site Our study was conducted at Best of All Lookout (28.2415°S, 153.2640°E) within Springbrook National Park, ~100 km south of Brisbane, Queensland, Australia.

106 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Located at the rim of the Tweed Shield Volcano caldera, Springbrook National Park has major remnants of UNESCO World Heritage listed ‘Gondwana Rainforests of Australia’ that provides important refugia for many relict and endemic species of flora and fauna [34]. Mean elevation of the site is 950 m and consists of complex notophyll vine forest (Regional Ecosystem 12.8.5) and simple microphyll fern forest (Regional Ecosystem 12.8.6) [35]. The study area was centred in a steep headwater gully containing the cloud-stripping stream lily, Helmholtzia glaberrima, and a small stand of Antarctic beech, Lophozonia moorei.

4.3.2 Camera-trapping design Camera trapping was conducted between 10 August and 14 October 2016, which encompassed the breeding period from pre-mating to post-mating male die-off of the two local Antechinus spp. Eleven Ltl Acorn® infrared digital cameras (Ltl-5310 series) were deployed at the study site and left in the same position for five consecutive deployments (ranging from 11-16 days in duration). Placement of the cameras utilized a pre-existing live-trapping grid consisting of 4 (200 m long and 20 m apart) parallel transects oriented down slope [30]. Cameras were randomly assigned to stations set ~50 m apart along each transect. We recognise this is a small area to deploy camera traps; however, A. arktos has a very limited known distribution and may occur patchily even within sites it is known to occur [30, 36]. Setting camera traps within the live- trapping grid, where the species has previously been captured, allowed us to be confident that A. arktos was present, and thus detectable, rather than absent due to unsuitable habitat. Cameras were mounted on wooden stakes that were hammered into the ground, with the lens positioned 35 cm above the ground surface and directed downward towards a bait container (Fig 4.1). To limit the occurrence of false positive camera detections, vegetation and leaf litter were cleared from within the camera’s infrared sensor zone. Several layers of cream-coloured masking tape were placed over the cameras’ LED lights to reduce the intensity of the flash when the cameras were triggered and avoid over-illuminating or possibly startling the target species. A seven- day field trial conducted in July 2016 at the study site was used to optimize camera settings and quantify the operational periods of the cameras. Based on information collected during this field trial, we configured all cameras to record a single photograph (JPEG format, 5 megapixels) when triggered, immediately followed by a 20-s video (AVI format, 640 by 480 pixels per frame), followed by a 10-min interval,

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 107

24 hrs per day, as some antechinus species are known to be at least partially diurnal [37, 38]. We considered a 10-min minimum interval to the next possible recording event of the cameras to be an acceptable compromise between detecting the same individuals repeatedly (which reduces battery life, as well as available storage of recorded media) and the likelihood of failing to record individuals that only infrequently visited the cameras’ monitoring areas [23]. To increase the probability of detecting target species (particularly A. arktos), we used a bait mixture of peanut butter, oats and bacon, which is commonly used in live trapping studies [3, 30] and has proven to be a superior bait type for antechinus [39], together with a single application (2-s spray) of FeralMoneTM, a generic carnivore attractant (Animal Control Technologies, Somerton, VIC, Australia; Jesse Rowland pers.comm.). Bait containers comprised 60 X 75 mm PVC vent cowls that were secured to the ground with tent pegs [40]. These devices permitted target species to see and smell the bait mixture but prevented them from directly accessing or removing it. Bait (including FeralMone), camera batteries and storage media (SD cards) were retrieved and replaced after each successive deployment.

108 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Figure 4.1. Schematic image of the vertical camera set-up showing camera positioning, with essential components and distances labelled

Animal recordings were identified to species level based on body size, body and head shape, tail length, ear morphology and behaviour (particularly movements captured on video). Reference photographs of A. arktos and A. stuartii taken during live trapping aided in identification. If identification was still uncertain after viewing both image and video files the animal was classified as an “unknown mammal”, “unknown antechinus” or “unknown murid” for the purposes of categorisation and the record was not used in subsequent analyses. Because successful and consistent identification between the two Antechinus spp. was an important element of our study, all antechinus observations were independently assessed by two researchers (ELG and

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 109

AMB) and only individuals positively identified by both researchers were categorised to species level.

Camera trap recordings were transcribed into separate binary response variables, one for each species per day of study, with a ‘1’ indicating species presence and a ‘0’ indicating species absence. Due to a high number of missing values (caused by the uneven deployment lengths ranging from 11-16 days) only the data from the first 11 days of each deployment were used for subsequent analyses, which was the minimum length of time that any of the cameras was set. We consider trimming deployment periods in the manner described above to be the best means of addressing missing observations, as plots of the capture proportions of each target species expressed as a function of the number of days of camera deployments clearly demonstrated that there were no marked changes in capture rates of any target species after day 11. Prior to statistical analysis we developed a list of explanatory variables that might reasonably influence detection probabilities (Table 4.1), and conducted exploratory data analysis on both the response and explanatory variables to test for outlier observations and evaluate the extent of collinearity among independent variables, as suggested by Zuur et al. [41].

Table 4.1. Explanatory variables used for modelling detection probabilities by remote camera traps of four species of small mammals at Springbrook National Park.

Variable Term Data type and values Name Cam Camera trap ID Categorical: 11 levels (1,2,3,4,5 etc.)

Dep Deployment Ordinal: 5 ordered levels (D1,D2,D3,D4,D5) each of 11 days length DayDep Days since Continuous: number of days since cameras were deployment deployed Moon Moon phase Categorical: 2 levels. M1= light, first quarter to third quarter, M2=dark, waning crescent to waxing crescent Rain Rainfall (mm) Continuous: total rainfall (mm) measured per day from nearest weather station

110 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

4.3.3 Data analysis The data are repeated occurrence observations of the target species at fixed sites, with one level of spatial structure (Camera trap ID) and two levels of temporal structure (day of deployment nested within deployment). ‘Deployment’ can be considered to be both a study design and seasonal variable that reflects the breeding stage and survival of the two antechinus species over time. Camera deployments #1 and #2 occurred prior to antechinus breeding periods, deployment #3 coincided with breeding, and camera deployments #4 and #5 took place post-breeding. ‘Day of deployment’ reflects the duration of bait effectiveness and possible neophobic responses of species to the cameras within each deployment. Rainfall and moon phase were also included as possible explanatory variables due to their reported influence on the activity of other small mammal species [42].

We used generalized linear mixed models (GLMMs) to determine which variables best accounted for the detection probabilities of the target species. Because the response variable was binary, it was fitted to a binary distribution through a log- link function. Camera trap ID was included as a random effect nested within ‘deployment’ and ‘day of deployment’ to account for pseudoreplication (repeated measures) of observations over both time scales. Significance of fixed effects was assessed by computing Wald statistics [43]. The ‘best’ (minimal adequate) model was determined by fitting the full statistical model and then excluding non-significant terms using a stepwise backward selection process recommended by Crawley [43]. To examine the effects of individual explanatory variables on detection probability we plotted the fitted relationships from the minimal adequate model of each species. Following Rendall et al. [12] we also used estimated daily detection probabilities from the minimal adequate model for each species to calculate the cumulative detection probability for each day since deployment using the formula:

P = 1 – (1 – p1) * (1 – p2) * (1 – p3)…(1 – pn) (1)

Where P is the cumulative nightly detection probability, p1 is the detection probability for night one, and n is the total number of survey nights per deployment. GLMMs were fitted using the R package lme4 [44].

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 111

4.3.4 Activity patterns and overlap Date and time records for each photo and video pair were used to investigate and compare the diel activity patterns of each target species. Because none of the target species had natural features or markings that allowed for individual identification within a species and because continued presence of species (after each 10-min interval) at a camera trap represented continued foraging activity, all capture records were included in the analyses as suggested by Carver et al. [45]. In addition, multiple records of individuals (e.g., two R. fuscipes individuals observed simultaneously at the same camera trap) were treated as separate events. All analyses of activity patterns were performed using the Overlap package [46] in R studio version 3.1.1 [47] using code adapted from Meredith and Ridout [46]. Probability density functions of activity for each species were estimated non-parametrically using kernel density estimates [13]. Then, to compare the activity times of each pair of sympatric species, the degree of overlap between the two estimated densities were measured. Various measures of overlap have been proposed; however, we used the ‘coefficient of overlapping (Δ)’ recommended by Ridout and Linkie [13]. This is a quantitative measure ranging from 0 (signifying no overlap) to 1 (signifying complete overlap). There are three alternative means of estimating the coefficient of overlapping, labelled Δ1, Δ4 and Δ5. Here, however, we used only Δ1 and Δ4 , which were recommended for sample sizes less than 50 and sample sizes greater than 75, respectively [13, 46]. Confidence intervals for coefficients of overlapping were obtained as percentile intervals from a recommended 10 000 bootstrap samples [46].

4.4 RESULTS

4.4.1 General findings In total, 8, 273 JPEG and AVI video pairs were recorded over 725 camera trap nights (Table 4.2). Of these pairs, 3, 207 (38.8 %) were deemed to be ‘false triggers’, with 49.9 % caused by flies active during daylight hours. Additionally, 334 (4 %) of observations could not be identified to species. This was overwhelmingly due to poor image quality, with less than six antechinus image and video pairs (across all cameras; 0.1 %) unable to be identified to species level due to disagreement between researchers. The remaining 5, 168 image and video pairs represented fauna from 10 different taxonomic groups, including our four target species: A. arktos, A. stuartii,

112 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

R. fuscipes and M. cervinipes. Non-target animals included: northern brown bandicoot (Isoodon macrourus), possums (Trichosurus spp.), short-beaked echidna (Tachyglossus aculeatus), feral (Felis catus), macropod, and various rainforest birds (not identified). R. fuscipes was detected most frequently, constituting 56 % of all mammal observations, with trap success (number of detections divided by number of camera trap nights) ranging from 77-98 % between deployments. Next most numerous was M. cervinipes (15.2 % of all observations; trap success ranging from 33-65 %), A. stuartii (13.2 % of all observations; trap success ranging from 15-59 %), I. macrourus (3.9 % of all observations; trap success ranging from 7-35 %) and finally A. arktos (2.1 % of all observations; trap success ranging from 3-21 %). The two antechinus species could easily be distinguished from the Muridae species by their smaller body sizes and pronounced, pointed snouts (Supplementary Figure 4.1). Muridae could also be easily identified to species level due to marked differences in species facial features and ear morphology (M. cervinipes has a shorter face, while R. fuscipes has conspicuously rounded ears; Supplementary Figure 4.1). Distinguishing between the two antechinus species was more challenging. Generally, the much larger body size and more rounded rump of A. arktos was sufficient for definitive identification (Supplementary Figure 4.1). However, very large A. stuartii males are similar in body size to small A. arktos females [30]. In such cases, video footage was often crucial in order to confidently assign antechinus to species level. The behaviour and movements of the two antechinus species were particularly diagnostic. A. stuartii typically exhibited rapid stop-start movements and regularly climbed the wooden stakes supporting the cameras; in contrast, A. arktos tended to move in a much slower, ‘shuffling’ gait, and always remained on the ground.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 113

Table 4.2. Summary of all camera trap image pairs recorded at Springbrook National Park over five successive deployments during 2016.

Total No. % of all mammal Trap success captures captures % Rattus fuscipes 2 439 56.0 77-98 Melomys cervinipes 662 15.2 33-65 Antechinus stuartii 574 13.2 15-59 Antechinus arktos 91 2.10 3-21 Isoodon macrourus 169 3.90 7-35 Trichosurus sp. 81 1.90 2-22 Tachyglossus aculeatus 2 0.05 - Felis catus 3 0.07 - Unidentified mammal 134 3.10 - Unidentified murid 134 3.10 - Unidentified antechinus 66 1.50 - Macropod 1 0.02 - Aves 812 - False trigger 3 207 - - Total 8 375

4.4.2 Temporal changes in detection probabilities GLMMs of detection probability were fitted for each of the four target species. The minimal adequate detection model for A. arktos, A. stuartii and M. cervinipes included the fixed effects of ‘deployment’ and ‘days since deployment’ (Table 4.3). The most parsimonious detection model for R. fuscipes, however, was the null (Table 4.3), with detection probability consistently high (close to 1) irrespective of effects from any of the explanatory variables. A. arktos and A. stuartii both had significant quartic relationships with deployment number (Table 4.3); detection probabilities were lower during camera deployments #2, #4 and #5 compared with deployments #1 and #3 (Fig 4.2). In comparison, M. cervinipes had a strong negative linear relationship with deployment number (Table 4.3; Fig 4.2).

114 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Table 4.3. Results of the minimal adequate model explaining detection of A. arktos, A. stuartii, M. cervinipes and R. fuscipes

Estimate Std. Error Z value P value A. arktos Intercept -1.481 0.274 -5.409 <0.001 Dep.L -1.041 0.362 -2.880 0.004 Dep.Q -0.054 0.326 -0.165 0.869 Dep.C -0.593 0.422 -1.405 0.160 Dep^4 1.565 0.343 4.560 <0.001 DayDep -0.147 0.044 -3.354 <0.001 A. stuartii Intercept 0.404 0.264 1.531 0.126 Dep.L -0.584 0.268 -2.181 0.029 Dep.Q -0.443 0.268 -1.653 0.098 Dep.C -0.076 0.288 -0.264 0.792 Dep^4 2.198 0.294 7.486 <0.001 DayDep -0.166 0.032 -5.121 <0.001 M. cervinipes Intercept 1.935 0.236 8.210 <0.001 Dep.L -1.115 0.261 -4.266 <0.001 Dep.Q -0.054 0.255 -0.213 0.831 Dep.C 0.207 0.260 0.798 0.425 Dep^4 -0.103 0.256 -0.401 0.688 DayDep -0.305 0.033 -9.167 <0.001 R. fuscipes Intercept 13.669 1.439 9.496 <0.001 Fixed factors include: Deployment (Dep) and days since deployment (DayDep). Deployment levels are labelled L, Q, C, and ^4, which stand for ‘linear’, ‘quadratic’, ‘cubic’, and ‘quartic’ polynomial terms respectively. Statistically significant values (P <0.05) are highlighted in bold font. *A. arktos AIC: 401.4, BIC: 440.9, loglik: -191.7, deviance: 383.4, residual DF: 585. *A. stuartii AIC: 666.6, BIC 706.1, loglik: -324.3, deviance: 648.6, residual DF: 585. *M. cervinipes AIC: 711.7, BIC: 751.2, loglik: -346.9, deviance: 693.7, residual DF: 585. *R. fuscipes AIC 157.8, BIC: 175.3, loglik-74.9, deviance: 149.8, residual DF: 590.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 115

Figure 4.2. Model estimates of the effect of camera deployment period on detection probabilities of our study’s four small-mammal target species: A. arktos (AA), A. stuartii (AS), M. cervinipes (MC) and R. fuscipes (RF). Vertical bars represent 95% confidence intervals.

Within individual camera deployments, there was a strong negative linear relationship between detection probabilities and ‘days since deployment’ for A. arktos, A. stuartii and M. cervinipes; the slopes of these relationships differed markedly between species (Table 4.3; Fig 4.3). A. arktos decreased from an average detection probability of 0.2 on the first day of each deployment to 0.06 on day 11, A. stuartii from 0.54 to 0.25 and M. cervinipes from 0.82 to 0.21 (Fig 4.3). Cumulative detection probability curves show that detection probability reached 95 % after just one night for R. fuscipes, two nights for M. cervinipes and five nights for A. stuartii (Fig 4.4). However, a cumulative detection probability of 95 % was never achieved for A. arktos. After 11 nights the cumulative detection probability of A. arktos was just 76 % (Fig 4.4).

116 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Figure 4.3. Relationships between detection probabilities estimated by GLMM and days since camera deployment (averaged across all five deployments) for our four target species: A. arktos (AA), A. stuartii (AS), M. cervinipes (MC) and R. fuscipes (RF). Bars represent 95% confidence intervals.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 117

Figure 4.4. Cumulative detection probability curves calculated for A. arktos (AA), A. stuartii (AS), M. cervinipes (MC) and R. fuscipes (RF) from GLMM minimal adequate detection models (averaged across all five deployments).

4.4.3 Activity patterns and overlap Periods of activity were strongly coincident for all four target species, as indicated by values of overlap coefficients being >0.7 (Table 4.4). A. arktos was predominantly nocturnal (91 % of observations were recorded during darkness), with a primary peak in activity between 18:00 and 19:00 hours, followed by smaller spikes of activity from 21:00 to 22:00 and at 03:00 hours, respectively (Fig 4.5). Similarly, A. stuartii had a primary peak of activity between 18:00 and 19:00 hours, with smaller spikes from 21:00 to 22:00 and at 02:00 hours (Fig 4.5). However, A. stuartii displayed stronger diurnal activity, with 115 detection events (20 % of all captures) recorded during daylight hours. Although not shown in Fig 4.5, the majority of this daytime activity (86 captures) occurred during the species’ breeding period within deployment

118 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

#3 from late August to mid-September. In comparison, R. fuscipes and M. cervinipes displayed a unimodal, peak in activity between 18:00 and 19:00 hours (Fig 4.5) and were strongly nocturnal, with 97 % and 96 % of captures recorded during darkness, respectively.

Table 4.4. Estimates of activity pattern overlap (0= no overlap, 1=complete overlap) between four co- occurring small mammal species, with sample size and approximate 95 % bootstrap confidence intervals

Species Kernel Sample CI 95% Density size A. arktos: A. stuartii 0.848 91 0.772 – 0.915 A. arktos: R. fuscipes 0.794 91 0.721 – 0.858 A. arktos: M. cervinipes 0.739 91 0.656 – 0.826 A. stuartii: R. fuscipes 0.804 570 0.771 – 0.837 A. stuartii: M. cervinipes 0.779 570 0.725 – 0.814 R. fuscipes and M. cervinipes 0.883 661 0.850 – 0.915

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 119

Figure 4.5. Estimates of the relative daily activity patterns for each pair of sympatric species pooled across all five deployments. On the x axis time is shown in 24-hour time. In each separate plot, the dashed and solid lines represent the kernel density estimates for the indicated species. The degree of activity overlap between the two species is the area under the minimum of the two density estimates, as indicated by the shaded area in each plot. The estimate of overlap and confidence intervals are given in Table 4.4

120Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

4.5 DISCUSSION

4.5.1 Effectiveness of the camera trapping design Our study confirms that remote infrared digital camera traps can be successfully used to detect and differentiate small, closely related, morphologically similar mammal species, including the endangered black-tailed dusky antechinus. Previous studies have highlighted the difficulties in distinguishing between small co-occurring mammalian species (including Murids and antechinus) from images recorded by horizontally oriented camera traps [25, 39, 40, 48]. However, using a modified vertical mounting design established by De Bondi et al. [5] and extended here we conclude we could confidently identify most detections of small mammals to species. The vertical cameras captured dorsal perspectives of animals against a cleared ground surface, allowing critically diagnostic features such as body and head shape, tail length and shape, and ear morphology to be easily distinguished. The standardized dimensions of the bait holders and fixed distance of the cameras to the ground also provided a means to accurately estimate body size. Additionally, the 20-s video recordings corresponding to each still image allowed multiple angles of each individual to be viewed and its behaviour closely observed. We found that video recordings of behaviour (especially close-ups) were particularly useful and often allowed us to discriminate antechinus species with confidence. Although our findings are specific to the taxa and study area, the method could be applied to other small mammals such as civets, martens, shrews and other rodents.

However, despite the general success of this approach, we caution that prior live trapping at the site and / or a certain degree of familiarity with the target species was essential when attempting to taxonomically classify recorded individuals, even from paired monochrome images and video clips. We recommend that live trapping and field-based observations be used in concert with camera trapping, especially at new sites that have not been previously assessed. The present study also used two independent experts to examine all antechinus images and video, to improve accuracy of identification and ensure that the reviewer of the camera footage did not get complacent with identifications. Disagreement between independent experts resulted in the removal of just five possible antechinus detections from 665 paired images and video footage. The inclusion of a third independent reviewer may have allowed for a

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 121

majority vote in these cases where the two experts could not agree. We advise that future studies employ as many independent reviewers as possible.

Our infrared cameras also recorded a large number of false triggers compared with other studies [5, 39]. This was likely due to the high infrared motion sensitivity of our particular camera model even at its lowest sensitivity. Most false triggers were of large flies moving near the bait stations during daylight hours. This issue can easily be avoided by configuring cameras to operate only during the night. However, in our case this would have meant foregoing collection of important data on the diurnal activity of A. stuartii. The wide detection zone of our particular camera model (Ltl Acorn®) may have also resulted in a number of false positive images, caused by animals triggering a camera trap outside of the cameras field of view [49]. In some instances, the trigger speed (time taken to take a photograph after it has detected heat/motion) of the cameras may also have been too slow for rapidly moving animals, resulting in out-of-focus or only partial images, which impeded classification to species. Use of ‘white-flash’ cameras may circumvent these problems. Unlike infrared illumination, white flash provides the ability to take colour images at night and thus enable unique pelage attributes of A. arktos (i.e., fuscous black hindfeet and tail, orange eye ring and cheek patch) to be used as an additional diagnostic feature in separating them from the more uniformly brown A. stuartii.

4.5.2 Factors influencing detection probability Changes to detection probability with time invariably differs from one species to another, depending on their level of attraction to baited camera traps and other key biological factors such as population density and activity [20]. Therefore, unsurprisingly R. fuscipes, which occurs at high population density at our and other sites where it is found [30, 50] and has a strong attraction to peanut butter and oat bait [39], had a near constant detection rate between and within camera deployments. In contrast, results of our GLMMs show that both deployment number and days since deployment are important factors influencing the detection probabilities of A. arktos, A. stuartii and M. cervinipes. Both A. arktos and A. stuartii exhibited marked differences in detection probabilities between camera deployments: probabilities were higher during camera deployments #1 and #3 (pre- and during breeding) than deployments #4 and #5 (post-breeding). This finding is consistent with previous live-

122 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

trapping studies that found Antechinus spp. were significantly more trappable just prior to and during their breeding periods (owing to increased activity and movement during this time) and lowest post-breeding following the period of male die-off, when only pregnant females remain in the population [51, 52]. However, such low detection rates of both Antechinus spp. in our study during camera deployment #2 (when males were still alive), was unexpected. This period during the study experienced the highest total rainfall of any camera deployment (76 mm), which may have reduced the extent of activity in the antechinus [42, 53]. Moreover, periods of rainfall occurring early in the camera deployment may have degraded the olfactory attractiveness of our baits, similar to the reduced toxicity level of ‘1080’ bait after rainfall [54]. In contrast, we found that M. cervinipes exhibited a strong negative linear relationship of detection probability with respect to camera deployment number, suggesting either: 1) natural declines in either activity periods or population size occur between deployments; and / or 2) a persistent learned decrease of attraction to the bait, because the baits were inaccessible and thus offered no food reward, leading M. cervinipes to forage elsewhere for better opportunities. Optimal GLMMs for A. arktos, A. stuartii and M. cervinipes also identified a significant negative linear relationship between detection probability and days since camera deployment. Distinct peaks in detection rates on the first night of each camera deployment, when baits were fresh and most effective, suggest that the attractiveness of baits declines markedly over time [23].

Remote cameras have a distinct advantage over alternate survey methods such as live trapping because of the former’s longer operational periods without the need of human intervention [5, 55]. Ideally, with remote cameras such as those used in the present study, several weeks or even months of data can be collected with only two visits to a site: one to deploy the cameras and the second to retrieve or reset them. We noted that the initial high peak of interest in the camera trap baits (and lure combination) only lasted a single evening and was followed by progressively lower detection rates until the bait was replaced. To remain optimally effective, in our study baits would have needed to be refreshed every two-to-three days. For species with high initial detection rates over this time such as R. fuscipes (0.96-0.98), M. cervinipes (0.72-0.82) and A. stuartii (0.48-0.54), we found that a single camera deployment would be sufficient to achieve 95 % cumulative detection probability if the species was present. Conversely, for rarer species such as A. arktos (detection rate 0.16-0.20),

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 123

at least 14 survey days (five visits or successive three-day baiting deployments) would be necessary to achieve 95 % cumulative detection probability if present. This extra effort would markedly add to the cost of an occurrence survey or long-term monitoring project for A. arktos and similarly rare species. This latter finding has general significance for other camera trapping studies and will be further discussed below.

4.5.3 Diel activity and overlap Information about the date and time that our cameras’ images and videos were recorded allowed us to examine the activity patterns of the four target species. Both R. fuscipes and M. cervinipes were found to be completely nocturnal, displaying a unimodal peak in periods of activity, consistent with previous studies of these species [37, 48, 50]. A. arktos and A. stuartii also exhibited primary peaks in activity, although smaller activity peaks were evident later in the night. Ours is the first study of activity patterns in A. arktos, a species which was discovered only very recently [27]. We found A. arktos to be primarily nocturnal, an activity pattern consistent with most of its congeners [26] but notably at odds with some other members of the Dusky antechinus species complex [28, 37, 38]. In contrast, A. stuartii showed greater diurnal activity during its breeding season (74 % of all daytime observations occurred during this period). Such a shift in the timing of activity periods has also been reported for island populations of A. minimus and may be necessary to maximize reproductive success when competition among males for mating opportunities within populations is strong [38].

In general, we noted strong overlap of the activity periods in all four target species, with each displaying a peak in activity during the same two-hour period following sunset. In south-east Queensland, R. fuscipes, M. cervinipes and several Antechinus spp. are often sympatric and a certain level of interspecific competition is believed to occur [56]. The patterns of activity observed for the target species in our study suggest that diel temporal partitioning is not a mechanism used to promote coexistence. Rather, differences in microhabitat use and diet are likely to be the principal factors limiting competition among these species [36]. However, although we found no evidence of temporal partitioning in our study, a degree of avoidance at proximity may still occur. On several occasions (six and two times, respectively) video footage clearly showed A. stuartii fleeing from R. fuscipes and A. arktos to avoid

124 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

confrontation. Dickman [57] suggested that evasive action by a subordinate before an encounter with a dominant species could be due to early detection, either by sound or smell. Such a behavioural strategy could occur between the larger R. fuscipes and A. arktos and the smaller A. stuartii.

4.5.4 Applicability of camera traps as a survey method for rare or elusive species Our study demonstrates that infrared digital camera traps can be used to detect and identify A. arktos and other small, morphologically similar mammals at a rate comparable to live-trapping. Although our study was not specifically designed to compare survey methods, three nights of live trapping were conducted in August 2016 (four days prior to the first camera trap deployment) and so some general comments are warranted. Live trapping (600 trap nights) captured 41 A. stuartii, 7 A. arktos, 77 R. fuscipes and 58 M. cervinipes (including recaptures; Gray unpublished work). In comparison, the first three nights of camera trapping (deployment #1, 33 trap nights) recorded 68 A. stuartii, 15 A. arktos, 228 R. fuscipes and 117 M. cervinipes. Because live traps are limited to one capture per night, but cameras can record multiple individuals (and recaptures) per night, we expected that camera traps would record more species at more trap stations than live traps [58]. Future studies will aim to formally test the relative efficacy of live versus camera trapping for this species.

Nevertheless, such a high number of camera recordings from multiple stations highlights the potential use of this technique for other small mammal surveys and monitoring in the future. It is particularly relevant for areas inaccessible to large numbers of live traps or of high conservation value, where minimal disturbance to the target species is preferable. However, to achieve consistently high detectability of A. arktos, our work suggests baits need to be refreshed every two-to-three days, making the remote cameras more maintenance demanding than might be assumed. This limitation may not only apply to rare small mammals. Strongly declining rates of detection were also recorded in our study for A. stuartii and M. cervinipes and have been previously documented for other small mammals [22, 23, 39]. Examining fifteen camera trapping studies conducted on small-medium mammals over the last 10 years, we found that 80 % deployed cameras longer than 1 week and 67 % deployed cameras for at least 2 weeks before rebaiting / collection. This is an important finding for camera trapping generally, indicating that such routine camera deployment lengths

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 125

may not be optimal when targeting some small mammals. For management of threatened mammals, the problem of declining bait effectiveness is more acute because on any given deployment, probability of detection is already low. It is uncertain if other attractants that have been used in camera trap studies, such as linseed oil, truffle oil or vanilla essence may enhance the time length of optimal attractiveness [39, 59, 60]. In any case, we recommend including the time since camera deployment as a covariate in future surveys of small mammals, especially when cameras are operational for extended periods. In cases of declining detection for rare fauna, a standard three night live trapping survey (where species can be identified at point of capture and other ecological information obtained) may be a more practical detection method than a camera trapping survey, which may require multiple, successive three-day baiting deployments.

4.6 ACKNOWLEDGEMENTS

We would like to thank Aila Keto (Australian Rainforest Conservation Society) and the rangers at Springbrook National Park (QPWS) who provided logistical support and accommodation. We would also like to thank Todd Landers for his generous help during initial field trials. Finally, we thank the residents of Springbrook for their interest and all the volunteers who ventured out with us to help collect this data.

4.7 REFERENCES

1. Barnosky AD, Matzke N, Tomiya S, Wogan GOU, Swartz B, Quental TB, et al. 2011. Has the Earth’s sixth mass extinction already arrived? Nature 471:51– 57.

2. McCallum ML. 2015. Vertebrate biodiversity losses point to a sixth mass extinction. Biodiversity Conservation 24: 2497-2519.

3. Tasker EM, and Dickman CR. 2002. A review of Elliott trapping methods for small mammals in Australia. Australian Mammalogy 23: 77-87.

4. 4. Cunningham RB, Lindenmayer DB, MacGregor C, Barry S, and Welsh A. 2005. Effects of trap position, trap history, microhabitat and season on capture probabilities of small mammals in wet eucalypt forest. Wildlife Research 32: 657-671.

126 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

5. De Bondi N, White JG, Stevens M, Cooke R. 2010. A comparison of the effectiveness of camera trapping and live trapping for sampling terrestrial small-mammal communities. Wildlife Research 37: 456-465.

6. Rovero F, Zimmermann F, Berzi D, and Meek P. 2013. “Which camera trap type and how many do I need?” A review of camera features and study designs for a range of wildlife research applications. Hystrix, the Italian Journal of Mammalogy 24: 148-146.

7. McCallum ML. 2013. Changing use of camera traps in mammalian field research: habitats, taxa and study types. Mammal Review 43: 196-206.

8. Meek PD, Ballard G, Vernes K, and Fleming PJS. 2015. The history of wildlife camera trapping as a survey tool in Australia. Australian Mammalogy 27: 1- 12.

9. Meek PD, Ballard G, Fleming PJS, Schaefer M, Williams W, and Falzon G. 2014. Camera traps can be heard and seen by animals. PLoS ONE 9(10): e110832. https://doi.org/10.1371/journal.pone.0110832.

10. Meek PD, Ballard G, Fleming PJS, and Falzon G. 2016. Are we getting the full picture? Animal responses to camera traps and implications for predator studies. Ecology and Evolution 6: 3216-3225.

11. Claridge AW, Mifsud G, Dawson J, and Saxon MJ. 2004. Use of infrared digital cameras to investigate the behaviour of cryptic species. Wildlife Research 31: 645–650.

12. Rendall AR, Sutherland DR, Cooke R, and White J. 2014. Camera trapping: a contemporary approach to monitoring invasive rodents in high conservation priority ecosystems. PLOS One 9: doi:10.1371/journal.pone.0086592

13. Ridout MS, and Linkie M. 2009. Estimating overlap of daily activity patterns from camera trap data. Journal of Agricultural, Biological, and Environmental Statistics 14: 322-337.

14. Rowcliffe JM., Kays R, Kranstauber B, Carbone C, and Jansen PA. 2014. Quantifying levels of animal activity using camera trap data. Methods in Ecology and Evolution 5: 1170-1179.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 127

15. Smith JK, and Coulson G. 2012. A comparison of vertical and horizontal camera trap orientations for detection of potoroos and bandicoots. Australian Mammalogy 34: 196-201.

16. Wang SW, and Macdonald DW. 2009. The use of camera traps for estimating tiger and leopard populations in the high altitude mountains of Bhutan. Biological Conservation 142: 606-613.

17. Tobler MW, Carrillo-Percastegui SE, and Powell G. 2009. Habitat use, activity patterns and use of mineral licks by five species of ungulate in south-eastern Peru. Journal of Tropical Ecology 25: 261-270.

18. Díaz-Ruiz F, Caro J, Delibes-Mateos M, Arroyo B, and Ferreras P. 2016. Drivers of (Vulpes vulpes) daily activity: prey availability, human disturbance or habitat structure? Journal of Zoology 298: 128-138.

19. Glen AS, Cockburn S, Nichols M, Ekanayake J, and Warburton B. 2013. Optimising camera traps for monitoring small mammals. PloS ONE 8(6): e67940. doi:10.1371/journal.pone.0067940.

20. Claridge AW, and Paull DJ. 2014. How long is a piece of string? Camera trapping methodology is question dependent. Pp. 205-214 in Camera trapping: wildlife management and research (Meek PD, Ballard AG, Banks PB, Claridge AW, Fleming P, Sanderson JG, and Swann DE eds.). CSIRO Publishing. Melbourne, Australia. Claridge, AW., Paull, DJ., and Barry, SC. 2010. Detection of medium-sized ground-dwelling mammals using camera infrared digital cameras: an alternative way forward? Australian Mammalogy 32: 165- 171.

21. Taylor BD, Goldingay RL, and Lindsay JM. 2014. Horizontal or vertical? Camera trap orientations and recording modes for detecting potoroos, bandicoots and pademelons. Australian Mammalogy 36: 60-66.

22. McDonald PJ, Griffiths AD, Nano C, Dickman CR. 2015. Landscape-scale factors determine occupancy of the critically endangered central rock-rat in arid Australia: the utility of camera trapping. Biological Conservation 191: 93- 100.

128 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

23. Weerakoon MK, Ruffino L, Cleary GP, Heavener S, Bytheway JP, Smith HM, et al. 2014. Can camera traps be used to estimate small mammal population size? Pp 307-316 in Camera trapping: wildlife management and research. (Meek PD, Ballard AG, Banks PB, Claridge AW, Fleming P, Sanderson JG, and Swann DE eds.). CSIRO Publishing, Melbourne, Australia.

24. Meek PD, Vernes K, and Falzon G. 2013. On the reliability of expert identification of small-medium sized mammals from camera trap photos. Wildlife Biology in Practice 9: 1-19.

25. Meek PD, and Vernes K. 2016. Can camera trapping be used to accurately survey and monitor the Hastings River mouse (Pseudomys oralis)? Australian Mammalogy 38: 44-51.

26. Van Dyck S, and Strahan R. 2008. The mammals of Australia. New Holland Publishers, Sydney, Australia.

27. Baker AM, Mutton TY, Hines H, Van Dyck S. 2014. The Black-tailed Antechinus, Antechinus arktos sp. nov.: a new species of carnivorous marsupial from montane regions of the Tweed Volcano caldera, eastern Australia. Zootaxa 3765:101–133.

28. Baker AM., Mutton TY, Mason ED, and Gray EL. 2015. A taxonomic assessment of the Australian Dusky Antechinus Complex: a new species, the Tasman Peninsula Dusky Antechinus (Antechinus vandycki sp. nov.) and an elevation to species of the Mainland Dusky Antechinus (Antechinus swainsonii mimetes (Thomas)). Memoirs of the Queensland Museum Nature 59: 75-126.

29. McAllan BM, Dickman CR, Crowther MS. 2006. Photoperiod as a reproductive cue in the marsupial genus Antechinus: ecological and evolutionary consequences. Biological Journal of the Linnean Society 87:365– 365.

30. Gray EL, Baker AM, and Firn J. 2017. Autecology of a new species of carnivorous marsupial, the endangered black-tailed dusky antechinus (Antechinus arktos), compared to a sympatric congener, the brown antechinus (Antechinus stuartii). Mammal Research 62: 47-63.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 129

31. Nature Conservation Act 1992. Available at: https://www.legislation.qld.gov.au/LEGISLTN/CURRENT/N/NatureConA92 .pdf

32. Threatened Species Conservation Act 1995 No 101. Available at: http://www.legislation.nsw.gov.au/inforce/b7b85c46-520e-40b3-9439- b9d3108df6f5/1995-101.pdf

33. McArdle BH. 1990. When Are Rare Species Not There? Oikos 57: 276-277.

34. Hunter RJ. 2003. World heritage and associative natural values of the Central Eastern Rainforest Reserves of Australia. NSW National Parks and Wildlife Service, Government of New South Wales.

35. Queensland Herbarium. 2013. Regional Ecosystem Description Database (REDD). Version 6.1. Queensland Department of Science, Information Technology, Innovation and the Arts, Brisbane.

36. Gray EL, Burwell CJ, and Baker AM. 2016. Benefits of being a generalist carnivore when threatened by climate change: the comparative dietary ecology of two semelparous marsupials, including a new endangered species (Antechinus arktos). Australian Journal of Zoology 64: 249-261.

37. Hall S. 1980. The diets of two coexisting species of Antechinus (Marsupialia: Dasyuridae). Wildlife Research 7: 365–378.

38. Sale MG, and Arnould JPY. 2009. Spatial and temporal organization in the swamp antechinus: comparison between island and mainland populations. Journal of Mammalogy 90: 347-355.

39. Paull DJ, Claridge AW, and Barry SC. 2011. There’s no accounting for taste: bait attractants and infrared digital cameras for detecting small to medium ground-dwelling mammals. Wildlife Research 38: 188-195.

40. Claridge AW, Paull DJ, and Barry SC. 2010. Detection of medium-sized ground-dwelling mammals using camera infrared digital cameras: an alternative way forward? Australian Mammalogy 32: 165-171.

41. Zuur A, Leno EN, Smith GM. 2007. Analysing ecological data. Springer- Verlag, New York, U.S.

130 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

42. Wróbel A, and Bogdziewiez M. 2015. It is raining mice and voles: which weather conditions influence the activity of Apodemus flavicollis and Myodes glareolus? European Journal of Wildlife Research 61: 475-478.

43. Crawley MJ. 2013. The R book. John Wiley & Sons, Ltd, UK.

44. Bates D, Maechler M, Bolker B, and Walker S. 2015. Fitting Linear Mixed- Effects Models Using lme4. Journal of Statistical Software 67:1-48.

45. Carver BD, Kennedy ML, Houston AE, and Franklin SB. 2011. Assessment of temporal partitioning in foraging patterns of synoptic Virginia opossums and raccoons. Jouurnal of Mammalogy 92: 134-139.

46. Meredith M, and Ridout MS. 2016. Overview of the overlap package. Available: http://cran.at.r- project.org/web/packages/overlap/vignettes/overlap.pdf.

47. R Core Team. 2014. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R- project.org/.

48. Meek PD, Zewe F, and Falzon G. 2012. Temporal activity patterns of the swamp rat (Rattus lutreolus) and other rodents in north-eastern New South Wales, Australia. Australian Mammalogy 34: 223-233.

49. Meek PD, Ballard G, Fleming PJS. 2015. The pitfalls of wildlife camera trapping as a survey tool in Australia. Australian Mammalogy 37: 13-22.

50. Wood DH. 1971. The ecology of Rattus fuscipes and Melomys cervinipes (Rodentia: Muridae) in a south-east Queensland rain forest. Australian Journal of Zoology 19: 371-392.

51. Wood DH. 1970. An ecological study of Antechinus stuartii (Marsupialia) in a south-east Queensland rain forest. Austral Journal of Zoology 18:185–207.

52. Leung L. 1999. Ecology of Australian tropical rainforest mammals. I. The Cape York antechinus, Antechinus leo (Dasyuridae: Marsupialia). Wildlife Research 26:287–306.

53. Williams CT, Wilsterman K, Kelley AD, Breton AR, Stark H, Humphries MM., et al. 2014. Light loggers reveal weather-driven changes in daily activity

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 131

patterns of arboreal and semifossorial rodents. Journal of Mammalogy 95: 1230-1239.

54. McIlroy JC, Gifford EJ, and Carpenter SM. 1988. The effect of rainfall and blowfly larvae on the toxicity of ‘1080’ treated meat baits used in poisoning campaigns against wild dogs. Australian Wildlife Research 15: 473-483.

55. Welbourne DJ, MacGregor C, Paull D, and Lindenmayer D. 2015. The effectiveness and cost of camera traps for surveying small reptiles and critical weight range mammals: a comparison with labour-intensive complementary methods. Wildlife Research 42: 414-425.

56. Dickman CR, and Woodside DP. 1983. A test of a competition model with reference to three species of small mammals in south-eastern Australia. Oecologia 60: 127-134.

57. Dickman CR. 1991. Mechanisms of competition among insectivorous mammals. Oecologia 85: 464-471.

58. Driessen MM, and Jarman PJ. 2014. Comparison of camera trapping and live trapping of mammals in Tasmanian coastal woodland and heathland. Pp 253- 262 in Camera trapping: wildlife management and research. (Meek PD, Ballard AG, Banks PB, Claridge AW, Fleming P, Sanderson JG, and Swann DE eds.). CSIRO Publishing, Melbourne, Australia.

59. Diete RL, Meek PD, Dixon KM, Dickman CR, Leung LKP. 2015. Best bait for your buck: bait preference for camera trapping north Australian mammals. Australian Journal of Zoology 63: 376-382.

60. Claridge AW, Paull DJ, Cunningham RB. 2016. Oils ain’t oils: can truffle- infused food additives improve detection of rare and cryptic mycophagous mammals? Australian Mammalogy 38: 12-20.

132 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia. 133

134 Chapter 4: Can remote infrared cameras be used to differentiate small, sympatric mammal species? A case study of the black-tailed dusky antechinus, Antechinus arktos and co-occurring small mammals in southeast Queensland, Australia.

Chapter 5: General discussion

5.1 SYMPATRIC RELATIONSHIP BETWEEN A. ARKTOS AND A. STUARTII

Antechinus arktos, like other members of the dusky antechinus species complex (including A. mimetes, A. swainsonii and A. vandycki), are relatively large (38-120 g), predominantly ground dwelling species, which occur at low apparent density (see Chapter 2). In comparison, A. stuartii are significantly smaller than A. arktos (by 23- 70 g), but orders of magnitude more abundant (Chapter 2). Although both species consumed a broad range of invertebrate prey, the composition of invertebrates taken differed significantly, indicating that the species were most likely foraging largely in different microhabitats and / or at different times (Chapter 3). Subsequent camera trapping revealed that the activity times of the two species overlapped considerably (Chapter 4). Therefore, the observed differences in diet composition are most likely the result of differential habitat use. A. arktos likely forages extensively in topsoil and subsurface leaf litter. This is evinced by the species’ long claws (3.9 – 4.3 mm) which are specialized for digging (Baker et al. 2014), and the frequency and volume of Diptera larvae in the diet (Chapter 3). In contrast, A. stuartii likely forages regularly in trees. A. stuartii were often observed to climb tree ferns upon release from Elliot traps and remain above ground for extended periods (chapter 2). A. stuartii also frequently consumed Coleoptera, which make up much of the invertebrate biomass of trees (Scarff and Bradley 2006) and adult Lepidoptera; taxa that primarily occur underneath tree bark (Dickman 1986b). A. stuartii (both the northern and southern clade) are known to be semi-arboreal, with the species exhibiting arboreal behaviour in dry forests that have little groundcover and in New South Wales where A. mimetes occupies the forest floor (Cockburn and Lazenby-Cohen 1992; Crowther and Braithwaite 2008). Generally, invertebrate biomass is much greater in leaf litter on the ground surface than on tree trunks, contains larger prey taxa such as larvae and worms and requires less area be covered by foragers for an equivalent return of food (Dickman 1988; Fisher and Dickman 1993; Scarff and Bradley 2006). Thus, while the ground surface is plausibly the preferred foraging level of antechinus (Dickman 1988; Fisher and Dickman 1993), competition forces each species to specialize in the niche

Chapter 5: General discussion 135

dimension in which it has a competitive advantage, and to relinquish those in which another species outcompetes it (Schoener 1974, 1986). In this case, by A. stuartii obtaining food resources from the tree surface and canopy, competition at ground level with their significantly larger and more specialized terrestrial congeners (A. arktos with northern A. stuartii in the Tweed Volcano caldera; A. mimetes with southern A. stuartii in NSW further south) could potentially be avoided or reduced.

There is compelling evidence for such competitive interaction between sympatric antechinus near Canberra. In a series of direct manipulation experiments, Dickman (1986 a, b, 1988) demonstrated the asymmetric competitive advantage of A. mimetes (37-178 g) over the significantly smaller A. agilis (16-40 g). In the first experiment, and in comparison with an unmanipulated (control) community, removal of A. mimetes led to increased numbers, movements and home range areas of A. agilis and a significant increase in their use of complex habitat on the ground surface (Dickman 1986a). These shifts also coincided with the consumption of more terrestrial prey (larvae, Amphipoda) and fewer arboreal prey (Coleoptera, Hemiptera) by A. agilis as they were able to move into the microhabitats previously dominated by their larger congener (Dickman 1986a). Subsequent reintroduction of A. mimetes was then shown to reverse these demographic and resource shifts (Dickman 1986a). In the second experiment, also in comparison with a control community, the numbers of A. mimetes were reduced (Dickman 1986b). The reduction produced the same demographic and resource shifts in A. agilis but were less in magnitude than those produced by the total removal of the larger species as per experiment one (Dickman 1986b). These results suggest that even low numbers of A. mimetes are able to exert qualitatively similar competitive effects on A. agilis. Dickman (1986b) therefore hypothesised that interference (or exclusion, according to Schoener 1983) was the mechanism by which competitive interactions occurred between the two species. In exploitation competition, a resource can be shared by two species if it is abundant relative to the density of the dominant competitor (Dickman 1986b; Morin 2011). In comparison, interference is positively related to the frequency of contact with the dominant species, so that some competitive effects should be evident in communities even when resources are not in short supply and / or the density of the dominant competitor is low (Dickman 1986b, 1988). Interference has also been experimentally demonstrated in sympatric shrews from England, Poland and Russia (Dickman 1991;

136 Chapter 5: General discussion

Churchfield et al. 1999; Rychlik and Zwolak 2006), and may be a mechanism of competition among insectivorous small mammals in South America (Vieira and Paise 2011).

Although there has been no direct experimental confirmation of competition between southern A. stuartii and A. mimetes, in staged encounters in captivity, A. stuartii avoided direct contact with A. mimetes on a moment-to-moment basis indicating interference competition may also occur between these species (Righetti et al. 2000). Based on the results of the present study, a similar relationship is proposed to exist between A. arktos and northern A. stuartii. Video footage obtained from camera traps clearly showed A. stuartii fleeing from A. arktos (and R. fuscipes) before direct confrontation on multiple occasions (Chapter 4). As per Dickman (1986a,b, 1988), niche differentiation in these closely related species is likely promoted by size, with the larger A. arktos and A. mimetes most likely able to exclude A. stuartii from otherwise preferred habitat via interference competition, and force them to spend more time above ground regardless of their density or the abundance of resources. In a phylogenetic revision of the genus, Mutton (2016) further theorized that, “Competition from these larger antechinus might have resulted in a shift in breeding timing in sympatric A. stuartii, leading to their reproductive isolation and thus speciation of A. stuartii north and south”. Notably, the present study observed close but separate breeding timing in sympatric A. arktos and northern A. stuartii at Springbrook National Park (Chapter 2). Breeding synchronously at a specific time allows female antechinus to coincide lactation and weaning of offspring with peak prey availability (Fisher et al. 2013). Plausibly, reproductive phase differences may have arisen between these sympatric species from their dietary preferences for terrestrial or arboreal invertebrates that peak in abundance at different times, with allochronic isolation occurring as an important consequence (Wainer 1976; McAllan et al. 2006; Baker et al 2012). However, future work correlating terrestrial and arboreal invertebrate abundance with parturition dates of each species would be required to formally test this hypothesis. Such research should also address the comparative habitat use of these taxa e.g., through both ground and tree trapping (see Wood 1970 and Smith et al. 2017) and / or relating trap captures to microhabitat features, and uncover the breeding timing of sympatric A. arktos / northern A. stuartii at Lamington National Park and A. mimetes / southern A. stuartii in NSW, which have yet to be determined.

Chapter 5: General discussion 137

5.2 RARITY AND CONSERVATION STATUS OF A. ARKTOS

5.2.1 Distribution and habitat specificity Antechinus arktos is known only from one restricted area (the Tweed Shield Volcano caldera on the Qld-NSW border), encompassing Springbrook National Park in the north east, through eastern Lamington National Park, and south to Border Ranges National Park (Figure 5.1). Records from the 1980’s exist of dusky antechinus from other disjunct regions of the caldera rim at Mount Warning and Nightcap National Parks; however, no voucher specimens or tissue samples were taken so it is impossible to verify identification (Baker et al. 2014).

Figure 5.1. Distribution map of A. arktos indicating all museum specimen locations (dots) and all present day populations (asterisks). Adapted from Baker et al. (2014).

138 Chapter 5: General discussion

Confirmed museum specimens from the 1970’s and 1980’s indicate A. arktos once occurred in mountain mallee heath, open forest and rainforest with wet sclerophyll elements above 780 m (Supplementary Table 5.1). However, as of 2017, the species has only been confirmed from two locations > 950 m in cool temperate / subtropical cloud forest, with high annual rainfall (> 3000 mm) augmented by fog drip (Supplementary Table 5.1; Bureau of Meteorology 2017). This represents one of the most restricted ranges of any mainland Australian mammal (Woinarksi et al. 2014).

At the first location, within Springbrook National Park, A. arktos occurs at two sites (Best of All and Bilborough Court Lookouts) which are both situated at the peak of the plateau at close proximity (2.5 km straight-line distance). Individuals from both sites were found to be genetically identical based on mtDNA (cytb) (Mutton 2016); however, there was no observed demographic interchange between the two groups during the mark-recapture study and so they are defined here as subpopulations (IUCN 2016). At the second location, discovered as part of this PhD, A. arktos was found along three proximate transects (100-500 m apart) running parallel to the main Border Track, within Lamington National Park (Supplementary Table 5.1). Despite the close proximity of these two locations (~ 8km straight-line distance), individuals from Lamington National Park were subsequently found to be 0.8 % genetically divergent to those captured at Springbrook National Park based on mtDNA (cytb) (Andrew Baker pers.comm.). This suggests the two subpopulations have not been connected in recent times (tens of thousands of generations). Plausibly, clearing of native vegetation and unsuitable climate conditions (e.g., lower rainfall) at lower elevations connecting the two locations (i.e., the Numinbah Valley) could be a significant barrier to dispersal (Mutton 2016), and may have effectively led to the isolation of A. arktos at these two mountaintop locations.

However, even within Springbrook and Lamington National Parks, A. arktos was found to occur patchily. All known subpopulations occur in complex notophyll vine forest / simple microphyll fern forest communities on steep slopes or along ridgelines which contain damp peaty soil, fine leaf litter and outcrops of Hobwee Basalt (rounded boulders). At Lamington National Park, one individual was caught in rainforest dominated by Antarctic beech but the majority of captures (10 out of 11) occurred within the two transects with dense Helmholtzia understorey. Similarly,

Chapter 5: General discussion 139

within Springbrook National Park, trap success and density of A. arktos was significantly greater at Best of All compared to Bilborough Court Lookout (Chapter 2). Best of All Lookout consists of old growth Gondwanan rainforest, whereas Bilborough Court Lookout consists of regrowth rainforest. The Best of All site also contains a headwater gully with dense Helmholtzia understorey and a small stand of Antarctic beech. Indeed, the majority of live and camera trap captures of A. arktos (50 % and 34 % respectively) occurred along a single transect that crossed the gully containing Helmholtzia. The apparent preference of A. arktos for these wet slopes supporting the cloud-stripping stream lily at both Springbrook and Lamington National Parks may be linked to their optimal foraging areas. A. arktos was found to consume a high frequency and volume of Diptera larvae, specifically tipulid (crane fly) larvae (Chapter 3), which require moisture to survive (CSIRO 1991). Thus, they predominantly occur near streams / gullies in moist leaf litter, soil or mud (CSIRO 1991). Few other locations (e.g., Border Ranges and possibly Mount Warning and Nightcap National Parks) have the combination of high altitude, cloud forest and vegetation of this type (DECCW 2010). Therefore, the distribution of A. arktos is most likely limited to small, isolated pockets at the summit of the caldera.

5.2.2 Abundance To date, only 48 A. arktos individuals have been captured. This includes the ten museum voucher specimens collected between 1966 -1989 (Baker et al. 2014), and all the live individuals captured in surveys during the present study from Springbrook and Lamington National Parks (Chapter 2). Trap success in capturing A. arktos was consistently low throughout its limited range, especially outside their short pre- breeding / breeding window of peak activity, never exceeding 3.5 % (or 7 individuals per 200 trap nights; Chapter 2). This indicates that at known sites, the species occurs at low density and / or it is difficult to trap for other reasons, such as the size of the trapping grids, trap shyness or trap saturation from non-target species. However, the latter possibilities are less likely.

The trapping grids at Springbrook National Park each enclosed an area of 0.6 ha, a smaller area than generally employed for capturing large-bodied antechinus (Watt 1997; Sale et al. 2008; Smith et al. 2017). This is arguably a smaller trapping area than the species home range size and may have contributed to the low number of individuals captured. However, as discussed in the previous section, A. arktos has specific habitat

140 Chapter 5: General discussion

requirements and appears to be patchily distributed even within known habitat. Both sites within Springbrook National Park occur on steep slopes, with limited scope to extend the trapping grids into areas of suitable habitat. Therefore, it is unlikely that a statistically significant number of individuals were missed during surveys. In addition, during the mark-recapture study (chapter 2), recapture probability of A. arktos was relatively high (53.8 %) compared to other members of the dusky antechinus species complex (see Watt 1997), and the residency of many individuals lengthy (3-6 months). Therefore, there seems to be little reticence from individuals to re-enter traps after capture, even if they may be initially cautious. The double trapping design of the study also ensured that at least half of the traps set across the grid remained open and accessible to A. arktos each night (Chapter 2). Therefore, the low trap success of A. arktos is most likely due to low abundance. The similarly low trap success of A. arktos from camera trapping (Chapter 4), a less invasive detection method, seems to support this.

Antechinus arktos thus shows all three forms of rarity as defined by Rabinowitz (1981; Rabinowitz et al. 1986). That is, a narrow geographic distribution, a restricted habitat specificity, and low abundance at all sites. However, unlike some naturally rare species, A. arktos also appears to have undergone a significant range contraction in recent decades (Figure 5.1), suggesting a threatening process or threatening processes may have or be continuing to act on the species.

5.2.3 Evidence of decline The contraction of A. arktos is inferred from the location and collection date of confirmed museum specimens and a review of past and present (this study) survey effort in these areas. To summarize, between 1966 – 1989, ten confirmed A. arktos specimens (then identified as A. mimetes) were recorded within the Tweed Shield Volcano caldera from Springbrook, Lamington and Border Ranges National Parks (Supplementary Table 5.1). It was a further 24 years until another specimen was collected, when Baker et al. (2014) caught four A. arktos at Best of All Lookout (~100 m from the previous 1968 capture site) and discovered a second subpopulation further east along the caldera rim at Bilborough Lookout. Such a lack of recent records is unlikely to be an artefact of poor survey effort. During the same survey, Baker et al. (2014) failed to detect A. arktos from several other sites within Springbrook National Park at lower elevations. In addition, Parra Faundes (2014) did not detect A. arktos

Chapter 5: General discussion 141

throughout her PhD study on A. stuartii, despite conducting live-trapping less than 1 km downslope from the Bilborough Court Lookout subpopulation. Annual surveys (over the period 1994 – 2013) involving a total of 22, 650 trap nights at O’Reilly’s Guest House and the surrounding Green Mountains section of Lamington National Park (encompassing three previous capture sites) by colleague Dr. Ian Gynther, also failed to detect A. arktos (Baker et al. 2014). As did more limited general surveys (~2000 trap nights) of the eastern section of Border Ranges National Park, including previous capture locations around Brindle and Grady’s Creeks (Baker et al. 2014). Baker et al. (2014) theorized that A. arktos may have retracted to the highest elevations of cloud forest in Lamington and Border Ranges National Parks, as seemed likely to have occurred at Springbrook National Park (Baker et al. 2014).

During the present study, four targeted live-trapping surveys were conducted to test this hypothesis (Supplementary Table 5.2). To be 95 % confident of detecting A. arktos (if present) in a single survey requires 600 – 1800 trap nights. This was calculated using the equation:

N = log (1 – a) / log (1-p)

, where N= the number of sampling units taken, p= the probability of the species appearing in a single sampling unit (determined from mark-recapture data from Springbrook National Park; Chapter 2) and a= the confidence that the species will be detected in the sample (McArdle 1990).

However, once the breeding timing of A. arktos had been determined (Chapter 2), all surveys were undertaken within the species’ short pre-breeding / breeding window (July-August), when they were determined to be most active (see Chapter 2 and 4). During these months only 600 trap nights were required to be 95 % confident of detecting A. arktos in a survey.

Two surveys were conducted at Border Ranges National Park (2014 and 2016). Trapping was concentrated around previous capture sites near Brindle Creek in dense Helmholtzia gullies (800 m elevation) and higher elevation habitat near a stand of Antarctic beech (1000 m) to confirm the findings of previous general surveys. However, 275 trap nights were also conducted opportunistically within suitable habitat (i.e., complex notophyll vine forest / simple microphyll fern forest) on Bar Mountain (1120 m elevation). Despite a combined 2, 375 trap nights across these elevational

142 Chapter 5: General discussion

gradients, no A. arktos were caught (Supplementary Table 5.2). Another survey was conducted at Lamington National Park. Because of the already extensive trapping effort made at O’Reilly’s Guest House and the surrounding Green Mountains, trap effort was concentrated at the peak of the plateau. This area had not previously been the target of any trapping effort due to the difficulty of walking traps 10 km in along the Border track to reach suitable habitat. Eleven A. arktos (10 individuals) were subsequently captured over 600 trap nights (Supplementary Table 5.2). Finally, one survey was conducted at Mount Nardi within Nightcap National Park, where dusky antechinus (A. arktos or A. mimetes) had historically been captured but species identification was not confirmed. Apart from Mount Warning, which is largely inaccessible to targeted live trapping surveys due to topography and accessible only by foot, Nightcap National Park contains the only other known suitable habitat within the Tweed Shield Volcano caldera. However, despite 800 trap nights, no A. arktos were recorded (Supplementary Table 5.2).

The discovery of A. arktos in pockets of suitable habitat along the border track, at the peak of Lamington plateau, supports the notion of Baker et al. (2014) that A. arktos has retracted to the highest, wettest and coolest parts of their range at both Springbrook and Lamington National Parks. In addition, the absence of the species from previous capture sites at Brindle Creek after multiple general and targeted surveys (totalling 3, 300 trap nights) strongly suggests the species has undergone extirpation at lower elevations within Border Ranges National Park. Limited clearing has occurred near confirmed records of A. arktos; therefore, it is unlikely the observed range decline is due to habitat disturbance, alteration or fragmentation. Instead, the pattern of decline seems consistent with the impact of anthropogenic climate change (Baker et al. 2014). In the region of the Tweed Shield Volcano caldera, mean annual maximum temperatures have increased by 1° C and total rainfall has dropped by 76 mm between 1950 and 2003, with the rate of increase accelerating since the 1970’s (Hennessy et al. 2004). Under such a changing climate, declines may have been caused by withdrawal of A. arktos to match the upward altitudinal shift of suitable (wetter, colder) habitat (Chapter 3). Similar upward altitudinal shifts have been recorded for several ringtail possum species in the Wet Tropics World Heritage Area of northeast Queensland (Williams et al. 2014), bird species in the cloud forests of Monteverde, Costa Rica (Pounds et al. 1997) and multiple small mammals, including Belding’s

Chapter 5: General discussion 143

ground squirrel (Urocitellus beldingi), the alpine chipmunk (Tamias alpinus), the Pacific jumping mouse (Zapus princeps) and the water (Sorex palutris) of montane California, USA (Rowe et al. 2015).

The absence of A. arktos at Nightcap National Park was therefore not unexpected. Although the region has the highest average annual rainfall in New South Wales (> 2500 mm; NSW Parks and Wildlife Service 2017), the only areas of suitable habitat were found on Mount Nardi at 800 m elevation. This is notably lower than known A. arktos habitat at Springbrook and Lamington National Parks, which indicates that if A. arktos was indeed present in the region during the 1980’s it may have had limited opportunities to disperse into suitable higher elevation habitat as a result of climate warming.

Future surveys should target high elevation (> 1000 m) regions of suitable habitat within Border Ranges National Park, particularly gullies with dense Helmholtzia that occur at the summit of Bar Mountain. A preliminary survey of 300 trap nights failed to detect the species at such sites; however, further surveys of 600 trap nights should be undertaken during July-August in order to be 95 % confident of the species absence from the area.

5.2.4 Threatening processes Anthropogenic climate change is most likely the greatest current and future long- term threat to the black-tailed dusky antechinus. Climate changes predicted for the Tweed Shield Volcano caldera include: an average temperature increase of 1.3º C ± 0.6º C by 2030, a drop in average rainfall of 3.5 % ± 11 %, more severe and frequent extreme weather events (e.g. drought and flash flooding), and a rise in the cloud base set to reduce mist and therefore affect the availability and consistency of moisture (Hutley et al. 1997; McJannet et al. 2007; Australian National University 2009). Although there has been no formal research into the potential impacts of such climate changes on Gondwana Rainforest in the region (Australian National University 2009), it may be possible to draw on experience in the Wet Tropics of Queensland to make some inferences. For example, Hilbert et al. (2001) determined that complex notophyll vine forest and simple microphyll fern forest communities are particularly sensitive to climate change. They estimated that warming of only 1º C would decrease this habitat type by up to 50 %. If a similar trend occurred within the study region, it would represent a significant loss of core habitat for A. arktos, which is already restricted to

144 Chapter 5: General discussion

three small isolated sites at low density. In addition, the species has likely already retracted to the highest elevations of cloud forest within Springbrook and Lamington National Parks as a result of past warming (see evidence of decline). This is a serious problem; because they are already at the limit of their elevational range, A. arktos would need to shift latitudinally. However, the nearest similar habitat is at Ebor (~250 km south of the Tweed Shield Volcano caldera through unsuitable habitat), which is already occupied by A. mimetes, a similarly sized terrestrial congener (Baker et al. 2014).

The increasingly unpredictable rainfall and more frequent drought events predicted under climate change scenarios may also have significant negative impacts on A. arktos through declines or changes in the availability or seasonality of invertebrate prey. In the present study, annual variability in rainfall was the most likely explanation for the significant differences observed in trap success and density of A. arktos between years (Chapter 2), with A. arktos found to consume a lower diversity and volume of invertebrate prey during 2014 when rainfall fell below the 34-year average (Chapter 3). Invertebrate abundance is controlled by climate variables such as rainfall and temperature (Wilson et al. 2007). Short periods of dry weather are known to limit invertebrate abundance in tropical rainforest (Frith and Frith 1985, 1990), which in turn has been demonstrated to have strong negative impacts on the survival, condition and overall abundance of other antechinus species (see Parrott et al. 2007; Sale et al. 2008; Recher et al. 2009). Antechinus may also be especially vulnerable to climate change due to their life history. Antechinus species typically inhabit highly seasonal, predictable habitats and synchronise their short annual mating season such that lactation and weaning of offspring coincides with peak invertebrate abundance in spring / summer that follows the seasonal pattern of rainfall (Lee et al. 1982; Fisher et al. 2013). If, due to climate change, this pattern of rainfall was altered to no longer coincide with timing of lactation, the effect on A. arktos could be devastating, as antechinus breeding is regulated by photoperiod and thereby fixed in time regardless of food availability (McAllan et al. 2006; Hagger et al. 2013). The small litter size of A. arktos of six young, in comparison to congeners, which can produce up to 14 young (Van Dyck and Strahan 2008), may also limit the species ability to maintain a viable population size following any such population fluctuations.

Chapter 5: General discussion 145

The impacts of climate change may also operate synergistically with other threats. Currently, all known subpopulations of A. arktos (and much of their potential habitat) occurs within national park; therefore, habitat disturbance is considered only a minor threat. However, future upgrades to walking tracks and / or lookouts may threaten the species given their close proximity to the Border track at Lamington National Park and Best of All and Bilborough Court Lookouts at Springbrook National Park. Both feral cats and foxes are also known to occur in Springbrook and Lamington National Parks (DEHP 2017 a,b), with cats also detected on cameras at Best of All Lookout (Chapter 4). Predation by feral cats and foxes is considered a primary driver of many small-medium mammal declines and extinctions in Australia (Woinarksi et al. 2015; Doherty et al. 2017). Recently, Frank et al. (2014) provided the first experimental evidence that feral cats can cause localised extinctions of small native mammals on the mainland of Australia. Although the threat of feral cat and / or fox predation on A. arktos has not been directly demonstrated, both species are known to consume antechinus (Green and Osborne 1981; Dickman 1996; Molsher et al. 1999). In addition, A. arktos is a larger, ground dwelling species, which likely further increases their susceptibility to predation from these introduced predators (Dickman 1996). Considering the small population and geographic range size of A. arktos, even low-moderate predation on the species by feral cats and foxes could cause significant impact.

5.2.5 Assessment against EPBC Act and IUCN criteria In Australia, IUCN criteria has been broadly adopted by most states and territories and forms the basis for federal threatened species categorization under the Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) and Environment Protection and Biodiversity Conservation Regulations 2000 (EPBC Regulations) (Department of Environment and Energy 2015). As reviewed in chapter one, there are five IUCN criteria (A-E) used to determine a species threatened status that require information relating mostly to distribution, abundance and the rate of change of these parameters (Supplementary Table 1.1). A notable difference is that under the EPBC Act there is no equivalent of IUCN criteria D2 (restricted area of occupancy) and no near threatened or data deficient categories (Department of Environment and Energy 2015). To qualify for listing in any of the threatened

146 Chapter 5: General discussion

categories, a species need only meet one of the criteria; however, it should still be assessed against as many criteria as the data allows (IUCN 2016).

Since the commencement of this PhD thesis, A. arktos has been formally listed as Endangered at both Queensland and New South Wales state level (Nature Conservation Act 1992; Threatened Species Conservation Act 1995) and been recommended for listing federally as Endangered by Woinarski et al. (2014). However, at the time of these listings, A. arktos had only just been discovered and was therefore assessed based on preliminary data from the species naming paper (Baker et al. 2014). Woinarksi et al. (2014) categorized A. arktos endangered under criterion B, citing both a low extent of occurrence (EOO) and area of occupancy (AOO), continuing decline in population size and severe fragmentation (occurrence at just one location). However, they noted that the available data was “too limited to reliably estimate population size or to indicate the rate of decline” and suggested that as further information became available, the species could be eligible for threatened status under criterion C and possibly Criterion A (Woinarski et al. 2014).

The present PhD study has significantly increased the amount of ecological information available for A. arktos (see chapter 2, 3, 4 and above); thus, a reassessment against EPBC Act and IUCN criteria seems warranted. A full evaluation of the five EPBC and IUCN criteria against what is now known for A. arktos follows below. This up to date data will then be used to outline the management actions and future research required to ensure the long-term viability of this rare new species.

Criterion 1 (Figure 5.2): not eligible Confirmed A. arktos specimens have been recorded at 12 sites within the Tweed Shield Volcano caldera, constituting an EOO of 136 km2 and AOO of 40 km2 (Supplementary Figure 5.1; Department of Environment and Energy 2017). The EOO is defined as “the area contained within the shortest continuous imaginary boundary that can be drawn to encompass all known, inferred or projected sites of present occurrence of a taxon”, (IUCN 2016), and was calculated using a minimum convex hull, as per IUCN guidelines (IUCN 2016). The AOO is defined as the area within its EOO, which is occupied by a taxon, and was calculated using a 2x2 km grid cell method, as per IUCN guidelines (IUCN 2016). However, as of 2017, the species has only been confirmed from two locations constituting an EOO of 60 km2 and AOO of 20 km2 (Supplementary Figure 5.2; Department of Environment and Energy 2017).

Chapter 5: General discussion 147

This represents an overall decline of 55.9 % and 50 % respectively. However, although a decline in EOO and AOO is inferred, there is insufficient information to assess the rate or magnitude of the decline over the last 10 years, which may limit the applicability of this criterion.

Similarly, although continued decline in both EOO and AOO is expected to occur in the near future due to anthropogenic climate change (see above), there is currently insufficient information to project the rate or magnitude of such a decline.

Figure 5.2. Criterion one of EPBC Act, population size reduction. Measured over the longer of 10 years or 3 generations based on any of A1 to A4.

Criterion 2 (Figure 5.3): Eligible under B1 abc (iv) for listing as Critically Endangered The current EOO of the species is estimated at 60 km2 (Supplementary Figure 5.2; Department of Environment and Energy 2017). The species occurs at two locations; however, both subpopulations within Springbrook National Park are 0.8 % genetically divergent to A. arktos captured at Lamington National Park (Andrew Baker pers.comm.). This suggests the two locations have not been connected in recent times and should therefore be considered severely fragmented and thus critically endangered (under B1 a). Continuing decline, specifically in geographic distribution, is inferred based on comparison between recent and historic trapping data and projections of future climate change in the region. Due to the life history of the species (male die- off) there is also a major (up to 50 % or more) inherent annual fluctuation in the number

148 Chapter 5: General discussion

of mature individuals. This could plausibly reach an order of magnitude when combined with variations resulting from seasonal resource availability, as has been demonstrated in other antechinus species (Recher et al. 2009).

A. arktos is additionally eligible under B2 abc (iv) as Endangered. The current AOO of A. arktos is estimated at 20 km2 (Supplementary Figure 5.2; Department of Environment and Energy 2017). The species occurs at two locations (constituting three subpopulations) but they are considered severely fragmented, as per B1. Continuing decline and extreme fluctuations in the number of mature individuals are also inferred, as per B1.

Figure 5.3. Criterion two of EPBC Act. Geographic distribution as indicators for either extent of occurrence and/or area of occupancy.

Criterion 3 (Figure 5.4): Eligible under C2 (b) for listing as Endangered The available data are too limited to reliably estimate total population size; however, based on capture records from Baker et al. (2014), Woinarksi et al. (2014) conservatively estimated the number of mature individuals to be <2000. The geographic distribution of A. arktos is expected to undergo continued decline based on comparison between recent and historic trapping data and projections of future climate change in the region. Extreme fluctuation in the number of mature individuals is also expected due the species life history and vulnerability to stochastic events (e.g., drought), which have been demonstrated to influence seasonal resource availability in other antechinus species (Parrott et al. 2007; Sale et al. 2008; Recher et al. 2009).

Chapter 5: General discussion 149

Figure 5.4. Criterion three of EPBC Act. Population size and decline.

Criterion 4 (Figure 5.5): not eligible Woinarksi et al. (2014) estimated the number of mature individuals to be <2000. Following male die-off, the number of mature individuals is effectively halved, indicating the species could be considered vulnerable <1000. However, based on recent capture records and updated EOO and AOO estimates (Department of Environment and Energy 2017), the abundance of A. arktos could be as low as 500 mature individuals (Andrew Baker pers.comm.). Following male die-off, the species could be eligible for listing as endangered under this criterion. Although, further survey would be required to increase the reliability of this estimate.

Figure 5.5. Criterion four of EPBC Act. Number of mature individuals.

Criterion 5 (Figure 5.6): not eligible Population viability analysis has not been undertaken for this species. Consequently, there is insufficient information to determine the eligibility of the species for listing under this criterion.

150 Chapter 5: General discussion

Figure 5.6. Criterion Five of EPBC Act. Quantitative analyses.

In summary, A. arktos was found to be eligible for listing as Critically Endangered under Criterion 2. As per IUCN criteria (2016), only the highest category of threat (Critically Endangered) is officially used. However, additional criteria that the species qualifies (or nearly qualifies) for has been included in the above documentation for completeness and to aid future assessments.

5.2.6 Recommended management actions 1. Increase the public profile of A. arktos, as the species is not yet recognised in any published field guide and many field operators may not be aware of its existence. This could be achieved by designing informational posters on the species and displaying them within Springbrook and Lamington National Parks to educate residents and visitors.

2. Establish an integrated monitoring program across all known subpopulations to assess the species’ total population size and to monitor population trends and distribution.

3. Conduct further targeted surveys to confirm the species’ presence / absence in suitable habitat using a combination of detection techniques, including: live trapping, camera trapping, species-specific pheromone lures, and detection dogs (see below for further details).

4. Prevent future habitat disturbance. Any infrastructure upgrades should implement a risk management plan to ensure all risks to the species are identified and managed appropriately.

5. Assess the population-level impacts of feral cats and foxes on A. arktos and if necessary undertake control programs to eradicate feral cats and foxes from A. arktos habitat to reduce the impact of predation on the species.

Chapter 5: General discussion 151

6. Investigate the feasibility of establishing a captive breeding program. A. arktos is highly vulnerable to the impacts of climate change (see above) and has likely already undergone extirpations in the wild. Maintaining an insurance population would therefore provide security for the species in the event of future declines, extirpations or extinction in the wild. Antechinus species have already been bred in captivity to study their breeding biology (Marlow 1961; Woolley 1966) and would likely be cheaper to maintain in appropriate numbers for genetic management in comparison to larger bodied taxa (Balmford et al.1996).

7. Develop a model to predict suitable habitat for A. arktos under present and future climates. This will provide a quantitative estimate of the impact climate change will have on the species, which will be important for planning management actions. In particular, determining the urgency for captive breeding programs and the viability of translocations.

5.2.7 Directions for future research Currently, both live trapping and camera trapping methods are limited by the relative low population density of A. arktos and the use of standard bait and attractants (Chapter 2 and Chapter 4). In order to be 95 % confident of the species absence at a site, a high number of live trap nights (600-1800) or visits (five, three-day camera deployments) are required, which limits the number of sites that can be surveyed and monitored effectively. There is clearly a serious need to develop species-specific lures that increase relative detection probability or capture rate of A. arktos while ideally minimizing the detection or by-catch of non-target species.

Several early studies have highlighted the key role that olfactory cues play in small mammal detection. For example, Rowe (1970) and Boonstra and Krebs (1976) found that house mice Mus musculus and field voles Microtus agrestis were much more likely to be captured in traps containing the odour of conspecifics. Antechinus males are believed to use olfactory cues from a cloacal scent to locate females during the breeding period (Adrian Bradley pers.comm.). Future studies should investigate the potential to collect this pheromone from A. arktos females during the breeding period and deploy it as an attractant via filter paper into each live trap or camera bait

152 Chapter 5: General discussion

container instead of traditional bait. This pheromone has been previously used by documentary filmmakers to coerce antechinus males into walking along specified pathways during the breeding season (Downer et al. 2002), and may increase the relative detectability of A. arktos. If effective, in live-trapping studies the reduced by- catch of species such as R. fuscipes and M. cervinipes would mean that a higher proportion of traps remain open and available to capture A. arktos. Thus, fewer traps could potentially be used with a greater detection probability in any given deployment. Similarly, during the camera-trapping study, bait sites were often dominated by R. fuscipes and there was evidence of interference competition between R. fuscipes and A. stuartii. Using a generic antechinus or even arktos-specific lure would potentially reduce such activity, allowing antechinus less impeded access to cameras. Considerably less time would also be spent by the researcher sifting through camera trap recordings, making the process more cost-effective in terms of both time spent in the field and post-field image analysis.

Another emerging method being used to survey rare and cryptic species is scat detection dogs. Long et al. (2007) found scat detection dogs to be substantially more effective and efficient than remote cameras and hair snares for documenting the presence of black bears (Ursus americanus), fishers (Martes pennant) and bobcats (Lynx rufus). Although little-used thus far on small mammals, scat detection dogs have shown promise for detecting a range of species, including Franklin’s ground squirrel, Poliocitellus franklinii (Duggan et al. 2011); the black-footed ferret, Mustela nigripes (Reindl-Thompson et al. 2006) and more recently Australia’s threatened smoky mouse, Pseudomys fumeus (Lynn Baker and Linda Broome, pers.comms.). Scat detection dogs are trained using species-specific scat or bedding material and permit sampling to occur quickly and over larger areas than generally afforded by live or camera traps set at fixed stations (Glen et al. 2016). Although training scent-detection dogs requires significant investment of time and money (Long et al. 2007), this method could ultimately prove an effective means of identifying potential occupancy sites for A. arktos, which could then be followed up by live or camera trapping with the targeted A. arktos pheromone.

5.3 OVERALL STATUS OF THE GENUS ANTECHINUS

Compared to other groups of Australian mammal, small, nocturnal marsupials such as antechinus, which occur predominantly in mesic habitat, appear to have fared

Chapter 5: General discussion 153

relatively well in terms of declines to date (Wilson et al. 2003). Indeed, no antechinus are known to have become extinct since European settlement (Dickman et al. 2001). However, within the last three years, three antechinus species, and one subspecies, have been listed as threatened under Australian legislation. Antechinus bellus and A. minimus maritimus have been listed as vulnerable federally; A. argentus has been listed as vulnerable in Qld and A. arktos has been listed as endangered in both Qld and NSW, with federal listings also currently in review for both latter species. In addition, A. mimetes insulanus, which is found only within the confines of Grampians National Park and A. godmani, which is restricted to the Wet Tropics of north Queensland, have been identified as near threatened by Woinarski et al. (2014). The most recently described antechinus, A. vandycki is data deficient but may also be threatened due to its known occurrence within a single forest block (encompassing just 40 km2) on eastern Tasman Peninsula (Baker et al. 2015). As part of his phylogenetic revision of Antechinus, Mutton (2016) identified that A. mysticus populations were more genetically fragmented than previously thought and suggested they too may warrant listing in a threatened category. Even apparently common and secure antechinus species have experienced severe or underestimated declines since European settlement (e.g., Bilney et al. 2010; Johnstone et al. 2010; Lada et al. 2014). Thus, at least one third of known antechinus species are most likely threatened, with more likely to become so in the near future.

The extent of small mammal declines may have been previously underestimated because historical information on the distribution and abundance of mammals was biased towards larger taxa that were more easily observed by explorers and early settlers (Bilney et al. 2010). Long-term monitoring to establish baseline abundance and distribution parameters and to assess current population trajectories is therefore critical for such species. However, even for threatened species, monitoring is seldom done adequately. For example, Woinarksi et al. (2014) found that there was no monitoring for nearly a quarter of all Australian threatened, near threatened and data deficient terrestrial mammal species and when there was, the programs were often limited in their scope, duration, periodicity and integration with management. Nisbet (2007 Pp 790) noted a good reason for this lack of scientific engagement in monitoring: “Monitoring does not win glittering prizes. Publication is difficult, infrequent and unread”. Similarly, studies of specific aspects of a threatened species’ biology are

154 Chapter 5: General discussion

likely to be turned away by the editorial boards of high impact journals as being “parochial and of limited interest” (Fleming and Bateman 2016). However, it is ultimately these studies that contribute directly toward the recovery planning and conservation management of individual species.

The present study has provided detailed ecological information on the breeding biology, diet, diel activity and detectability of a new species of threatened marsupial, Antechinus arktos, and its sympatric relationship with northern A. stuartii about which little was previously known. Because of the differing size range and habitat preferences of the two species examined (A. arktos large and ground dwelling; A. stuartii smaller and partially arboreal) and the similar life history strategy of all antechinus species, many of the results of this thesis (e.g., optimal camera and live trapping methods) are applicable to other members of the genus. Thus, the findings of this study may not just have important implications for recovery plans aimed at A. arktos but may also assist in the planning of future conservation initiatives for A. argentus, A. bellus, and A. minimus maritimus.

5.4 REFERENCES

Australian National University. 2009. Implications of climate change for Australia’s World Heritage properties: a preliminary assessment. A report to the Department of Climate Change and the Department of Environment, Water, Heritage and the Arts by the Fenner School of Environment and Society, the Australian National University.

Baker AM., Mutton TY., Hines H., and Van Dyck S. 2014. The Black-tailed Antechinus, Antechinus arktos sp. nov.: a new species of carnivorous marsupial from montane regions of the Tweed Volcano caldera, eastern Australia. Zootaxa 3765:101– 133.

Baker AM., Mutton TY., Mason ED., and Gray EL. 2015. A taxonomic assessment of the Australian Dusky Antechinus Complex: a new species, the Tasman Peninsula Dusky Antechinus (Antechinus vandycki sp. nov.) and an elevation to species, the Mainland Dusky Antechinus (Antechinus mimetes new status). Memoirs of the Queensland Museum 59: 75-126.

Chapter 5: General discussion 155

Baker A., Mutton T., Van Dyck S. 2012. A new dasyurid marsupial from eastern Queensland, Australia: the buff-footed antechinus, Antechinus mysticus sp.nov. (Marsupialia: Dasyuridae). Zootaxa 3515:1–37.

Balmford A., Mace GM., and LeaderWilliams N. 1996. Designing the ark: setting priorities for captive breeding. Conservation Biology 10: 719-727.

Bilney RJ., Cooke R., and White JG. 2010. Underestimated and severe: small mammal decline from the forests of south-eastern Australia since European settlement, as revealed by a top-order predator. Biological Conservation 143: 52-59.

Boonstra R., and Krebs CJ. 1976. The effect of odour on trap responses in Microtus townsendii. Journal of Zoology 9: 467-476.

Bureau of Meteorology. 2017. Daily rainfall: Springbrook Road. Australian Government. Available at: http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=136&p_displa y_type=dailyDataFile&p_stn_num=040607&p_startYear

Churchfield S., Nesterenko VA., and Shvarts EA. 1999. Food niche overlap and ecological separation amongst six species of coexisting forest shrews (Insectivora: Soricidae) in the Russian Far East. Journal of Zoology 248: 349-359.

Cockburn A., and Lazenby-Cohen KA. 1992. Use of nest trees by Antechinus stuartii, a semelparous lekking marsupial. Journal of Zoology 226: 657-680.

Crowther MS., and Braithwaite RW. 2008. Brown antechinus, Antechinus stuartii. In: Van Dyck S., and Strahan R (eds). The mammals of Australia. New Holland Publishers, Australia.

CSIRO. 1991. The Insects of Australia 2nd edition. Melbourne University Press, Melbourne.

Department of Environment, Climate Change and Water New South Wales (DECCW). 2010. Border Ranges rainforest biodiversity management plan New South Wales and Queensland, Department of Environment, Climate Change and Water New South Wales, Sydney.

156 Chapter 5: General discussion

Department of Environment and Energy (DoEE). 2015. Threatened species status assessment manual. Australian Government Department of the Environment and Energy, Canberra. Available at: http://www.environment.gov.au/system/files/resources/b561c693-7412-423a-b82d- aefcefaaaee2/files/seap-manual.pdf

Department of Environment and Energy (DoEE). 2017. Area of occupancy and extent of occurrence for Antechinus arktos. Unpublished report, Australian Government Department of the Environment and Energy, Canberra.

Department of Environment and Heritage Protection (DEHP). 2017a. Introduced animals of Springbrook National Park, Wetland info. Queensland Government. Available at: https://wetlandinfo.ehp.qld.gov.au/wetlands/facts-maps/wildlife/?AreaID=national- park-springbrook&Kingdom=animals&SpeciesFilter=Introduced

Department of Environment and Heritage Protection (DEHP). 2017b. Wildlife of Lamington National Park, Wetlandinfo. Queensland Government. Available at:

https://wetlandinfo.ehp.qld.gov.au/wetlands/facts- maps/wildlife/?AreaID=national-park-lamington

Dickman CR. 1986a. An experimental study of competition between two species of dasyurid marsupials. Ecological Monographs 56: 221–241.

Dickman CR. 1986b. An experimental manipulation of the intensity of interspecific competition: effects on a small mammal. Oecologia 70: 536–543.

Dickman CR. 1988. Body size, prey size and community structure in insectivorous mammals. Ecology 69:569–580.

Dickman CR. 1991. Mechanisms of competition among insectivorous mammals. Oecologia 85: 464-471.

Dickman CR. 1996. Impact of exotic generalist predators on the native fauna of Australia. Wildlife Biology 2: 185-195.

Dickman CR. 2001. Ecological attributes and conservation of dasyurid marsupials in New South Wales, Australia. Pacific Conservation Biology 7: 124-133.

Chapter 5: General discussion 157

Doherty TS., Dickman CR., Johnson CN., Legge SM., Ritchie EG., and Woinarski JCZ. 2017. Impacts and management of feral cats Felis catus in Australia. Mammal Review 47: 83-97.

Downer J., Honeyborne J., Brownlow M., and Scholey K. 2002. Weird Nature, episode 2. John Downer Productions, BBC, UK.

Duggan JM., Heske EJ., Schooley RL., Hurt A., and Whitelaw A. 2011. Comparing detection dogs and livetrapping surveys for a cryptic rodent. The Journal of Wildlife Management 75: 1209-1217.

Fisher DO., and Dickman CR. 1993. Body size-prey size relationships in insectivorous marsupials: tests of three hypotheses. Ecology 74: 1871-1883.

Fisher DO., Dickman CR., Jones ME., and Blomberg SP. 2013. Sperm competition drives the evolution of suicidal reproduction in mammals. PNAS 110:17910–17914.

Fleming PA., and Bateman PW. 2016. The good, the bad, and the ugly: which Australian terrestrial mammal species attract the most research? Mammal Review 46: 241-254.

Frank A., Johnson C., Potts J., Fisher A., Lawes M., Woinarksi J., Tuft K., Radford I., Gordon I., Collis M., and Legge S. 2014. Experimental evidence that feral cats cause local extirpation of small mammals in Australia’s tropical savanna. Journal of Applied Ecology 51: 1486-1493.

Frith C., and Frith D. 1985. Seasonality of insect abundance in an Australian upland tropical rainforest. Australian Journal of Ecology 10: 237-248.

Frith D., and Frith C. 1990. Seasonality of litter invertebrate populations in an Australian upland tropical rainforest. Biotropica 22: 181–190.

Glen AS., Anderson D., Veltman CJ., Garvey PM., and Nichols M. 2016. Wildlife detector dogs and camera traps: a comparison of techniques for detecting feral cats. New Zealand Journal of Zoology 43: 127-137.

Green K., and Osborne WS. 1981. The diet of foxes, Vulpes vulpes, in relation to abundance of prey above the winter snowline in New South Wales. Australian Wildlife Research 8: 349-360.

158 Chapter 5: General discussion

Hagger V., Fisher D., Schmidt S., and Blomberg, S. 2013. Assessing the vulnerability of an assemblage of subtropical rainforest vertebrate species to climate change in south-east Queensland. Austral Ecology 38: 465–475.

Hennessy KJ., McInnes KL., Abbs DJ., Jones RN., Bathols J., Suppiah R., Ricketts J., Rafter T., Collins D., and Jones D. 2004. Climate change in New South Wales. Part 2: Projected changes in climate extremes. New South Wales Greenhouse Office, Climate Impact Group, CSIRO Atmospheric Research and National Climate Centre, Australian Government Bureau of Meteorology, Canberra.

Hilbert DW., Ostendorf B., and Hopkins MS. 2001. Sensitivity of tropical forests to climate change in the humid tropics of north Queensland. Austral Ecology 26: 590- 603.

Hutley LB., Doley D., Yates DJ., and Boonsaner A. 1997. Water balance of an Australian subtropical rainforest at altitude: the ecological and physiological significance of intercepted cloud and fog. Australian Journal of Botany 45: 311–329.

IUCN. 2016. Guidelines for using the IUCN red list categories and criteria. Version 12.

Johnstone CP., Reina RD., Lill A. 2010. Impact of anthropogenic habitat fragmentation on population health in a small, carnivorous marsupial. Journal of Mammalogy 91: 1332-1341.

Lada H., Thomson JR., Cunningham SC., Mac Nally R. 2014. Relating demographic characteristics of a small mammal to remotely sensed forest-stand condition. PLos ONE 9: doi:10.1371/journal.pone.0091731.

Lee AK., Woolley PA., and Braithwaite R. 1982. Life history strategies of dasyurid marsupials. In: Archer M (ed). Carnivorous marsupials. Sydney Royal Zoological Society of New South Wales Vol 1: 1-11.

Long RA., Donovan TM., Mackay P., Zielinski WJ., and Buzas JS. 2007. Comparing scat detection dogs, cameras, and hair snares for surveying carnivores. The Journal of Wildlife Management 71: 2018-2025.

Marlow B. 1961. Reproductive behaviour of the marsupial mouse, Antechinus flavipes (Marsupialia) and the development of the pouch young. Australian Journal of Zoology 9: 203-220.

Chapter 5: General discussion 159

McAllan BM., Dickman CR., and Crowther MS. 2006. Photoperiod as a reproductive cue in the marsupial genus Antechinus: ecological and evolutionary consequences. Biological Journal of the Linnean Society 87: 365-365.

McArdle BH. 1990. When Are Rare Species Not There? Oikos 57: 276-277.

McJannet D., Fitch WP., Disher M., and Reddell P. 2007. Water balance of tropical rainforest canopies in north Queensland, Australia. Hydrological Processes 21: 3473–3484.

Molsher R., Newsome A., and Dickman CR. 1999. Feeding ecology and population dynamics of the feral cat (Felis catus) in relation to availability of prey in central eastern New South Wales. Wildlife Research 26: 593-607.

Morin PJ. 2011. Community Ecology (2nd edition). John Wiley & Sons Publication, UK.

Mutton TY. 2016. Evolutionary biology of the Australian carnivorous marsupial Genus Antechinus. PhD thesis, Queensland University of Technology, Australia.

Nisbet E. 2007. Earth monitoring Cinderella science. Nature 450: 789-790.

NSW Parks and Wildlife Service. 2017. Nightcap National Park. NSW Parks and Wildlife Service. Available at: http://www.nationalparks.nsw.gov.au/visit-a- park/parks/nightcap-national-park/visitor-info#Weather,-temperature-and-rainfall

Parra Faundes D. 2014. Sex allocation, terminal investment and the effects of environmental constraints on maternal investment in subtropical antechinus. PhD thesis, University of Queensland, Australia.

Parrott ML., Ward SJ., Temple-smith PD., and Selwood L. 2007. Effects of drought on weight, survival and breeding success of agile antechinus (Antechinus agilis), dusky antechinus (A. swainsonii) and bush rats (Rattus fuscipes). Wildlife Research 34: 437-442.

Pounds JA., Fogden MPL., Campbell JH. 1999. Biological response to climate change on a tropical mountain. Nature 398: 611-615.

Rabinowitz D. 1981. Seven forms of rarity In: Synge H (ed) The biological aspects of rare plant conservation. John Wiley & Sons Ltd, USA.

160 Chapter 5: General discussion

Rabinowitz D., Cairns S., and Dillon T. 1986. Seven forms of rarity and their frequency in the flora of the British Isles. In: Soulé ME (ed) Conservation biology, the science of scarcity and diversity. Sinauer Associates INC, USA.

Recher HF., Lunney D., and Matthews A. 2009. Small mammal populations in a eucalypt forest affected by fire and drought. I. Long-term patterns in an era of climate change. Wildlife Research 36: 143-158.

Reindl-Thompson SA., Shivik JA., Whitelaw A., Hurt A., and Higgins KF. 2006. Efficacy of scent dogs in detecting Black-footed ferrets at a reintroduction site in South Dakota. Wildlife Society Bulletin: 34: 1435-1439.

Righetti J., Fox BJ., and Croft DB. 2000. Behavioural mechanisms of competition in small dasyurid marsupials. Australian Journal of Zoology 48: 561-576.

Rowe FP. 1970. The response of wild mice (Mus musculus) to live traps marked by their own and foreign mouse odour. Journal of Zoology 162: 517-520.

Rowe KC., Rowe KM., Tingley MW., Koo MS., Patton JL., Conroy CJ., Perrine JD., Beissinger SR., and Moritz C. 2015. Spatially heterogeneous impact of climate change on small mammals of montane California. Proceedings of the Royal Society B 282: 20141857. http://dx.doi.org/10.1098/rspb.2014.1857

Rychlik L., and Zwolak R. 2006. Interspecific aggression and behavioural dominance among four sympatric species of shrews. Canadian Journal of Zoology 84: 434-448.

Sale MG., Wilson BA., Arnould JPY. 2008. Factors influencing population dynamics in island and mainland populations of the swamp antechinus (Antechinus minimus). Australian Journal of Zoology 56: 187-194.

Scarff FR., and Bradley JS. 2006. Invertebrate prey of the bark-foraging insectivore Phascogale tapoatafa: distribution of biomass amongst alternative foraging substrates within south-western Australian woodlands. Australian Journal of Zoology 54: 335-341.

Schoener TW. 1974. Resource partitioning in ecological communities. Science 185: 27-39.

Schoener TW. 1983. Field experiments on interspecific competition. American Naturalist 122: 240-285.

Chapter 5: General discussion 161

Schoener TW. 1986. Resource partitioning. In: Kikkawa J., and Anderson DJ (eds) Community ecology: pattern and process. Blackwell Scientific Publications, Australia.

Smith GC., Means K., and Churchill S. 2017. Aspects of the ecology of the Atherton antechinus (Antechinus godmani) living in sympatry with the rusty antechinus (A. adustus) in the Wet Tropics, Queensland- a trapping and radio-tracking study. Australian Mammalogy https://doi.org/10.1071/AM16050

Van Dyck S., and Strahan R. 2008. The Mammals of Australia. New Holland Publishers, Sydney.

Vieira EM., and Paise G. 2011. Temporal niche overlap among insectivorous small mammals. Integrative Zoology 6: 375-386.

Wainer JW. 1976. Studies of an island population of Antechinus minimus (Marsupialia: Dasyuridae). Australian Journal of Zoology 19:1–7.

Watt A. 1997. Population ecology and reproductive seasonality in three species of Antechinus (Marsupialia: Dasyuridae) in the wet tropics of Queensland. Wildlife Research 24: 531–539.

Williams SE., Scheffers BR., VanDerWal J., Roslan N., Anderson A., Storlie C., Brunner R., Welbergen J., and Parsons S. 2014. Species resilience: the key to understanding biodiversity in the rainforests of the Australian Wet Tropic. Report to the National Environment Research Program. Reef and Rainforest Research Centre Limited, Cairns.

Wilson BA., Dickman CR., and Fletcher TP. 2003. Dasyurid dilemmas: problems and solutions for conserving Australia’s small carnivorous marsupials. In: Jones M., Dickman CR., and Archer E (eds). Predators with pouches. CSIRO Publishing, Australia. Pp 407-421.

Wilson RD., Trueman JWH., Williams SE., and Yeates DK. 2007. Altitudinally restricted communities of schizophoran flies in Queensland’s Wet Tropics: vulnerability to climate change. Biodiversity Conservation 16: 3163-3177.

Woinarski JCZ., Burbidge AA., and Harrison PL. 2014. The action plan for Australian mammals 2012. CSIRO Publishing, Australia.

162 Chapter 5: General discussion

Woinarski JCZ., Burbidge AA., and Harrison PL. 2015. Ongoing unravelling of a continental fauna: decline and extinction of Australian mammals since European settlement. PNAS 112: 4531-4540.

Wood DH (1970) An ecological study of Antechinus stuartii (Marsupialia) in a south-east Queensland rain forest. Australian Journal of Zoology 18:185–207.

Woolley PA. 1966. Reproduction in antechinus spp. And other dasyurid marsupials. Symposium of the Zoological Society of Australia 15: 281-294.

Chapter 5: General discussion 163

Supplementary material

Supplementary Table 1.1 IUCN criteria for listing species in a threatened category (taken from IUCN 2016).

Supplementary material 165

Supplementary Figure 4.1. Example of the type of images obtained, and the ease of identifying diagnostic features (e.g., body size, body and head shape, and ear morphology) from vertically oriented cameras. Species include (clockwise from top left); A) A. stuartii, B) A. arktos, C) R. fuscipes, and D) M. cervinipes. Note antechinus are smaller in size and have more pronounced pointed snouts compared to the Muridae. A. arktos is larger, with a more rounded rump than A. stuartii; while, M. cervinipes has a shorter face than R. fuscipes and is smaller in size. R. fuscipes also has distinctive large rounded ears and coarser looking fur compared to M. cervinipes.

166 Supplementary material

Supplementary Table 5.1. Geographic location of all A. arktos captures including year, habitat, proximity to type locality, and altitude (m) information.

Location Year Habitat Proximity Altitude ID to type (m) locality Queensland: Dave’s Creek, 1966 Mountain mallee heath on rhyolitic soils surrounded by rainforest 6km north- 840 JM836 (QM) Lamington NP, 28° and wet sclerophyll forest west 13˝ S, 153° 13´ E Best of All Lookout, 1968 Patch of Antarctic beech in complex notophyll vine forest on ~100m 1000 JM1594 Springbrook NP, 28° Cainozoic igneous rocks 14' S, 153° 16' E Warrie Circuit, 1971 Graded walking track immediately adjacent to an area of tall, 2.5km 780 JM1595 Springbrook NP, 28° layered open forest on Cainozoic igneous rocks dominated by Blue north-east 13' S, 153°16' E Mountain ash (Eucalyptus oreades) and New England ash (E. campanulata) Binna Burra, 1974 Patch of rainforest containing a number of mature eucalypts and 9km north- 780 JM835 Lamington NP, 28° brush box, adjacent to the road at the park entrance. Open forest on west 12´ S, 153° 11´ E Cainozoic igneous rocks Near O’Reilly’s 1975 Stand of regenerating rainforest (previously complex notophyll vine 13km west 900 JM834 Guest House, forest on Cainozoic igneous rocks). Dominant plant species was Lamington NP, 28° native guava, Rhodomyrtus psidioides. 13' S, 153° 07' E Lamington NP, 28° 1975 Walking track surrounded by complex notophyll vine forest on 11km west 1000 N11506 15' S, 153° 09' E Cainozoic igneous rocks

Supplementary material 167

Near O’Reilly’s 1977 Cleared area 100 m from rainforest at O’Reilly’s Guesthouse, 12.5km 900 JM2281 Guest House, 28° 14' adjacent to Lamington National Park. This site was previously west S, 153° 08' E complex notophyll vine forest on Cainozoic igneous rocks Best of All Lookout, 2013 Complex notophyll vine forest on Cainozoic igneous rocks, beside Type 950 QMJM20009 Springbrook NP, 28° a headwater gully with stream lily, Helmholtzia glaberrima. Also a locality 14´ S 153° 15´ E prominent stand of simple microphyll fern forest with Antarctic beech. Bilborough Court 2013 Regrowth complex notophyll vine forest and simple microphyll 2.5 km east 950 QMJM20010 Lookout, fern forest on Cainozoic igneous rocks. Contains a low canopy, Springbrook NP, 28° dense vines and rocky substrate. 14´ S 153° 17´ E New South Wales: Brindle Creek, 1988 Cool-subtropical rainforest equivalent to complex notophyll vine 24 km 800 JM7949 Border Ranges NP, forest on Cainozoic igneous rocks south-west 28° 22´ 36˝ S 153° 04´ 06˝ E Grady’s Creek, 1989 Walking track ~100 m north of Brindle Creek Rd surrounded by 21 km 1000 M20605 Border Ranges NP, Cool-subtropical rainforest equivalent to complex notophyll vine south-west 28° 22´ 32˝ S 153° forest on Cainozoic igneous rocks 06´ 22˝ E Border Ranges NP, 1989 Cool temperate rainforest equivalent to simple microphyll fern 21 km 900 M20606 28° 21´ 50˝ S 153° forest with Antarctic beech on Cainozoic igneous rocks south-west 06´ 15˝ E Border Track, 2015 Simple microphyll fern forest with Antarctic beech on Cainozoic 8 km west 1165-1200 QM Lamington NP, 28° igneous rocks. Some sites had dense Helmholtzia glaberrima 15´ S 153° 10´ E understorey, outcrops of rounded boulders and peaty soil. Type locality highlighted in bold. QM=Specimen held at the Queensland Museum, M=Specimen held in the Australian Museum, N=Specimen held by Queensland Parks and Wildlife Service

168 Supplementary material

Supplementary Table 5.2 Additional live-trapping field data at Lamington, Border Ranges and Nightcap National Parks, 2014-2016.

National Park General location/s Altitude (m) Date No. of No. of Trap No. of A. transects nights arktos captures Border Ranges Brindle Creek walking track near Antarctic 800-1120 10-15th June 2014 4 1525 0 beech picnic area, Helmholtzia loop walking track and Bar Mountain circuit Lamington Border track (north-eastern slope of Mt 1165-1200 12-14th August 8 600 11 (4F, 7M) Bithongabel) between Toolona and Wanungra 2015 Lookouts Border Ranges Brindle Creek walking track near Antarctic 800-900 26-29th July 2016 4 800 0 beech picnic area and Helmholtzia loop walking track Nightcap Mount Nardi along three walking tracks: 800 13-16th August 6 800 0 Mount Matheson Loop, Pholis Gap and 2016 Historic Nightcap walking track

Supplementary material 169

Supplementary Figure 5.1. Extent of occurrence (EOO) and area of occupancy (AOO) of A. arktos (all records). Provided by Department of the Environment and Energy (2017).

170 Supplementary material

Supplementary Figure 5.2 Extent of occurrence (EOO) and area of occupancy (AOO) of A. arktos 1997-2017. Provided by Department of the Environment and Energy (2017).

Supplementary material 171